Intel Corp. v. Tela Innovations, Inc.

Full title: INTEL CORPORATION, Plaintiff, v. TELA INNOVATIONS, INC., Defendant.

Court: UNITED STATES DISTRICT COURT NORTHERN DISTRICT OF CALIFORNIA

Date published: Dec 22, 2020

Case No. 3:18-cv-02848-WHO (N.D. Cal. Dec. 22, 2020)

Opinion

Case No. 3:18-cv-02848-WHO

12-22-2020

INTEL CORPORATION, Plaintiff, v. TELA INNOVATIONS, INC., Defendant.

William H. Orrick United States District Judge

ORDER ON MOTIONS FOR SUMMARY JUDGMENT; MOTIONS TO STRIKE AND EXCLUDE; ADMINISTRATIVE MOTIONS TO SEAL

Re: Dkt. Nos. 217, 218, 219, 221, 228, 230, 234, 236, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 253, 255, 257, 259, 260, 261, 262, 270, 272, 274, 276

INTRODUCTION

Tela Innovations, Inc. ("Tela"), the defendant and counterclaimant, owns (as relevant here) five patents related to the design and manufacture of semiconductor chips. Intel Corporation ("Intel"), the plaintiff and counter-defendant, filed this action seeking declaratory judgment that several of its products did not infringe Tela's patents and that the patents are unenforceable.

Before me are eight substantive motions. Each party has moved for summary judgment on several issues. Intel seeks summary judgment that its products do not infringe on two grounds and that two of Tela's patents are invalid. Tela seeks summary judgment on a group of Intel's claims and defenses that depend on Intel's contention that Professor Lawrence Pileggi—who assigned his putative rights in the patents to Intel—is an unnamed inventor of the patents. Tela also seeks summary judgment on Intel's claims and defenses that allege Tela breached a covenant not to sue between the parties. Additionally, each party has filed three motions to strike the opinions or exclude the testimony of the other party's experts.

Intel's motion for summary judgment is granted. It has shown its products do not infringe Tela's patents because they lack "diffusion regions" and "gate contact structures" that the asserted claims require. It has also shown that two of Tela's patents are invalid for failure to satisfy the statutory written description and enablement requirements. Tela's motion for summary judgment is granted in part and denied in part. It has shown that it did not breach the covenant not to sue. The inventorship issue, however, is dependent on genuine disputes of material fact not appropriate for summary judgment. The motions to strike and exclude are resolved as described below.

BACKGROUND

I. THE TECHNOLOGY

These motions concern U.S. Patent Nos. 7,943,966 ("the '966 Patent"), 7,948,012 ("the '012 Patent"), 10,141,334 ("the '334 Patent"), 10,141,335 ("the '335 Patent"), and 10,186,523 ("the '523 Patent") (collectively, the "Asserted Patents"). The Asserted Patents are all owned by Tela, are part of the same patent family, and claim priority to U.S. Provisional Patent Application No. 60/781,288 ("the '288 Provisional"). The Asserted Patents were issued between May 2011 and January 2019; the '288 Provisional was filed March 9, 2006.

I describe the aspects of the Asserted Patents relevant to this dispute in greater detail below. As a general matter, the Asserted Patents concern fabrication of semiconductor chips. See Claim Construction Order [Dkt. No. 175] 4. The fabrication of circuits like those described in the Asserted Patents occurs layer by layer. Id. "Lithography" is one part of the fabrication process in which light is used to pattern shapes onto a material. Id. 4-5. In general, as technology has advanced, the size of the features on the devices has gotten smaller and smaller. The "lithographic gap" is the phenomenon in which feature sizes are smaller than the wavelength of the light that is patterning them. Id. 5. The Asserted Patents aim to improve the design and manufacturability of integrated circuits by helping resolve the lithographic gap. Id.

This summary judgment dispute relates to semiconductor chips made by Intel using what it refers to as 22 nm, 14 nm, and 10 nm fabrication technologies (collectively, the "Accused Products"). Again, the aspects of these technologies relevant to the dispute are discussed in detail below. Generally, Intel argues that the state of the art, and its Accused Products, have progressed beyond what is reflected in the Asserted Patents and that the Accused Products do not infringe.

Several concepts require explanation for context and to aid understanding. The substantive discussion of the relevant aspects of the technology is below; this explanation should not be understood to describe in detail integrated circuits generally or the Asserted Patents or Accused Products specifically. Integrated circuit chips are a fundamental element of modern technology like computers and smartphones. As noted, chips like those at issue here are fabricated in layers. The base layer is a semiconductor material (usually silicon) called the "substrate." To generate an electrical current, impurities (for instance, boron) are introduced within the substrate. These impurities act as "sources" and "drains" that are charged, permitting an electrical current to flow. The flow is controlled by a "gate," sometimes referred to as "poly" (because they were often made with polysilicon); in the Asserted Patents, the gate is in a different layer than the substrate. Additionally, a "gate contact" layer or structure is placed above the gate. Last, interconnect layers—sometimes referred to as "metal" or "M" with designators such as "Metal 1" or "M1" depending on the layer number—connect the bottom layers to higher ones that are added.

II. THE PARTIES' HISTORY

On March 1, 2005, Scott Becker, the CEO of Tela and named inventor on the Asserted Patents, met with Carnegie Mellon Professor Lawrence Pileggi. What occurred at this meeting is described below in the discussion on Tela's motion for summary judgment. Both parties agree that Pileggi and his graduate students presented about technology they were working on—some of which Pileggi was attempting to commercialize through a company he founded called Fabbrix. The meeting occurred under a confidentiality agreement that Becker signed. Pileggi and Intel claim that, in this meeting, Pileggi presented concepts that Becker then used as an integral part of the Asserted Patents. Tela disputes this. Pileggi sold Fabbrix to a company called PDF in 2007. Pileggi worked with PDF for four more years, although the parties dispute his level of involvement with the company.

In June 2008, PDF filed a patent protest to an application that was then pending to which the Asserted Patents claim priority. The protest included alleged prior art references, including from Pileggi and Fabbrix. In May 2007, Intel invested in Tela. The parties also entered into a covenant not to sue (the "CNTS"). In June 2008, Intel considered a follow-on investment; it requested and was given a legal analysis from Tela regarding PDF's protest. In January 2019, after this suit was filed, Pileggi licensed his purported rights in the Asserted Patents to Intel.

III. PROCEDURAL HISTORY

I have discussed the history of this case in several previous orders. See, e.g., Dkt. Nos. 162, 175. Intel originally filed this action in May 2018 for a declaratory judgment of noninfringement or unenforceability of a group of patents owned by Tela and has twice amended its complaint, altering the patents at issue. Dkt. Nos. 1, 38, 119. Tela eventually brought counterclaims against Intel and Intel brought counterclaims in reply. Dkt. Nos. 83, 101, 118, 164, 169, 170. There have since been several motions to dismiss, a motion to transfer venue, a series of discovery disputes, and a claim construction. Additionally, there has been a co-pending proceeding before the International Trade Commission ("ITC") about much of the same subject matter.

Here, Intel moves for summary judgment of noninfringement on the Asserted Patents. It also moves for summary judgment that the '334 and '335 patents are invalid. Tela seeks summary judgment on several of Intel's claims and defenses. Specifically, some of Intel's claims and defenses argue that the technology in the Asserted Patents was invented, in part, by Pileggi. See Second Amended Complaint ("SAC") [Dkt. No. 119] at 39-49; Intel's Answer to Tela's Counterclaims and Counterclaims in Reply ("CIR") [Dkt. No. 164] at 18-19. Tela seeks summary judgment on these inventorship claims and defenses because, it argues, Pileggi did not invent the technology, the claims fail as a matter of law, and the claims are barred by laches and equitable estoppel. Intel also argues that infringement claims against it based on the Asserted Patents are barred by the CNTS. Tela seeks summary judgment on Intel's claims and defenses based on the CNTS because, it contends, it does not cover the Asserted Patents. Each party has also filed three motions to strike the reports or exclude the testimony of the other party's experts on numerous grounds. I held a hearing on December 2, 2020, on all pending motions.

LEGAL STANDARD

I. SUMMARY JUDGMENT

A. Generally

Summary judgment on a claim or defense is appropriate "if the movant shows that there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law." FED. R. CIV. P. 56(a). In order to prevail, a party moving for summary judgment must show the absence of a genuine issue of material fact with respect to an essential element of the non-moving party's claim, or to a defense on which the non-moving party will bear the burden of persuasion at trial. See Celotex Corp. v. Catrett, 477 U.S. 317, 323 (1986). Once the movant has made this showing, the burden then shifts to the party opposing summary judgment to identify "specific facts showing there is a genuine issue for trial." Id. The party opposing summary judgment must then present affirmative evidence from which a jury could return a verdict in that party's favor. Anderson v. Liberty Lobby, 477 U.S. 242, 257 (1986).

On summary judgment, the Court draws all reasonable factual inferences in favor of the non-movant. Id. at 255. In deciding a motion for summary judgment, "[c]redibility determinations, the weighing of the evidence, and the drawing of legitimate inferences from the facts are jury functions, not those of a judge." Id. However, conclusory and speculative testimony does not raise genuine issues of fact and is insufficient to defeat summary judgment. See Thornhill Publ'g Co., Inc. v. GTE Corp., 594 F.2d 730, 738 (9th Cir. 1979).

B. Noninfringement

Summary judgment of noninfringement requires a two-step analysis. "First, the claims of the patent must be construed to determine their scope. Second, a determination must be made as to whether the properly construed claims read on the accused device." Pitney Bowes, Inc. v. Hewlett-Packard Co., 182 F.3d 1298, 1304 (Fed. Cir. 1999) (internal citations omitted). "The determination of infringement, both literal and under the doctrine of equivalents, is a question of fact." Lockheed Martin Corp. v. Space Sys./Loral, Inc., 324 F.3d 1308, 1318 (Fed. Cir. 2003); see also Kilopass Tech. Inc. v. Sidense Corp., No. 10-cv-02066-SI, 2012 WL 3545286, at *4 (N.D. Cal. Aug. 16, 2012). Because the ultimate burden of proving infringement rests with the patentee, an accused infringer may show that summary judgment of noninfringement is proper either by producing evidence that would preclude a finding of infringement, or by showing that the evidence on file fails to create a material factual dispute as to any essential element of the patentee's case. See Novartis Corp. v. Ben Venue Labs., Inc., 271 F.3d 1043, 1046 (Fed. Cir. 2001). "Summary judgment of noninfringement may only be granted if, after viewing the alleged facts in the light most favorable to the nonmovant and drawing all justifiable inferences in the nonmovant's favor, there is no genuine issue whether the accused device is encompassed by the patent claims." Id. "Literal infringement requires the patentee to prove that the accused device contains each limitation of the asserted claim(s)." Bayer AG v. Elan Pharm. Research Corp., 212 F.3d 1241, 1247 (Fed. Cir. 2000). "If any claim limitation is absent from the accused device, there is no literal infringement as a matter of law." Id.

II. MOTIONS TO EXCLUDE

Federal Rule of Evidence ("FRE") 702 governs testimony by expert witnesses. Under the Rule, a witness is an expert when "qualified . . . by knowledge, skill, experience, training, or education." FED. R. EVID. 702. An expert may testify to expert opinions if "(a) the expert's scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert has reliably applied the principles and methods to the facts of the case." Id. Unlike lay witnesses, expert witnesses may base their opinions on underlying facts and data that would otherwise be inadmissible on their own "[i]f experts in the particular field would reasonably rely on those kinds of facts or data in forming an opinion on the subject." FED. R. EVID. 703.

In Daubert v. Merrell Dow Pharm., Inc., the Supreme Court recognized "a gatekeeping role for the judge" in assessing expert testimony. 509 U.S. 579, 597 (1993). The FRE "assign to the trial judge the task of ensuring that an expert's testimony both rests on a reliable foundation and is relevant to the task at hand." Id. This judicial assessment into the reliability of expert testimony, the Court explained, hinges on "whether the reasoning or methodology underlying the testimony is scientifically valid and of whether that reasoning or methodology properly can be applied to the facts in issue." Id. at 592-93. Answering that question, in turn, is a "flexible" inquiry, id. at 594, that is guided by "[m]any factors" including "whether a theory or technique . . . can be (and has been) tested," "whether the theory or technique has been subjected to peer review and publication," "the known or potential rate of error," and the level of acceptance within the relevant expert community. See id. at 593-94.

DISCUSSION

I recognize that my resolution of some issues below, if final, would render others moot. I address all of the parties' disputes for the sake of streamlining future proceedings in case any of my determinations are set aside on appeal.

I. SUMMARY JUDGMENT ON INFRINGEMENT AND INVALIDITY

Intel moves for summary judgment on three grounds. First, it argues that the Accused Products do not satisfy the "diffusion region" requirement of the Asserted Patents. Second, it argues that the '334, '335, and '523 Patents are not infringed because of their "gate contact structure" requirement. Third, it argues that the '334 and '335 patents are invalid for their failure to satisfy the statutory written description and enablement requirements. I agree with Intel on all grounds.

As explained above, literal infringement ultimately "requires the patentee to prove that the accused device contains each limitation of the asserted claim(s)." Bayer, 212 F.3d at 1247. If a claim limitation is "absent from the accused device, there is no literal infringement as a matter of law." Id.¹ The two terms at issue here have already been construed, so all that remains is to determine whether the claims read on the Accused Products.

1 The doctrine of equivalents is not at issue. Tela previously attempted to reserve the ability to assert a doctrine of equivalents claim; I held in June 2019 that "[i]f Tela wants to amend to add contentions under this theory, it must seek leave." Dkt. No. 162 at 19. It did not do so until it sought to amend its infringement contentions in February 2020. Dkt. No. 191. I denied it leave, explaining that my previous order had given Tela the opportunity to amend but, by then, "it was simply too late." Dkt. No. 199 at 1-2.

A. Noninfringement Based on Diffusion Regions

As the parties agree, every asserted claim requires a "diffusion region." See Intel's Motion for Summary Judgment ("Intel SJ Mot.") [Dkt. No. 218] 5; Tela's Opposition to Intel SJ Mot. ("Intel SJ Oppo.") [Dkt. No. 231] 2. The claims either expressly require a "diffusion region" or they require "gate electrodes"; under my claim construction, a "gate electrode" is "a portion of the conductive shape that extends over the diffusion region and is used to control the flow of electrical current between the source and drain regions of a transistor." Claim Construction Order 14. The parties' dispute is over whether Intel's products have "diffusion regions" as that term has been construed. I conclude that Intel is entitled to summary judgment of noninfringement because the Accused Products do not have "diffusion regions" under the agreed construction.

The parties agreed on the construction of "diffusion region" as "selected portions of the substrate within which impurities have been introduced to form the source or drain of a transistor." See id. 5. There is no material dispute of fact about several features of the Accused Products. The parties and their experts agree about the nature of the Accused Products' substrate. In general, a semiconductor's (usually silicon) substrate is the base layer of material. See, e.g., Intel SJ Mot. Ex. 4 (Tela's ITC demonstratives) at ECF 5; '966 Patent at Fig. 2; see also Claim Construction Order 4. The substrate of the Accused Products is likewise a base layer of silicon. See, e.g., Intel SJ Oppo. 8 (collecting citations); Intel SJ Mot. 6 (collecting citations); Intel SJ Oppo. Ex. C at ¶ 155. The Accused Products also have [Redacted] formed from this silicon substrate. See, e.g., Intel SJ Oppo. 8; Intel SJ Mot. 6. Despite Tela's attempts to show otherwise, the parties and their experts agree—as I describe below—about the materially relevant aspects of the sources and drains of the Accused Products.

Summary judgment is warranted because the regions that contain impurities that form sources and drains on the Accused Products are not, as they must be, selected portions "of the substrate." They are, instead, a different material entirely. Both parties' experts agree that, in the Accused Products, a part of the silicon substrate [Redacted] See, e.g., Intel SJ Mot. Ex. 5 ¶ 155 (Tela's expert report describing how source and drain regions are "formed by [Redacted] "); Ex. 6 ¶ 251 (Intel's expert report illustrating the process of [Redacted] a portion of the substrate). Then, the impurities are introduced via a process called "epitaxial deposition" or "epitaxy." In this process, a [Redacted] . See id. Ex. 5 ¶ 155; Declaration of Christopher Auth ("Auth Decl.") [Dkt. No. 217-5] ¶¶ 22-24. It is this [Redacted] that contains impurities (such as [Redacted] ), not the silicon substrate. Taking the plain language of the agreed construction, the epitaxially deposited material is not a "portion[] of the substrate." To hold otherwise would mean that materials become "portions of the substrate" merely because they are [Redacted] .

This conclusion is reinforced because the Accused Products' impurities are not introduced "within" the substrate, as the agreed construction requires. In the diffusion regions described in the Asserted Patents, the impurities exist "within" the substrate itself. See, e.g., '966 Patent at 12. This was done to "modify[] the electrical properties of the substrate." Id. (emphasis added). In the Accused Products, however, the impurities are in a physically distinct material that has, in effect, [Redacted] . Accordingly, the Accused Products do not literally infringe the asserted claims.

Tela makes a number of arguments in response, but none demonstrate a genuine dispute of material fact. Tela relies heavily on internal Intel documents that show that Intel sometimes referred to the epitaxial deposition process as "diffusion." See Intel SJ Oppo. 3-9. The relevant question in this literal infringement analysis, however, is whether the Accused Products have "diffusion regions" as the asserted claims use that term. Pitney Bowes, 182 F.3d at 1304. The patents' use of that term was conclusively established by the parties' agreed construction. Intel's internal nomenclature cannot change either the nature of its epitaxial deposition process or the agreed construction. Cf. Plastic Omnium Advanced Innovation & Research v. Donghee Am., Inc., 943 F.3d 929, 936 (Fed. Cir. 2019) ("[T]he patentee's definition . . . as construed by the court—not Donghee's product literature—controls whether the accused product falls within the scope of the claim."). Intel has produced evidence, moreover, that its internal use of "diffusion" to describe the process is a historical holdover. See Auth Dec. ¶ 28. Some older Intel products (like products across the field) possessed "diffusion regions" and so the areas in which impurities are introduced into the Accused Products are still sometimes referred to as "diffusion regions" and the process as "diffusion." Id. Intel's internal terminology does not create a genuine dispute of material fact.

Tela also argues, as it must, that the agreed construction of "diffusion region" does apply to Intel's epitaxially deposited impurities. See Intel SJ Oppo. 7-9. Tela points out that, in an epitaxial deposition process, impurities are (in a sense) introduced into portions of the substrate. Id. 8-9. In essence, it argues that "[b]ecause the impurities are introduced into [Redacted] , they are introduced within the substrate." Id. 8. Tela's proposed interpretation, however, ignores the two features of the agreed construction identified above. Most importantly, the diffusion regions must themselves be, as explained, "selected portions of the substrate" and the epitaxially deposited areas are not. Additionally, diffusion regions are substrate "within which" impurities have been introduced. When the epitaxial material is placed into [Redacted] , it is not "within" the substrate, it is where substrate [Redacted] . Again, Tela's argument would mean that anything becomes a "portion of the substrate" merely because it is [Redacted] , despite being a different material altogether.²

2 Perhaps tellingly, Tela's argument often subtly shifts the language from "within"—as the agreed construction uses—to "into," which has connotations more in line with its argument. See, e.g., Intel SJ Oppo. 8.

Tela next contends that "diffusion region" would be "understood by POSITAs . . . to refer to the area where sources and drains are formed using various different processes." Intel SJ Oppo. 3, 6-7. In other words, Tela argues that the areas containing sources and drains are understood by those skilled in the art to be diffusion regions even if they are created via epitaxial deposition. Even if this were true, this understanding is not relevant to the extent it departs from the claim construction. I agree with Tela that some POSITAs might well refer to Intel's epitaxially grown sources and drains as "diffusion regions" (many, perhaps, in the same holdover way Intel does). But that does not mean that those sources and drains are "diffusion regions" under the asserted claims as construed by the parties.

Tela also argues that the agreed construction "does not require a specific 'diffusing' process" and that finding for Intel would effectively impose one. Intel SJ Oppo. 5-6 (emphasis added). While I agree that no particular process is necessarily contemplated by the claims, the agreed construction does require that the diffusion region be a "portion[] of the substrate"; the only reason that Intel's process is relevant is that the result is the purported "diffusion region" is not a "portion[] of the substrate," it is a different material entirely.

Finally, Tela attempts to reframe this as a factual debate—and therefore one that is inappropriate to resolve at summary judgment. See Intel SJ Oppo. 7-9. The relevant facts, however, are not in dispute. There is no factual dispute about the relevant aspects of Intel's epitaxially deposited sources and drains. The parties agreed on a construction of "diffusion regions." Tela's arguments are better geared toward claim construction, which already occurred. The Accused Products do not possess "diffusion regions" as the asserted claims use that term.³

3 Tela further contends that a feature of the process called [Redacted] aside from the epitaxially grown impurities. Intel SJ Oppo. 6. For the reasons explained below, I giant Intel's motion to strike the portion of Tela's expert report that forms the basis of this argument because it was never disclosed as an infringement theory. Additionally, Tela signaled at the hearing on these motions that it was no longer pursuing the theory.

B. Noninfringement Based on Gate Contact Structures

Intel also argues that the 14 nm and 10 nm Accused Products do not infringe the '334, '335, and '523 patents' asserted claims.⁴ Intel SJ Mot. 9. These claims require "gate contact structure(s)" or "contact structure(s)." See Claim Construction Order 16. The parties disputed the proper construction and I construed both terms to mean "conductive structure(s) in a gate contact layer above and separate from the gate layer and below and separate from interconnect layers." Id. For present purposes, these structures must therefore possess three salient features: they must be (1) "separate from" the gate layer and interconnect layers, (2) "above" the gate layer, and (3) "below" interconnect layers. I use the denotation "[gate] contact structure(s)" when referring to the claimed structures in all three patents.

4 Tela did not accuse the 22 nm products of infringing these claims.

i. 14 nm Products

Intel argues that its 14 nm products do not have "[gate] contact structures" as that term has been construed. It asserts that the 14 nm products instead have [Redacted] '" Intel SJ Mot. 11. Accordingly, Intel asserts that [Redacted] ) is not "below and separate from interconnect layers" because [Redacted] . For instance, Intel points to the Declaration of Christopher Auth, Ph.D., Intel's vice president who also worked for many years on the technical aspects of Intel's semiconductors. Auth Decl. ¶¶ 1-4. Auth states that, rather than using a " [Redacted] " as in older technology, the [Redacted] " Id. ¶ 33-34. Intel also points to the testimony of Dr. Terrence Hook, Tela's expert in the ITC proceeding, that the [Redacted] Intel SJ Mot. Ex. 11 at 1153:1-13, not that it is "separate from" the interconnect layers as the claim construction requires.⁵

5 At the hearing on these motions, the parties disputed whether Hook's statements are adoptive admissions by Tela. None of my present decisions depend on this issue. I look to Hook's testimony only because it is evidence the parties rely on.

Tela primarily responds by pointing to Intel's layout files, design abstractions, and design rules that contemplate the 14 nm products as having a gate contact layer that is separate and below the interconnect. See Intel SJ Oppo. 9-13. For the reasons explained below, I giant Intel's motion to strike this evidence. The evidence does not create a genuine dispute of material fact about the 14 nm products' physical structure.

Tela does not point to any evidence that contradicts the reality that the [Redacted] itself—that is, the actual physical structure— [Redacted] (and, as noted, Tela's expert in the ITC proceeding agreed that it does). Tela does passingly point to the deposition of Dr. Vivek Subramanian, Intel's technical expert. Intel SJ Oppo. 12. But the portion it references begins with Subramanian reinforcing that " [Redacted] " Id. Ex. L at 159:24-160:2. The distinction Tela is apparently relying on Subramanian for is that the " [Redacted] ," that will ultimately be [Redacted] ; but that does not contradict the testimony [Redacted] . Id. at 160:3-6.

Tela also argues that diagrams of the 14 nm products supporting its position are consistent with those I relied on in reaching my construction of "gate contact layer." Intel SJ Oppo. 10-11. But in that claim construction, I necessarily relied on the contents of the patents themselves. Now, the task is to compare the construction to the Accused Products. Tela has, as explained, showed no genuine dispute of material fact about the 14 nm products' physical features.⁶

6 Because I conclude that the 14 nm products do not infringe on this ground, I do not address whether the [Redacted] is "above and separate from the gate layer."

ii. 10 nm Products

Intel also argues that its 10 nm products do not have the required "[gate] contact structures" because they have a [Redacted] . See Intel SJ Mot. 14-15; Auth Decl. ¶¶ 36-37. This "unified structure" occurs because [Redacted] . Auth Decl. ¶¶ 36-27; see also Intel SJ Mot. Ex. 10 at 268:22-269:16 (Tela's expert agreeing that [Redacted] ). In other words, the record shows that the accused gate contacts are not "separate from" the interconnect layer because they are a [Redacted] .

Tela points to no record evidence that casts doubt on the finding that the physical structure is [Redacted] —and therefore not multiple structures that are separate. Instead, it tries to fashion a distinction between different parts of this [Redacted] structure; none of its arguments show that the gate contact structure is "separate from" the interconnect. Because it is difficult to discuss Tela's arguments without naming different parts of the [Redacted] differently, I adopt the label of [Redacted] to describe the portion of the structure that contacts the gate—that is, the portion that Tela accuses of meeting the claim limitation. See Intel SJ Oppo. 13-15.

First, Tela and Dr. Daniel Foty, its expert, rely on Intel's layout and upstream design files. Id. 14. As explained, that evidence has been struck and does not create a dispute of material fact.

Second, Foty attempts to show that there are actually multiple structures in existence. Foty, as alluded to above, agrees that the process is performed in a [Redacted] . Id.Ex. C ¶ 171. He states, however, that the [Redacted] Id. This opinion stems not from any argument that the [Redacted] but instead from his assertion that [Redacted] Id. ¶ 475. It is as a result of this, he states, that "two separate structures are formed." Id. Even in the best light for Tela, this does not create a dispute of material fact about the structures. At most it shows, as it plainly says, that [Redacted] is formed. Id. ¶ 475. In other words, different areas into which [Redacted] . See id. His opinion is, therefore, not aimed at a factual dispute but at a definitional one: he has adopted an unusual definition of "separate" that does not actually require structures to be separate. Foty does not dispute that the [Redacted] are a [Redacted] in the sense that they are made from [Redacted] with no identifiable physical separation; accordingly, he does not present any fact showing they are separate structures.

Tela's arguments on both the 14 nm and 10 nm products are, once again, better aimed at claim construction. During claim construction Intel argued that the [gate] contact structures were required to be separate from the gate and interconnect layers. I agreed. See Claim Construction Order 16. Tela's proposed construction had no separateness requirement. Id. Now, Intel has shown sufficient evidence that the accused [gate] contact structures are not, physically, separate as the claim construction requires. Tela has not shown a genuine dispute of material fact on this point. Intel is entitled to summary judgment because neither the 14 nm nor the 10 nm products have "[gate] contact structures" as that term has been construed.

C. Validity of the '334 and '335 Patents

Intel argues that the '334 and '335 patents are invalid because they fail to satisfy the statutory written description and enablement requirements. 35 U.S.C. § 112(a) provides that a patent's specification "shall contain a written description of the invention, and of the maimer and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art . . . to make and use the same." Accordingly, as relevant here, the specification must perform two functions: enablement and description.

i. Enablement

The requirement of enablement stems from Section 112's provision that the specification "enable any person skilled in the art . . . to make and use" the invention. 35 U.S.C. § 112(a). "Enablement is a question of law based on underlying factual findings." MagSil Corp. v. Hitachi Glob. Storage Techs., Inc., 687 F.3d 1377, 1380 (Fed. Cir. 2012). A party must prove invalidity based on failure to enable by clear and convincing evidence. Microsoft Corp. v. I4I Ltd. Partnership, 564 U.S. 91, 95 (2011). "To be enabling, the specification of a patent must teach those skilled in the art how to make and use the full scope of the claimed invention without 'undue experimentation.'" MagSil, 687 F.3d at 1380 (internal quotation marks omitted). Enablement is an "important doctrine [that] prevents both inadequate disclosure of an invention and overbroad claiming that might otherwise attempt to cover more than was actually invented." Id.

These general principles apply to patents that contemplate a range of technologies. The patent itself need not teach a person skilled in the art to practice the full scope of the claims if that person is taught to do so without undue experimentation. Wyeth & Cordis Corp. v. Abbott Labs., 720 F.3d 1380 (Fed. Cir. 2013). A range, like the different nanometer sizes here, is not improper merely because it "open-ended"; instead, the Federal Circuit has held that if those skilled in the art would recognize an inherent limit on the range, the patent need only enable them to practice to that limit. See, e.g., Andersen Corp. v. Fiber Composites, LLC, 474 F.3d 1361, 1376-77 (Fed. Cir. 2007). And, because enablement is directed to those skilled in the art, "the artisan's knowledge of the prior art and routine experimentation can often fill gaps, interpolate between embodiments, and perhaps even extrapolate beyond the disclosed embodiments, depending upon the predictability of the art." Chiron Corp. v. Genentech, Inc., 363 F.3d 1247, 1253 (Fed. Cir. 2004). Accordingly, the patent need not teach what is already known in the art. Id. But "[w]hen you claim a [range], you can't simply disavow the invalid portion and keep the valid portion of the claim." Alcon Research, Ltd. v. Apotex Inc., 687 F.3d 1362, 1368 (Fed. Cir. 2012).

Intel argues that the '334 and '335 patents are invalid because they would not enable a person of skill in the art to practice the full range of technologies without undue experimentation. The claims of the '334 and '335 patents require a "gate . . . width of less than or equal to about 45 nanometers." See '334 Patent at 30, 35, 36; '335 Patent at 30, 36, 37. They also require a "gate pitch of less than or equal to about 193 nanometers." See '334 Patent at 30, 35, 36; '335 Patent at 30, 36, 37. Gate widths are the "short" dimension of the gate features and gate pitches are the center-to-center distance between gates. Intel contends that the patents do not describe or enable these ranges.

In its Motion, Intel initially argued that the '288 Provisional itself (which, as noted, the '334 and '335 Patents claim priority to) showed that the technology described was not written or enabled. The '288 Provisional, Intel pointed out, "depict[ed] the smallest feature size available in 2006 as being approximately 65 nm, and predict[ed] that 45 nm feature sizes would be reached around the year 2012." Intel SJ Mot. 16. It also stated that "[t]oday minimum feature sizes are approaching 45 nm." Id. (emphasis added by Intel). Intel argued that, as a result, it was clear that the '334 and '335 Patents' claims to a gate "width of less than or equal to about 45 nanometers" failed to describe or enable.

In its Opposition, Tela made three points that have caused Intel to narrow the argument in its Reply. The first is legal. Tela pointed out that the patent need not itself enable the entire feasible range if it sufficiently taught those skilled in the art to do so without undue experimentation; as discussed, that is correct. The second point is factual. Tela cited evidence that people had, at the time of the '288 Provisional, "fabricated devices with gate line widths as small as 10nm." Intel SJ Oppo. 24 (citing Ex. CC ¶ 421). In fact, a 2002 paper—cited in a presentation given by Intel's expert, Subramanian, in 2012—announced fabrication of a 10 nm gate width. See id. Ex. CC ¶ 407, Ex. W at ECF 30. The third point is both legal and factual: Tela contends that there is a theoretical lower limit of 7 nm for gate widths and that the enablement only needs to be to this limit. See Intel SJ Oppo. 22 (citing Ex J at 111). In its Reply, Intel appears to treat (perhaps for the sake of argument on summary judgment) 7 nm as the lower bound for gate widths. See Intel SJ Reply 11.

In response to these points, Intel narrowed its focus; it argues that the '334 and '335 patents do not enable "single-nanometer gate widths." See Intel SJ Reply 11. Because Tela treats 7 nm as the theoretical lower limit and there is evidence that 10 nm fabrication was known in the art at the time of patents, Intel's argument is that the '334 and '335 Patents do not enable a person skilled in the art, without undue experimentation, to practice gate widths between 10 nm and 7 nm. See id.

I agree with Intel that there is no genuine dispute of material fact that the patents are, by clear and convincing evidence, invalid for failure to enable. They do not, on their face, teach gate widths below 10 nm—indeed, they do not even purport to teach gate widths below 45 nm, which they classify as being a future development. The question therefore becomes whether, based on the knowledge from the patents and the art, the work necessary to practice gate widths smaller than 10 nm would require "undue experimentation."

Intel has presented evidence that the experimentation required to reach smaller and smaller sizes is, in fact, extremely demanding, rigorous, time-consuming, and resource-intensive. Indeed, there is no evidence on this record that single nanometer gate widths have been achieved today, fifteen years after the application. Hook, for instance, agreed with Intel's characterization of reaching each progressively smaller size as requiring "considerable innovation, considerable invention, thousands and thousands of engineers, and billions and billions of dollars in R&D." Intel SJ Mot. Ex. 11 at 1105:5-1106:13. Because of the way the technology develops, it has historically moved to progressively smaller sizes—so this effort is all geared at each shrinkage, not a jump to 7 nm. Auth attests that he has spent 19 years scaling down these sizes—and has even now not reached Tela's theoretical lower bound of 7 nm gate widths. Auth Decl. ¶ 14. Moreover, it is not disputed that Intel itself employed over 2,000 engineers—95% percent of whom have PhDs—to achieve its 10 nm products. Id. ¶15.

While I agree with Tela—especially when viewing the evidence in the best light to it—that Intel's efforts by themselves may not be adequately illustrative because it is engaged in mass-manufacturing and commercialization, the evidence as a whole paints a clear picture that the "experimentation" required to reach each progressively smaller size is "undue." Tela itself admits that reaching each progressively smaller "process node" (e.g., moving from a 45 nm to 30 nm gate width) has historically taken two years—two years of the industry-wide experimentation described above. See Intel SJ Oppo. 17. Cf. White Consol. Indus., Inc. v. Vega Servo-Control, Inc., 713 F.2d 788, 791 (Fed. Cir. 1983) (classifying one-and-a-half to two years of experimentation as "clearly unreasonable").

Tela makes several arguments that (1) the patents sufficiently enable the full range and (2) any required experimentation is not "undue." Because those arguments are often coextensive, I address them together. First, Tela and Foty argue that "Moore's Law" and CMOS scaling show that a person skilled in the art would be able to practice widths smaller than 10 nm without undue experimentation. As relevant here, Moore's Law is the historical tendency—and the associated projection of the future that is born from it—for the area of new semiconductors to shrink by approximately 50% every two years as innovation occurs. See, e.g., Intel SJ Oppo. Ex. CC ¶¶ 381-81. CMOS scaling is the shrinking of chips. It has, historically, roughly aligned with Moore's Law. See, e.g., id. ¶¶ 390-91. According to Tela, the historical shrinking of chips and that CMOS scaling is known in the art shows that the patents sufficiently teach a POSITA how to fabricate chips with gate widths down to 7 nm. I disagree.

That argument is silly. "Moore's Law" is, as Foty himself puts it, not a "law of physics." Intel SJ Oppo. Ex. CC ¶ 381. It is, instead, an "observation." Id. Tela's expert in the ITC proceeding, Hook, agreed with the characterization of Moore's Law as the "observation of reducing circuit chip area by 50 percent." Intel SJ Mot. Ex. 11 at 1105:5-8 (emphasis added). He also agreed with the statements that Moore's Law "requires a lot more than just scaling by .7," the CMOS scaling that Tela tries to present as a foregone conclusion. Id. at 1164: 6-19. It "requires constant innovation, constant invention, and constant R&D work." Id. Tela has pointed to no record evidence that Moore's Law is more than a historical tendency that leads to extrapolated future projections. It has also not indicated any case in which a court has held that a patent can claim smaller sizes down to the theoretical limit purely on the basis of Moore's Law or similar historical trend. I note—though my finding is not based on this—that such a holding would have dramatic consequences for innovation given the prevalence of the technology in modern society and the expansive ramifications of using Moore's Law to extend the reach of patents. In the best light for Tela, these principles show that it is likely the industry will, in the proper timeframe, reach the relevant sizes. But that is precisely because there is, the record shows, great experimentation to reach this goal.

Relatedly, Tela argues that CMOS scaling is so well known in the art that a POSITA looking at its patents would understand how to scale what they teach down to sufficiently small sizes. That argument is incorrect for largely the same reasons just described: to achieve this scaling (as Tela's ITC expert admitted), the undue experimentation described above has always been required. In support of its scaling argument, Tela also at points contends that the '334 and '335 Patents disclose only a "layout methodology" or "layout technique" and, to the extent the patents do not teach down to certain sizes, they are merely not teaching "embodiments" of that technique or methodology. The Patents, however, are for "semiconductor chip[s]" as both the terminology and the substance of their claims make clear. In any event, Tela's argument is unpersuasive because the fundamental question is whether the patents would teach the technology without undue experimentation: Regardless of whether it is framed as patents for the chip itself or merely a layout methodology for it, they do not—as described above—teach a POSITA how to achieve gate widths smaller than 10 nm without undue experimentation. If the patents are framed as teaching layout techniques, the same experimentation described above is required to practice them at extremely small sizes.

Aside from this, Tela argues that its expert testimony shows that any experimentation would not be "undue." Its broad argument, again, is that because scaling is known in the art, the patents teach a scalable layout technique that POSITAs could simply scale by shrinking components. It points to a statement from Hook regarding achieving the scaling predicted by Moore's Law:

In my opinion, the elements that you just enumerated, while challenging and interesting, are, in fact, business as usual for the industry. And they are—if routine means what one normally does. I'm not consulting a dictionary, but it's exactly what would be anticipated. I consider all of the things that you said, despite the large numbers, billions [of dollars] and thousands [of people], that's routine. That's exactly what everybody expects to do and does every day and every year. It's routine inasmuch as—routine here doesn't mean simple. It means as per usual.

Intel SJ Oppo. Ex. 11 at 1124:14-24. This statement does not show that the patents teach how to fabricate gate widths smaller than 10 nm without undue experimentation. Hook's statement is not—nor could it be—a legal conclusion; instead, in the best light to Tela, he was testifying that the industry "expects" Moore's Law to be adhered to if the industry does what has become "routine" in the industry—namely, the enormous amount of work and experimentation that Intel identifies. And while he may view this striking level of experimentation as "routine" in the industry, it is still undue experimentation.

Tela also relies on Foty's opinions, which largely echo Tela's Moore's Law argument. Foty adds the additional gloss that, aside from arguments over Moore's Law itself, CMOS scaling is so well-recognized in the industry that a POSITA in 2006 could, from Tela's patents, be taught the technology down to the 7 nm gate width size. Intel SJ Oppo. Ex. CC at ¶¶ 422-34. There are several problems with this. Despite being more detailed than the Hook statement described above, Foty's opinion suffers the same fundamental flaw: The record shows that the innovation and experimentation Foty classifies as quotidian and inevitable is, as I have explained, undue experimentation for enablement purposes. Foty does not dispute the nature or quantity of the work that goes into producing scaling; he simply characterizes those facts as being only routine. But, as explained, enablement is a question of law. Tela points to no fact that Foty relies on that would transform the experimentation described above into "routine."

Moreover, Foty's report is to some extent undermined by the facts described in Hook's testimony. Hook testified that he agreed with the statement that "since 2006, this priority date, there have been 15 years of extensive and exceedingly complicated innovation and invention that has gone on." Intel SJ Mot. Ex. 11 at 1285:1-5. He agreed with the statement that "even if we think about the most cutting edge products today . . . the technology just isn't there to get it down to the ranges that you say are the limitations for purposes of the Tela claims." Id. at 1283-84:21-3. Tela is correct that Hook's ultimate conclusion is that such experimentation is "expected" and "routine" in the industry, but a conclusory characterization does not govern the analysis and it is no reason to disregard the evidence of what actually occurs in the field.

Aside from these arguments about how, broadly speaking, semiconductor technology is advanced, Tela has several specific arguments related to enablement of small gate widths. As I explain, none of them create a genuine dispute of material fact about whether sizes below 10 nm were enabled by the '334 and '335 Patents. Primarily, Tela contends it "will present evidence that POSITAs had actually made semiconductors with gate widths in substantially the entire range from 45nm to 7nm as of the time of the inventions of the '334 and '335 Patents." Intel SJ Oppo. 22. It cites six paragraphs of Foty's report in support. And it argues that "Dr. Foty will further explain that the actual construction of devices with these gate widths likewise shows that POSITAs could practice substantially the entire pitch range at the time." Id. It cites five of those six paragraphs in support. Tela's argument itself, even aside from its supporting evidence, reveals a key infirmity by arguing that "substantially" the entire range of width and pitch was enabled. As explained, the smallest size of gate width Tela has identified in the record at the relevant time was 10 nm. See Intel SJ Oppo. 20. This aside, the evidence Tela cites does not create a genuine dispute of material fact. Those six paragraphs of Foty's report do not show that the entire range down to 7 nm was enabled; instead, the smallest size Tela points to is, again, 10 nm. See id. Ex. CC ¶¶ 378, 407-11. The bulk of these paragraphs, in fact, repeats the earlier arguments about scaling and commercial production.

These single-digit sizes are not trivial or peripheral. Achieving smaller and smaller sizes, the record shows, is key in the field. This is not a case in which, for instance, the range of the technology is, as a practical matter, entirely enabled even if there is not proof of every last immaterial embodiment. See, e.g., Scripps Clinic & Research Found. v. Genentech, Inc., 927 F.2d 1565, 1572 (Fed. Cir. 1991), clarified on denial of reconsideration, No. 89-1541, 1991 WL 523489 (Fed. Cir. Apr. 30, 1991), overruled on other grounds by Abbott Labs. v. Sandoz, Inc., 566 F.3d 1282 (Fed. Cir. 2009). Instead, scaling gate widths down to single nanometer sizes—in addition to requiring an intense amount of experimentation—would be a valuable development. As both parties make clear, size is key in the integrated circuit industry and reaching down toward a lower limit is still the subject of costly, time-consuming innovation.

Tela also argues that this argument depends on the "hotly contested issue of how a POSITA would understand the disclosures," in particular, "whether those documents disclose what was possible generally at the time of the inventions (Tela's position) or what was commercially available (Intel's position)." Intel SJ Oppo. 19. Parenthetically, that statement does not reflect the position Intel advances. In any event, Tela has shown no evidence that the patents disclose single-nanometer sizes in any way, including "generally." Relatedly, Tela argues that Intel shows, at most, what was commercially available at the time of the Provisional, not what could be done generally. Id. 20. But, as explained, there is no record evidence that single nanometer gate widths could be achieved without undue experimentation at all.

There is clear and convincing evidence that the '334 and '335 Patents would not enable a person of skill in the art to practice the full claimed range down to Tela's proffered theoretical limit without undue experimentation.

ii. Written Description

"The essence of the written description requirement is that a patent applicant, as part of the bargain with the public, must describe his or her invention so that the public will know what it is and that he or she has truly made the claimed invention." AbbVie Deutschland GmbH & Co., KG v. Janssen Biotech, Inc., 759 F.3d 1285, 1298 (Fed. Cir. 2014). The inventor must, through the written description, "convey with reasonable clarity to those skilled in the art that, as of the filing date sought, he or she was in possession of the invention." Nuvo Pharm. (Ireland) Designated Activity Co. v. Dr. Reddy's Labs. Inc., 923 F.3d 1368, 1376 (Fed. Cir. 2019) (internal alteration omitted). The requirement accordingly "plays a vital role in curtailing claims that have not been invented, and thus cannot be described." AbbVie, 759 F.3d at 1299 (internal quotation marks and alteration omitted). It helps give substance to the Supreme Court's admonition that "a patent is not a hunting license. It is not a reward for the search, but compensation for its successful conclusion." Brenner v. Manson, 383 U.S. 519, 536 (1966).

As in many cases, the written-description and enablement analyses are, on the facts here, closely related. For the reasons explained above, a POSITA would not, from the '334 and '335 Patents, be aware that the named inventor "was in possession of the invention": the record is clear that there is no evidence that single-digit nanometer gate widths were possible to create at the time, and the patents describe those small sizes—and indeed much larger sizes—entirely in futuristic, aspirational terms. And, for the reasons described above, neither the knowledge in the art nor the possible (undue) experimentation changes this conclusion. At bottom, the '334 and '335 Patents describe technology that had not yet been invented; they express a "mere wish or plan for obtaining the claimed invention[, which] is not adequate written description." Centocor Ortho Biotech, Inc. v. Abbott Labs., 636 F.3d 1341, 1348 (Fed. Cir. 2011).

D. Consistency of Invalidity and Infringement Arguments

Tela argues that Intel's positions regarding, on the one hand, the invalidity of the '334 and '335 patents and, on the other hand, noninfringement of the 14 nm products are inconsistent with each other. See Intel SJ Oppo. 15-16. This argument stems from a conclusion of Intel's expert, Subramanian, in his initial expert report on invalidity. See Intel SJ Oppo. Ex. BB ¶¶ 790-91. In that report, Subramanian states that he "disagree[s] that Intel's Accused Products infringe for the reasons that I will explain in my rebuttal report," but that if the 14 nm products were found to infringe "under Tela's infringement theory," they would anticipate the '334 and '335 Patents. Id. ¶ 791. I disagree with Tela that Intel has taken inconsistent positions that are "fatal" to some or all of its arguments. Intel SJ Oppo. 15. Subramanian's report is explicit that he believes the Accused Products do not infringe the '334 and '335 Patents. He merely states that, if Tela's theory of infringement prevailed, his view is that the 14 nm products would therefore anticipate those patents. Id. Ex. BB ¶ 790-91. That position is consistent and permissible. See Peters v. Active Mfg. Co., 129 U.S. 530, 537 (1889) ("That which infringes, if later, would anticipate, if earlier.").

E. Infringement and Invalidity Conclusion

For the reasons described, Intel is entitled to summary judgment that the Accused Products do not infringe the asserted claims and that the '334 and '335 Patents are invalid.

II. SUMMARY JUDGMENT ON INVENTORSHIP AND THE CNTS

Some of Intel's claims and defenses are based on its allegation that Pileggi should rightfully be named an inventor of the Asserted Patents. Additionally, some of Intel's claims depend on its allegation that Tela violated the parties' CNTS—that they entered into around the time of Intel's investment into Tela—by pursuing its claims regarding the Asserted Patents here. Tela moves for summary judgment on those claims and defenses.

A. Laches

I first consider Tela's argument that Intel's inventorship-based claims and defenses are barred by laches. "Laches is an equitable defense that may bar an inventorship claim." Lismont v. Alexander Binzel Corp., 813 F.3d 998, 1002 (Fed. Cir. 2016). To prevail on a laches claim, Tela must establish that (1) the delay in filing a suit was "unreasonable and inexcusable" and (2) it suffered "material prejudice" as a result of the delay. Id. In the inventorship context, "a rebuttable presumption of laches attaches whenever more than six years passes from the time a purportedly omitted inventor knew or should have known of the issuance of the relevant patent." Id. Like all matters to be settled on summary judgment, any underlying material facts must not be in genuine dispute. See id.

i. Presumption of Laches

Tela points to three pieces of evidence that Pileggi knew or should have known of the issuance of the patents more than six years prior to this suit. None establishes its entitlement to the presumption of laches as a matter of law. First, Tela points to a phone call in January 2006 between Pileggi and Becker, and an email Pileggi wrote to himself afterward. Tela's Motion for Summary Judgment ("Tela SJ Mot.") [Dkt. No. 221] 14. In that email, Pileggi wrote,

Told him that I had heard that he [Mr. Becker] was starting a company, and that the technology was similar to the CMU technology . . . . I reminded him that our meeting was covered by our NDA, and that our IP belonged to CMU and MARCO, but was licensed to Fabbrix through CMU.

See Tela SJ Mot. Ex. 20 ¶ 127; Intel's Opposition to Tela's Mot. ("Tela SJ Oppo.") [Dkt. No. 229] Ex. 36.

Second, Tela points to a letter that a Fabbrix attorney sent to Becker on July 24, 2006. See Tela SJ Mot. Ex. 23. Tela characterizes that letter as "impl[ying] that Tela misused Fabbrix's confidential and/or proprietary information." Tela SJ Mot. 15. The letter states, in relevant part,

As you are also aware, various individuals have been exposed to Intellectual Property of Fabbrix, as well as certain of its confidential business information. This includes yourself . . . and John Malecki.

While verbal assurances have been made by you to Fabbrix that the business of Telas [sic] is different than and not in competition with the business of Fabbrix, such assurances are viewed with skepticism, since "Telas" in Spanish means "fabrics," and information that

Fabbrix has received from third party sources is inconsistent with those verbal assurances. Tela SJ Mot. Ex. 23 at ECF 2. The letter "reminded" Becker of Tela's confidentiality obligations and alerted it to a patent application that Fabbrix had published. Id. at ECF 2-3.

Third, Tela points to a protest that PDF—which bought Fabbrix—filed under 37 C.F.R. § 1.291 in March 2007. See Tela SJ Mot. Ex. 28. That protest was filed with the U.S. Patent and Trademark Office ("PTO") and sent in an email to Fabbrix. See id. It related to the '402 Application, the parent to the Asserted Patents. As noted, Pileggi worked at PDF in some capacity for approximately four years after the sale. See id. Ex. 30 at 263:22-264:22.

None of these pieces of evidence—nor the combination of them—show that Tela is entitled to summary judgment on this issue. At this stage, I must view all facts in the light most favorable to Intel; in that light, a reasonable person could conclude that Pileggi neither knew nor should have known of the Asserted Patents at their time of issuance. There is no direct evidence that Pileggi was aware of the Asserted Patents more than six years prior to filing this suit. He did admit in a deposition that he became aware of Tela's patent filings but could not recall when, which is insufficient to base a summary judgment determination on. Tela's Reply in Support of Tela SJ Mot. ("Tela SJ Reply") [Dkt. No. 237] Ex. A at 254:9-13.

Tela's other evidence, likewise, has been shown to be subject to material factual disputes. There is a clear dispute about Pileggi's knowledge of the PDF protest: Pileggi states that he was unaware of it. Tela SJ Oppo. Ex. 1 ¶ 31. Indeed, Tela bases its argument that Pileggi knew of it on the fact that he was associated with the company at that point. See Tela SJ Mot. 15. To show he had knowledge, Tela depends on inference; it argues that he was "deeply involved" in the company and so must have known. See Tela SJ Reply 13. At this stage, Tela is not entitled to rely on that inference, and I must draw the contrary one in favor of Intel—namely that Pileggi (as he stated) had no knowledge of the protest until this litigation. Tela SJ Oppo. Ex. 1 ¶ 31.

Tela's other evidence are the January 2006 phone call and the June 2006 email. Those statements might make a reasonable person believe that Pileggi was aware that his ideas were being utilized. There are, however, two problems. First, I cannot say on summary judgment that Pileggi was or necessarily would have been on notice that his ideas were being appropriated and, allegedly, patented. The evidence is not sufficiently conclusive at this stage. Relying on the email would require accepting an inferential chain to reach Tela's interpretation. The letter might, as Tela itself says, "impl[y]" misuse, but a reasonable jury would not automatically accept it as conclusive evidence that Pileggi was aware his inventions were in fact being misused. This aside, even if Pileggi suspected Tela were making use of some portion of his ideas, it is a leap from that belief to actual or constructive knowledge that the patents should claim him as an inventor.

ii. Entitlement to Laches

Tela also argues that laches bars Pileggi's inventorship claim, regardless of whether the presumption attaches or not. I cannot say as a matter of law that any delay in filing the suit was "unreasonable or inexcusable." On the record before me, as explained above, there is not sufficient evidence (on summary judgment) to conclude that Pileggi delayed for unjustifiable reasons. Tela's argument to the contrary, again, relies on drawing several important inferences in its favor. Specifically, Tela's argument depends on the inferences that Pileggi or Intel were aware of Pileggi's alleged contributions to the Asserted Patents for some length of time prior to filing that would make delay unjustified. It would be improper to draw that inference with respect to Pileggi, for the reasons discussed above.

It would also be improper—on this record, at summary judgment—to conclude that there was an unjustifiable delay on Intel's part. Tela offers two pieces of evidence that Intel (as distinct from Pileggi) knew or should have known of Pileggi's alleged inventorship: the PDF protest and the diligence it conducted prior to investing in Tela. Tela SJ Mot. 16-18. Tela's argument with respect to Intel is less convincing than with respect to Pileggi because Intel, at the time of the PDF protest, had no relationship to Pileggi's claims. That protest, moreover, was not that Pileggi should have been named an inventor, it concerned the patent's validity. As a result, Intel would have no reason to know of Pileggi's claims to inventorship at the time of the protest. The diligence prior to the investment also does not discuss inventorship issues, it merely states that a protest is pending and that Tela would be submitting prior art references to the PTO. Even more fundamentally, Intel would have no reason to pursue Pileggi's inventorship claim prior to Pileggi granting Intel a license to his rights in the Asserted Patents. In other words, there was no "delay" in Intel bringing an inventorship claim because it had no inventorship claim.

B. Equitable Estoppel

Relatedly, Tela argues that Intel is equitably estopped from asserting its inventorship-based claims and defenses. Like laches, equitable estoppel is an equitable defense. Equitable estoppel bars claims when there is "(1) unreasonable and inexcusable delay in filing suit, (2) prejudice to the defendant as a result of the delay, (3) affirmative conduct by the party against whom estoppel is asserted inducing the belief that it had abandoned its claim, and (4) detrimental reliance by the party asserting estoppel." MCV, Inc. v. King-Seeley Thermos Co., 870 F.2d 1568, 1571 (Fed. Cir. 1989).

For the reasons just discussed, there was no "unreasonable or inexcusable delay" in filing this suit that can be determined as a matter of summary judgment. This aside, equitable estoppel is a worse fit for this case than laches because Tela has not presented suitable evidence showing affirmative inducement or detrimental reliance. Tela's only argument that Intel took affirmative steps to induce it to abandon a claim is the purported delay in filing the suit. See Tela SJ Mot. 20. Tela points to no case in which these facts would suffice to show "affirmative conduct" aimed at "inducing" a party's reliance. There is no allegation, for instance, that Intel (or Pileggi) affirmatively represented that no suit would be filed, or even took any action that would be cause a reasonable person to believe no suit would be filed. Tela's theory would effectively collapse the "delay" portion of the equitable estoppel test with the final two prongs. Tela has pointed to no evidence that even arguably—let alone that is sufficient for summary judgment—shows that any alleged delay in filing the suit was done to induce Tela's reliance.

C. Inventorship Claims and Defenses

Tela argues that Pileggi is, as a matter of law, not an inventor of the technology in the Asserted Patents and, consequently, that Intel's claims and defenses that depend on that allegation fail.⁷ Although the underlying technology in this dispute is complex, Tela's argument on this point is narrow: It contends that Pileggi's assertion that he contributed to the technology is not, as it must be, corroborated. I agree with Intel that there are genuine disputes of material fact about the alleged corroborative evidence that preclude summary judgment.

7 Tela also moved for summary judgment on several of Intel's fraud-based claims and defenses related to the inventorship issue. See Tela SJ Mot. 21-22. Intel represents, and Tela does not dispute, that it is no longer pursuing those claims, so the motion is to that extent moot. See Tela SJ Oppo. 1 n.1.

Inventorship is a question of law based on underlying facts. Gen. Elec. Co. v. Wilkins, 750 F.3d 1324, 1329 (Fed. Cir. 2014). The issuance of a patent creates a presumption that the named inventors are the "true and only inventors." Id. Accordingly, a claim of incorrect inventorship must be proven by clear and convincing evidence. Hess v. Advanced Cardiovascular Sys., Inc., 106 F.3d 976, 980 (Fed. Cir. 1997). "A joint invention is the product of a collaboration between two or more persons working together to solve the problem addressed." Burroughs Wellcome Co. v. Barr Labs., Inc., 40 F.3d 1223, 1227 (Fed. Cir. 1994). "Conception is the touchstone of inventorship" and is "complete only when the idea is so clearly defined in the inventor's mind that only ordinary skill would be necessary to reduce the invention to practice, without extensive research or experimentation." Id. at 1227-28. But the alleged unnamed inventor need not have invented the entire invention; he or she must instead have "made a contribution to the claimed invention that is not insignificant in quality, when that contribution is measured against the dimension of the full invention, and did more than merely explain to the real inventors well-known concepts and/or the current state of the art." Acromed Corp. v. Sofamor Danek Grp., Inc., 253 F.3d 1371, 1379 (Fed. Cir. 2001) (internal quotation marks and alteration omitted).

Based largely on practical concerns about the unreliability inherent in depending on "conception" in an inventor's mind—and the temptation to fabricate or inflate such evidence—the Federal Circuit has developed a rule of corroboration. "[T]he inventor must prove his conception by corroborating evidence, preferably by showing a contemporaneous disclosure." Burroughs Wellcome, 40 F.3d at 1228. "[A]n alleged inventor's testimony cannot satisfy the 'clear and convincing evidence' standard." Price v. Symsek, 988 F.2d 1187, 1194 (Fed. Cir. 1993). The Federal Circuit has applied a "rule of reason" analysis to determine whether the alleged inventor's testimony and corroborative evidence are satisfactory. Ethicon, Inc. v. U.S. Surgical Corp., 135 F.3d 1456, 1464 (Fed. Cir. 1998). While that analysis can be settled on summary judgment if there are no genuine disputes of material fact and the movant is entitled to judgment, the Federal Circuit has cautioned that this analysis will often require, "consider[ing] corroborating evidence in context, mak[ing] necessary credibility determinations, and assign[ing] appropriate probative weight to the evidence to determine whether clear and convincing evidence supports a claim of co-inventorship." Id. Corroborating evidence is usually "contemporaneous documents" of the unnamed inventor, the testimony of someone other than the alleged inventor, or circumstantial evidence. Id. at 1461.

The parties' argument boils down to whether Intel's corroborating evidence is insufficient as a matter of law. I conclude that that evidence is subject to genuine disputes of material fact that preclude summary judgment.

As Tela's argument focuses on corroboration, I briefly outline Pileggi's claims. As all agree, Pileggi and Becker met in March 2005 and Pileggi and his graduate students presented about then technology to Becker. See, e.g., Tela SJ Oppo. Ex. 1 ¶ 10. By then, Pileggi was in the process of forming Fabbrix. Id. ¶ 6. Pileggi says that he met with Becker at the suggestion of one of his venture capitalists, who thought Becker might make a good Fabbrix employee. Id. The meeting took place under a non-disclosure agreement ("NDA") that Pileggi suggested and Becker signed. See id. ¶¶ 7-8, 10. Pileggi states that, at the time, his research focused on addressing the lithographic gap, described above. See, e.g., id. ¶ 3. He represents that he was experimenting with layout techniques of chips that would help alleviate the problems with manufacturability. In particular, he was developing techniques to "increase geometric regularity" of the designs—that is, using "regular, repeating structures" that were, therefore, "easier to print" with lithography. Id. ¶¶ 3-4. One technique was to have the [Redacted] Id. ¶ 4. The [Redacted] above that gate layer would also be [Redacted] . Id. ¶ 5. Pileggi says he revealed all of this to Becker. Three months later, Becker incorporated Tela. See Tela SJ Oppo. Ex. 30.

The Asserted Patents disclose a "grid" to "facilitate parallel placement" of the features in each layer to "optimize[]" the lithography process. See '523 Patent at 11:1-17, Fig. 3A. They also provide for unidirectional gate features and disclose that metal interconnect layers can be added above the unidirectional gates. Id. at 14:20-28, 16:36-41, 17:39-44. Hook agreed that "the use of linear features [was] a significant aspect" of the Asserted Patents as a whole. Tela SJ Oppo. Ex. 22 at 178:1-179:9.

In response to Tela's argument, Intel points to three types of corroboration of Pileggi's testimony: the sworn statements of Dr. Veerbhan Kheterpal, then a graduate student of Pileggi's who attended the meeting; several documents from the time period, including from the meeting itself; and circumstantial evidence. In the light most favorable to Intel, the existence of this evidence precludes summary judgment. The facts here demand credibility determinations and weighing of contested evidence in context.

First, Intel points to Kheterpal's statements. At that time, Kheterpal was a graduate student under Pileggi. See Tela SJ Oppo. Ex. 2 ¶¶ 3-4. He also co-founded Fabbrix. Id. ¶ 2. Today, he is the co-founder and CEO of a supercomputer company. Id. ¶ 1. Kheterpal's account of events supports Pileggi's in many ways. He too says that Pileggi's research focused on the [Redacted] . Id. ¶¶ 5-6. He states that he, Pileggi, and several others met with Becker and that Pileggi did, in fact, present to Becker about "all aspects of the regular fabrics research." Id. ¶14. In particular, he agrees that Pileggi "showed and explained in detail [Redacted] ." Id.

Tela attacks Kheterpal's corroboration as insufficient. Tela appears to attempt to group together Pileggi and Kheterpal in its analysis and argue that then statements together need other corroboration. See, e.g., Tela SJ Reply 2 ("Intel seeks to fill in the holes in its theory with self-serving declarations from Prof. Pileggi and Dr. Kheterpal. But these declarations are insufficient as a matter of law because they are contradicted by the contemporaneous documents."). The argument seems to be that, because Kheterpal was Pileggi's student and co-founder at the time, he should be treated as a mere extension of Pileggi. Kheterpal is not now in that role to Pileggi, and it is the weight and credibility of his testimony today that is at issue. Moreover, Kheterpal is not trying to be named as an inventor, so his testimony is not suspect for that reason. While the Federal Circuit has sometimes held that the testimony of a non-inventor was insufficient because of the relationship between the witness and inventor, I cannot say that being someone's graduate student and commercial partner a decade-and-a-half ago is the sort of relationship that, as a matter of law, renders testimony not reliable. Indeed, one of Tela's leading cases on this point makes clear that the relationship must be evaluated on numerous dimensions, including the extent to which the witness has been "impeach[ed]" and the extent to which their credibility is affected by their interest in the litigation—determinations that are not appropriate for summary judgment on this record. See Sandt Tech., Ltd. v. Resco Metal & Plastics Corp., 264 F.3d 1344, 1351 (Fed. Cir. 2001).

Tela's other arguments about Kheterpal go to weight. It contends that Kheterpal is being paid by Intel for his involvement in the suit. Tela SJ Reply 6-7. That is a classic issue that goes to credibility and is not for summary judgment. It asserts that Kheterpal's statements are not detailed enough. Id. But his testimony, as it stands, supports Pileggi's; Tela points to no evidence in the record when it showed that more details from Kheterpal undermined his general point. That is a matter for cross-examination.⁸

8 Tela attacks some of Kheterpal's statements as inadmissible hearsay. See Tela SJ Reply 6. But they are not hearsay to the extent they merely report that things were or were not said, because the underlying statements are not being used for the "truth of the matter asserted," they are being used to show the statement was made. FED. R. EVID 801(c)(2).

Intel also points to contemporaneous documents. Primarily, it relies on the slide presentation purportedly given to Becker at the meeting. One slide states [Redacted] " and that [Redacted] " Tela SJ Oppo. Ex. 8 at 1136. Another states [Redacted] " Id. at 1139. The slides also include images that could arguably be construed as showing the [Redacted] . See id. at 1125. Intel also presents other, less direct, evidence that the pertinent information was communicated. For instance, the slides contain what Pileggi states are the test results from using his [Redacted] approach. Id. at 1125, 1127.

Tela argues that the corroboration in the slides is insufficient. It contends that the presentation is dated ten days after the meeting. Tela SJ Reply 4. But whether one believes that nothing material was changed by that revision is a matter of weight and credibility; in the best light for Intel, it is plausible that the material aspects of the presentation were shown to Becker and that the revision that occurred (if any) did not concern them. Tela also asserts that the presentation does not actually disclose one-dimensional "gridded layouts." Tela SJ Mot. 8-12; Tela SJ Reply 4. But the corroborative documents need not live up to any set bright line or disclose all aspects of the contribution to the invention; instead, all of the evidence taken together must be corroborative. In the best light for Intel, the slides arguably corroborate Pileggi's account because they disclose his [Redacted] . That the slides do not, as Tela argues, disclose every aspect of the claimed invention or point to disclosure of a different technology entirely does not mean there is no genuine dispute of material fact here—indeed, it underscores that there is.⁹

9 Intel also points to circumstantial evidence. I conclude, however, that Kheterpal's declaration and the slide presentation create at least a triable issue of fact, so I do not address the purported circumstantial evidence.

Tela relies on several Federal Circuit cases to attempt to show that the slide presentation cannot serve as corroborative evidence because it itself requires Pileggi's interpretation or, in Tela's view, corroboration. In other words, Tela argues that this evidence results in a "catch-22" in which the slides are needed to corroborate Pileggi and Pileggi's testimony is needed to explain why the slides corroborate him. I take Tela's point that this documentary evidence is not as strong or self-evident as is many cases, but none of Tela's authorities render it unreliable as a matter of law.

In Tela's leading case, Apator Miitors ApS v. Kamstrup A/S, the purported unnamed inventor relied on emails and drawings to corroborate his testimony. 887 F.3d 1293, 1296 (Fed. Cir. 2018). There, the key alleged corroborative evidence about each email came from the inventor's testimony rather than the emails themselves. See id. For instance, one email merely said a "sample is attached" and the inventor's testimony provided all of the detail that the attachment supported his inventorship claim. Id. More helpful for Tela is the Federal Circuit's discussion of the drawings. Like the slides here, the drawings indicate they were "modified" at a later date but the unnamed inventor claimed they were created earlier.

Here, however, there are two important differences that are inappropriate to resolve at summary judgment. First, unlike in Apator, Pileggi's testimony is not the only support that the subject matter of the slides was presented; Kheterpal testifies to that effect too. Second, there is some circumstantial evidence that these slides were shown to Becker, including that they fit with the agenda provided for the meeting and the names of the various slide presentations concern Becker. Accordingly, this issue is for the finder of fact, not summary judgment.

A reasonable factfinder could credit (1) Pileggi's representations, (2) Kheterpal's representations, and (3) the slide presentation to Becker as Intel characterizes it. If all of that evidence were credited, a reasonable factfinder could see clear and convincing evidence of Pileggi's contribution to the Asserted Patents. While I agree with Tela that the evidence is somewhat mixed, genuine disputes of material fact preclude summary judgment. Tela's motion for summary judgment on the inventorship-based claims and defenses is DENIED.

D. Breach of Contract and Covenant Not to Sue

As I explained above, the parties entered into the CNTS on May 9, 2007. Count X of the SAC requests declaratory judgment because, it alleges, the CNTS bars Tela from claiming that Intel infringed the Asserted Patents. See SAC ¶¶ 189-93. Count XI of the SAC alleges breach of contract based on Tela allegedly breaching the CNTS by arguing the Accused Products infringe the Asserted Patents. Id. ¶¶ 194-201. Two of Intel's counterclaims and one of its affirmative defenses mirror these allegations. See CIR ¶¶ 144-45, 214-26.

The parties' dispute on this issue is over the interpretation of the CNTS. The CNTS provides that Tela agrees that it "shall not Assert any Company Patent Right against Intel . . . for the manufacture, use, import, offer for sale or sale of any of Intel's Products or any process or method employed in the manufacture, testing, distribution, or use thereof for so long as Intel does not Assert any Intel Patent Right against [Tela]." CNTS at ECF 3. The parties agree that the broad definition of "Assert" includes Tela's position in this case. Their disagreement is about whether the Asserted Patents fall within the definition of "Company Patent Right." The CNTS defines "Company Patent Right" to mean "any Patent Rights of [Tela] that claim priority during the term, or within two (2) years of the expiration of termination of this Agreement." Id. (emphasis added).

Tela argues that all of the Asserted Patents "claim priority" to March 9, 2006, the date of the '288 Provisional, which is before the CNTS went into effect on May 9, 2007. See Tela SJ Mot. 22-24. Intel contends that "claim priority" does not mean that the patents themselves state that they claim priority on their face, but instead that the patent must be legally entitled to priority outside of the term of the CNTS. See Tela SJ Oppo. 23-25. In other words, Intel argues that "if Tela's alleged pre-CNTS priority documents do not satisfy" the statutory requirements of Section 112, "the Asserted Patents [do] not 'claim priority' before the CNTS's May 9, 2007 effective date." Id. 24. Intel further argues that, as a result, there is a genuine dispute of material fact. Id.

I conclude that Tela's interpretation of the CNTS is correct and, consequently, it is entitled to summary judgment. The phrase "claim priority," the parties agree, is a specialized term in this context.¹⁰ The Federal Circuit has frequently differentiated between "claiming" priority and actually being "entitled" to it. In Nat. Alternatives Int'l, Inc. v. Iancu, for instance, the court critiqued an argument as "conflat[ing] properly claiming priority and demonstrating entitlement to priority." 904 F.3d 1375, 1380 (Fed. Cir. 2018). In In re NTP, Inc., the court stated that "a patent's claims are not entitled to an earlier priority date merely because the patentee claims priority." 654 F.3d 1268, 1277 (Fed. Cir. 2011). The PTO, too, appears to often discuss the two as separate. See, e.g., U.S. Patent and Trademark Office, Manual of Patent Examining Procedures, Section 211 ("See MPEP § 211 et seq. for the detailed requirements for claiming (and entitlement to), the benefit of an earlier-filed application under 35 U.S.C. 119(e), 120, 121, 365(c), or 386(c)."). When a patent "claims priority," the ordinary understanding in this context is that the patent claims priority on its face. I agree with Intel to the extent that the phrase "claim priority" would not always and everywhere mean "claim priority on the face of the patent." And the Federal Circuit will often explicitly contrast "claiming priority on the face of the patent" with entitlement. But Intel has presented no Federal Circuit case—or other authority that would shed light on the usual meaning of the phrase in this context—that conflates a patent "claiming priority" with being adjudicated to be entitled to priority.

10 The CNTS provides that it "shall be governed by and construed in accordance with the laws of the State of Delaware." CNTS at ECF 4. Neither party applies the law of a particular state, perhaps because the question of interpretation turns on a term of art specific to patents. In any event, on this basic question, Delaware law is consistent with general principles of contract interpretation that I and the parties apply. See, e.g., United Rentals, Inc. v. RAM Holdings, Inc., 937 A.2d 810, 830 (Del. Ch. 2007) (laying out fundamental interpretive principles).

To reach a contrary interpretation, Intel points to 35 U.S.C. § 120, which governs when patent applications are "entitled to the benefit" of the date on an earlier filed application. It is true that claims of priority in patents are made to meet the requirements of that statute, so Intel's argument is not wholly groundless. But it is unconvincing. Section 120 does not itself use the phrase "claim priority." Instead, patents themselves "claim" priority, and, often times, they are "entitled" to priority under Section 120. Intel's interpretation, moreover, stretches the plain language of the CNTS significantly further than Tela's. And Intel's argument would mean that the parties imposed odd logistical barriers on themselves. For a patent to qualify under the CNTS under Intel's interpretation, it would presumably need to be adjudicated to be entitled to priority; otherwise, there would be no conclusive determination on that front.

Not only is Intel's interpretation out of step with the CNTS's plain language, it makes little sense given that document's purpose and structure. In consideration for Intel's investment in Tela—as the CNTS explicitly contemplates, see CNTS at ECF 2—Tela gave up, among other things, the ability to sue Intel during the period of their agreement based on its patents. But the '228 Provisional was filed prior to the CNTS. The parties did not even arguably intend Tela to relinquish any rights related to that Provisional. Yet, under Intel's interpretation, Tela would be giving up rights related to it merely because it happened to file the other applications within the period of the agreement. The agreement, it is clear, is intended to prevent assertions of rights that grew out of the term of the parties' relationship, which would not include applications that claim priority before the relationship.

If the parties had wanted to adopt the more expansive approach Intel urges, the straightforward way of doing so would be to have provided that Tela could not assert rights under any patent application "filed" during the period of the CNTS, or "filed and not found to be entitled to the earlier priority date they claim." Instead, the parties chose to use the familiar language of "claim[ing] priority" during the relevant period.

Intel also argues that this interpretation is incorrect because it "would grant Tela unilateral discretion to evade the CNTS by making an unsubstantiated allegation of pre-CNTS priority on the face of a patent" which would "render this provision of the CNTS meaningless." Tela SJ Oppo. 24. Its argument is overstated. There are significant incentives not to make "unsubstantiated allegations" to the PTO. See, e.g., 37 C.F.R. 1.56 (discussing the duty to disclose material information). Parties do not need to rely solely on private contracts to prevent each other from misleading the government. That provision is no more "meaningless" than any other contractual provision that could be theoretically thwarted by a party acting in sufficiently bad faith.

Although it does not repeat the assertion in its substantive argument section, Intel's "Introduction" argues that it "negotiated the 'claim priority' language precisely to restrict the patents falling outside the CNTS's scope to those developed before the parties entered into the agreement." Tela SJ Oppo. 7. One party's internal, unmanifested desire, however, does not dictate the meaning of a contract—especially in the absence of ambiguity. See, e.g., Restatement (Second) of Contracts § 202. And Intel cites no record evidence in support of this contention, or that shows that this was the parties' mutual understanding.

Finally, Intel is incorrect that there is a genuine dispute of material fact because of this dispute over interpretation. Tela SJ Oppo. 24-25. All of the cases it cites in support stand for the well-established principle that when there is a genuine ambiguity in the contract, there may be underlying disputed questions of fact about the parties' intent. See, e.g., Whittlestone, Inc. v. Handi-Craft Co., No. C 08-04193 SBA, 2012 WL 3939629, at *9 (N.D. Cal. Sept. 10, 2012) (finding the contract ambiguous and the parties' extrinsic evidence contradictory). Here, however, I do not find the contact ambiguous.

Tela's motion for summary judgment on this issue is GRANTED.

E. Inventorship and CNTS Conclusion

Genuine disputes of material fact preclude summary judgment on Intel's inventorship-based claims and defenses. Tela has, however, established it is entitled to summary judgment on the claims based on its alleged violation of the CNTS.

III. MOTIONS TO EXCLUDE AND STRIKE

A. Intel's Motions to Strike and Exclude Opinions of Dr. Foty

Intel moves to exclude the opinions of Tela's technical expert, Dr. Daniel Foty. See Intel's Daubert Motion to Exclude Opinions of Dr. Daniel Foty ("Foty Mot.") [Dkt. No. 246]. Separately, Intel moves to strike portions of Dr. Foty's report for Tela's alleged failure to properly disclose the theories they relate to and on grounds of timeliness. See Intel's Motion to Strike Portions of Tela's Expert Reports ("Foty/Greene Mot.") [Dkt. No. 244].

i. Failure to Disclose

"Patent Local Rule 3 requires patent disclosures early in a case and streamlines discovery by replacing the series of interrogatories that parties would likely have propounded without it." ASUS Computer Int'l v. Round Rock Research, LLC, No. 12-CV-02099 JST (NC), 2014 WL 1463609, at *1 (N.D. Cal. Apr. 11, 2014) (internal quotation marks and alteration omitted). This rule and others "are designed to require parties to crystallize their theories of the case early in the litigation and to adhere to those theories once they have been disclosed." Verinata Health, Inc. v. Sequenom, Inc., No. C 12-00865 SI, 2014 WL 4100638, at *1 (N.D. Cal. Aug. 20, 2014). As a result, "[i]t is well settled that expert reports may not introduce theories not set forth in [infringement] contentions." Finjan, Inc. v. Cisco Sys. Inc., No. 17CV00072BLFSVK, 2020 WL 2322923, at *3 (N.D. Cal. May 11, 2020) (internal quotation marks and alteration omitted). Courts in this District frequently strike portions of expert reports that attempt to circumvent the disclosure rules. See e.g., id.; Verinata Health, Inc. v. Sequenom, Inc., No. C 12-00865 SI, 2014 WL 4100638 (N.D. Cal. Aug. 20, 2014), at *3.

1. Thermal Annealing

Intel argues that Tela did not disclose, in its infringement contentions, a theory advanced by Foty to support Tela's infringement Opposition. At the hearing, Tela represented it is no longer pursuing the theory based on its understanding of Intel's theories. For avoidance of potential future doubt or confusion, I nonetheless grant the motion to strike it.

As explained above, one dispute in this case is over "diffusion regions," which the asserted claims require. The parties agreed to a construction of "diffusion region" as "selected portions of the substrate within which impurities have been introduced to form the source or drain of a transistor." Claim Construction Order 5. After fact discovery had closed, Tela moved to submit supplemental infringement contentions. See Dkt. No. 191. Among other things, Tela sought to add its argument that [Redacted] , a step in the process by which Intel creates sources and drams. See Foty/Greene Mot. Ex. 12. I denied Tela's request to supplement its infringement contentions. Dkt. No. 199. That proposed amendment, I explained, was untimely. Id . at 1. Foty's expert report now states that the [Redacted] " See Foty/Greene Mot. Ex. 1 ¶ 156. In its papers, Tela relies on that theory (and others) in responding to Intel's motion for summary judgment. See Intel SJ Oppo 6.

I agree with Intel that Tela is attempting to use its expert to put forward a new infringement contention that it did not originally advance and was denied leave to add via amendment. Tela did not disclose this theory in any of its disclosures and only asked for leave to do so after fact discovery was closed. This is the sort of situation the Patent Local Rules seek to avoid. See, e.g., Finjan, 2020 WL 2322923, at *3.

Tela's primary response is that it adequately disclosed the theory in its contentions. See Tela's Opposition to Foty/Greene Mot. ("Foty/Greene Oppo.") 6-7. It correctly notes that it need not have disclosed the theory word-for-word and it asserts that Foty's opinion only adds "details" to its theories about diffusion regions. Id . 1. But the purported "disclosures" of the theory are not disclosures at all; it points to nothing in the record showing that it previously disclosed that the diffusion regions Tela urges exist were caused as a [Redacted] , rather than the intentional epitaxial deposition that Intel uses. Those are quite different mechanisms for why Tela believes its products are infringed. It also argues the information was already in Intel's possession, id . 7-8, but that would be true with any owner of an Accused Product; the reason to disclose is not that the underlying technology is unknown, it is that the infringement theory is not.

Tela also argues that Foty's opinion about thermal annealing is necessary to rebut a theory that Intel disclosed on the last of day of discovery—which would mean that Tela could not have responded before the discovery cutoff. Id . 6-7. This argument does not sit entirely comfortably with Tela's argument that Foty's theory was adequately disclosed. Setting that aside, there is nothing in the record that Tela has pointed to that supports its argument that Intel advanced a new theory on the final day of discovery. Tela asserts that Intel only stated on the final day of discovery that "its epitaxial deposition process for forming source and drain regions does not meet the construction of 'diffusion region.'" Id . But in an August. 14, 2019, deposition, Tela asked Auth about diffusion regions and he replied, in part, that " [Redacted] " Intel's Reply in Support of Foty/Greene Mot. ("Foty/Greene Reply") [Dkt. No. 271] Ex. 45 at 133:12-17. And, of course, one of Intel's noninfringement contentions has long been that its products lacked diffusion regions; Tela was on notice to conduct discovery into why. Importantly, it is Tela's burden to show infringement and put forward infringement contentions. I recognize that the disclosure posture of this case is somewhat different than in many patent cases: Intel brought noninfringement claims against Tela, while Tela was required to then put forward its infringement contentions. All the same, it was Tela's burden to make clear (with sufficient time to conduct discovery) that it viewed its products as infringed because thermal annealing [Redacted] . It did not.

The motion to strike on this ground is GRANTED.¹¹

11 Intel also moves to exclude Foty's opinions on this subject as being irrelevant, misleading, and inadequate under Daubert. Foty Mot. 8-12. Because I conclude that the opinions should be struck on the basis described, there is no need to address these arguments.

2. First-Metal Vertical Grid

Intel moves to strike portions of Dr. Foty's report that, it contends, present a theory of infringement that was undisclosed in Tela's infringement contentions. Foty states that [Redacted] " for 14 nm products and [Redacted] " define the claim requirement of a "first-metal grid." See Foty/Greene Mot. Ex. 3 ¶¶ 325, 458. In its infringement contentions, Tela defined the first-metal grid and stated, "[f]or example, the metal structures are centered about a metal grid separated by a metal pitch and extend lengthwise in an x-direction." Id. Ex. 8 at 23. According to Intel, Tela's contention "was that the claimed 'grid' is defined by where the features are positioned, whereas Dr. Foty now opines that the 'grid' is defined by the size of the features and by how close their edges can be to each other." Foty/Greene Reply 5. The motion to strike is DENIED. Foty's explanation only adds detail to Tela's overarching theory of infringement. At most, he explains how the first-metal features (identified in the contentions) are positioned, rather than introducing a new argument for why Intel's products infringe.

3. Transistor Configurations

In its claim charts, Tela asserted that a claim was met on Intel's 10 nm products based on the layout of cells. It then mapped transistor configurations onto those cells in support. It states that this "is evidenced, for example, by" one cell layout in particular. It then maps the purported claim components onto the cell:

[Redacted] Foty/Greene Mot. Ex. 10. Intel objects that, in his report, Foty includes 33 additional cells with this claim mapping. Intel argues that this inclusion amounts to advancement of new infringement theories not previously disclosed. Foty/Greene Mot. at 7-9. I disagree. Unlike Foty's thermal annealing opinions, the overall theory of infringement was disclosed by Tela: that the cells met the requirements of the claim for the stated reasons. The claim chart said that the image above was an example. Tela was not obligated to disclose every particular manifestation of its theory. I recognize that Tela attempted to amend its infringement theories to add, among other things, claim mapping of the new cells in Foty's report. That, however, does not necessarily mean that its arguments about these cells were not fairly encompassed by its broader infringement theory. Tela could, for instance, have wanted assurance that the cells were included in the case prior to submitting Foty's report. The motion to strike these opinions is DENIED.

4. First-Metal Layer

Intel argues that Foty's depiction of what constitutes a "first-metal layer" in the 14 nm products differs from the depiction in Tela's infringement contentions. Tela replies that the parties were aware that the image Intel relies on from its contentions was inadvertently mislabeled and that other portions of the contentions contained images that are correctly labelled and align with Foty's report. The images below are, respectively, the image Intel relies on from the infringement contentions, one of the images Tela relies on from its infringement contentions (and its legend), and what appears in Foty's report:

[Redacted] See Foty/Greene Mot. Ex. 11 at 22, Ex. 5 ¶ 319; Foty/Greene Oppo. Ex. C at 47.

This request to strike is DENIED. First, Intel does not dispute it was on notice, based on Foty's deposition, that the labelling of the lower protrusions in the first image was a mistake. See Foty/Greene Oppo. Ex. H at 492:14-19. Indeed, it has no answer at all on this point. Second, Intel's only response to Tela's inclusion of the second image above in its infringement contentions is that it appeared in a different set of contentions than the image Foty now includes. While that may make a difference in some circumstances, I do not find it sufficient to strike this portion of the report. Even if it was related to other infringement contentions, it put Intel on sufficient notice that Tela was marking the structure in the second image as the first-metal layer; Intel has offered no reason that that designation would not be equally relevant to all claims, especially if it was aware that the first image was incorrect.

ii. Supplemental Report

Tela was required to submit expert reports on infringement by April 8, 2020. On June 24, 2020, Tela submitted what it styled as a "supplemental" expert report by Foty that analyzed the ITC administrative law judge's ("ALJ") conclusions, which issued on May 22, 2020. See Foty/Greene Mot. Ex. 6. Intel argues that report should be stricken on three grounds. I agree with Intel on one: the report is improper under the Federal Rules of Civil Procedure. Tela asserts it was able to submit the report past the cutoff because of FRCP 26(e), which governs supplementing discovery responses. See Foty/Greene Oppo. 15. Under that rule, when a party has made a disclosure under the FRCP, the party, "must supplement or correct its disclosure or response . . . in a timely manner if the party learns that in some material respect the disclosure or response is incomplete or incorrect." FED. R. CIV. P. 26(e)(1). Foty's supplemental report does not, however, correct any "incomplete" or "incorrect" prior disclosure. Instead, the report exclusively discusses Foty's areas of agreement and disagreement with the ALJ.

Tela argues that Foty could not have issued the report prior to the ALJ's determination. That is true, but it does transform a filing into one that corrects past "incomplete" or "incorrect" disclosures. The parties do not have an automatic right to file expert reports simply to respond to unfavorable adjudications; they have the right to file expert reports in accordance with the case schedule. Especially in the patent context, the cutoffs for expert reports are not an empty formality. The patent rules are concerned, among other things, with preventing suits from becoming sprawling and unmanageable by, for instance, the boundless filing of new expert reports. Rule 26(e) exists to impose a duty on parties to correct past filing, it does not permit end-running case schedules. The motion to strike this report is GRANTED.

iii. Layout Files, Design Documents, and Fabrication Steps

As alluded to above, Foty relies on certain design-stage documents and fabrication steps of Intel's products to show that their "gate contact structures" are separate layers from the gate layer and interconnects, as the claim construction requires. Foty also bases some opinions on "diffusion regions" on layout files rather than the physical chips. Intel argues that these opinions should be excluded under Daubert because they are irrelevant and will create confusion.

"A patentee may prove infringement by any method of analysis that is probative of the fact of infringement and circumstantial evidence may be sufficient." Martek Biosciences Corp. v. Nutrinova, Inc., 579 F.3d 1363, 1372 (Fed. Cir. 2009) (internal quotation marks and citations omitted). But "to infringe an apparatus claim, the device must meet all of the structural limitations." Cross Med. Prod., Inc. v. Medtronic Sofamor Danek, Inc., 424 F.3d 1293, 1311-12 (Fed. Cir. 2005). Accordingly, the parties' dispute must be centered on the structure of the device. This is not to say that other evidence cannot, in some circumstances, act as a stand-in for the device's physical structures. As Tela correctly points out, courts may compare an "industry standard" to the claims to determine infringement. Fujitsu Ltd. v. Netgear Inc., 620 F.3d 1321, 1327 (Fed. Cir. 2010). Tela leaves out, however, the reason behind this: "claims should be compared to the accused product to determine infringement. However, if an accused product operates in accordance with a standard, then comparing the claims to that standard is the same as comparing the claims to the accused product." Id. (emphasis added). And, the Federal Circuit held, the other party could show that the structure did not, in fact, operate the same as the industry standard. See id.¹²

12 Tela points out that Intel did not identify specific paragraphs it seeks to strike with these motions. I therefore describe the substance of what will be struck.

1. Upstream Documents

As I explained above, some of the asserted claims require "[gate] contact structure(s)" that the 14 nm and 10 nm Accused Products do not have. Without repeating my analysis in detail, I find that the accused contact structures are, for somewhat different reasons for each product type, not "separate from" the interconnect layers. Foty opines that they are separate structures for several reasons. One is based on what Intel calls "upstream layout shapes" or "design-stage abstractions." The motion to strike reliance on these is GRANTED. Tela had and has the composition and structure of the accused products available to it, yet these opinions focus an attenuated abstractions, not the completed structures. This is not a case in which the structure itself is unavailable or in which there is conclusive evidence that a proxy for it is necessary or reflective of the physical device. See Fujitsu, 620 F.3d at 1327 ("[C]laims should be compared to the accused product to determine infringement.").

Tela makes two types of counterarguments. The first is to essentially repeat its substantive arguments on the merits. See Foty Oppo. 8-9. I have addressed the merits above. The second is that the Federal Circuit and other courts have approved of consideration of this type of evidence. One of Tela's Federal Circuit cases is Fujitsu, discussed above. That case permitted a court to rely on an industry standard where the product conformed to that standard. Fujitsu, 620 F.3d at 1327. The court there was clear that: "We agree that claims should be compared to the accused product to determine infringement." Id. (emphasis added). Consideration of the standard was permissible precisely because, in that case, it was "the same" as considering the product. Id. And, the court made clear that if an industry standard did not shed sufficient light on the accused product, it would not be considered in the same way. Id. at 1327-28.

Tela also relies on Martek. There, however, the experts "conceptually analyzed the accused process and testified that it must meet the functional claim limitation based on the composition of Lonza's culture medium and the known effects of chloride concentration on stainless steel corrosion." Martek, 579 F.3d at 1373-74. In other words, the experts analyzed the components of the accused product and applied an established scientific principle to them. Indeed, the court there explicitly approved of an earlier precedent's admonition that the patentee "did not prove infringement because she presented no testimony based on the accused products themselves that supported a finding of infringement." Id. (internal quotation marks omitted). Again, then, the Federal Circuit's approach is not so rigid as Tela would have it; instead, the court has permitted evidence of infringement when it actually reflected the reality of the accused products themselves. It is worth underscoring that what is excluded here are either (1) simple diagrams that map structures in abstract ways or (2) design "rules" that conceptualize structures as different layers even when both parties experts agree that a single structure is present. The motion to strike is GRANTED to the extent described. For clarity, I do not strike these documents merely because they are renderings; both parties rely on renderings of the physical device that have not been objected to.

Even if this evidence were admissible, it would not create a genuine dispute of material fact. If the actual physical structure of an accused product does not infringe but an abstraction of the product found in internal design literature would, it does not change the fact that the accused product does not infringe.

2. Fabrication Steps

Intel groups together the above upstream documents with the actual steps taken to fabricate the devices. I separate the analysis because different concerns are implicated. Foty opines that because [Redacted] are made in multiple steps, the structures are separate. That is quite different from relying on abstractions unconnected to the physical device. These steps do relate to the physical structure of the device. The motion to strike reliance on them is DENIED. For the reasons explained above, however, that [Redacted] in multiple steps does not create a genuine dispute of material fact.

3. Diffusion Region Layout Shapes

Intel also moves to strike Foty's reliance on certain layout shapes to form his opinions about diffusion regions. For largely the same reasons described in the first subsection above, I will strike Foty's reliance on these layout shapes. They are, again, abstractions that do not necessarily reflect the structure of the Accused Products. The physical structure of the products is entirely available, so use of these shapes contradicts Federal Circuit precedent and would be unduly misleading and confusing. Indeed, Foty also relies on teardown images. If the evidence were admitted, it would not create a genuine dispute of material fact because the physical structure of the Accused Products is established and the claim term has been construed; Intel's internal labelling of areas as "diffusion regions" or layout shapes showing the same does not change those realities. At the healing, Tela argued that the patents themselves have layout shapes, so it relied on them here. But in an infringement analysis, the task is to compare the construed claim to the accused product. The motion is GRANTED. Foty's reliance on the layout shapes is struck but not his reliance on other evidence or opinions that do not depend on them.

iv. 45 nm Products as Prior Art

Foty opined that Intel's 45 nm products—a previous generation than the Accused Products—are not prior art. Intel seeks to exclude his opinions on this topic for several reasons.

1. Abandonment, Suppression, and Concealment

Foty argues that Intel abandoned, suppressed, and/or concealed the 45 nm products, contravening Intel's expert's opinion that Intel did not do so. Abandoned, suppressed, or concealed art does not entitle an inventor to a patent and cannot invalidate a patent. 35 U.S.C. § 102(g); Checkpoint Sys., Inc. v. U.S. Int'l Trade Comm'n, 54 F.3d 756, 761 (Fed. Cir. 1995).

In Fox Group, Inc. v. Cree, Inc., 700 F.3d 1300 (Fed. Cir. 2012), the Federal Circuit addressed the standard for abandonment, suppression, or concealment. It differentiated there between what is required for a process claim and a product claim. One way that a product can be shown to not have been abandoned, concealed, or suppressed is through commercialization of that product. See Fox Grp., 700 F.3d at 1306. In Fox Group, the party arguing that the claim had been suppressed contended that, even though there the product had been publicly disclosed, it was not done so in a way as to enable it. Id. The Federal Circuit said that a requirement that the commercialized product be enabling applied to process claims. See id. at 1306-07. But, it said, there was a distinction between product and process claims on this front. Id. at 1307. The purpose of Section 102(g), the court said, "is to bar an inventor from receiving a patent on an invention that has already been invented and was not abandoned, suppressed, or concealed." Id. The rationale appears to be that the disclosure of a product—as opposed to a process—is sufficient to show there was no abandonment, concealment, or suppression. Accordingly, the court held that the appellee "ma[king] its invention . . . known to the public," even if not enabling, was sufficient. Id.

Intel argues that Foty's opinions effectively contravene Fox Group because he opines that Intel was required to make an enabling disclosure of its 45 nm products, as opposed to simply a public disclosure. Foty Mot. 17-18. Foty opines that Intel's technology was not prior art because it was abandoned, suppressed, or concealed; the public, commercial release of its 45 nm chips, Foty argues, it not sufficient to enable because Intel kept secret test chips and layout files. To the extent that Foty's opinions rest on this basis, I agree that they squarely contradict Fox Group and are, accordingly, contrary to law; they would, by the same token, be inadmissible under Daubert and FRE 403 as it would confuse the issues and mislead the trier of fact. See, e.g., Apple, Inc. v. Samsung Elecs. Co., No. 11-CV-01846-LHK, 2012 WL 2571332, at *6 (N.D. Cal. June 30, 2012).

Tela counters that this issue turns on whether the alleged suppression was "intentional" or "inferential." Foty Oppo. 15-16. But Intel's argument is that Foty's opinion is contrary to law to the extent it states that enablement is required—regardless of what form of suppression occurred. Tela then relies on the dissent from Fox Group.¹³ Puzzlingly, Tela states, "[f]or avoidance of doubt, Tela is not asking this Court to adopt [the] dissent in order to deny Intel's Motion." Id. 16. But it then goes on to say, that "Intel's entire Daubert Motion with respect to Dr. Foty's 45nm opinions is based off of a 2-judge majority decision that the third judge on the panel confirmed is in contradiction with prior established Federal Circuit law shows just how tenuous Intel's arguments are" and previously said that the dissent "confirms the flaws in Intel's reasoning." Id. A two-judge majority opinion is, of course, just as binding on me as a three-judge majority. The dissent may well have adopted Tela's position, but I must follow the majority's guidance.

13 That opinion concurred in part and dissented in part, but Tela relies on portion that dissented.

The motion to exclude Foty's opinions at trial to the extent they rely on the principle (or purport to conclude) that enablement is required for public disclosure under Section 102(g) is GRANTED.

2. Use of Confidential Documents

Foty states that Subramanian's prior-art analysis is incorrect, among other reasons, because it relies on confidential information not available to the public. Intel argues that such opinions are incorrect as a matter of law. Foty Mot. 18-19. The two cases it cites as authority, however, do not compel that conclusion. In one, the court's discussion of confidentiality was about whether the documents could have been obtained sooner for purposes of amendment. Illumina, Inc. v. Natera, Inc., No. 18-CV-01662-SI, 2019 WL 2077850, at *1 (N.D. Cal. May 10, 2019). The allegedly confidential information in the other was apparently used to establish the date of invention, not the state of the art. Fox Grp., Inc. v. Cree, Inc., 819 F. Supp. 2d 524, 531-32 (E.D. Va. 2011), aff'd in part, vacated in part, 700 F.3d 1300. The motion is DENIED.

3. Differentiation Between the Test Chip and Product

One of Intel's arguments is that the 45 nm test chip and the Penryn processor that it ultimately sold qualify as the "same invention" for purposes of the prior-art analysis. The former, Intel asserts, is a prototype of the latter. See Foty Mot. 19-20. Intel's experts have introduced evidence they claim supports this view. See, e.g., Foty Mot. Ex. 19 ¶ 285. Tela, on the other hand, maintains that the test chip and Penryn processor are sufficiently different that they do not qualify as the same invention. Foty opines that the two are different and offers various theories in support. See, e.g., Foty Mot. Ex. 3 ¶¶ 491-94.

Intel seeks to strike Foty's opinions on the basis that he erroneously treats the two chips as separate inventions. Intel relies principally on Eolas Technologies Inc. v. Microsoft Corp., 399 F.3d 1325 (Fed. Cir. 2005). There, the district court held that, because a later version of the product was made, the prototype was "abandoned." Id. at 1333. The Federal Circuit held that the definition of "invention," however, could encompass "improved version[s]" of the products. Id. Intel argues that Eolas shows that Foty's treatment of the 45 nm test chip and the commercialized Penryn product is contrary to law

This argument begs the question. Eolas was on appeal from summary judgment after full evidentiary findings about the earlier and later versions were made. Intel effectively seeks to turn a factual dispute—whether a previous version is simply "improved" such that it is the same invention or whether it was so changed as to be a different one—into a legal one. Here, there have been no findings of fact about the differences between the two. Indeed, the differences (or lack thereof) is the heart of the disagreement between the experts and parties on this topic. All of Intel's arguments relate to whether Foty's opinions are ultimately correct, not to whether they have a sound basis under Daubert. Intel argues, for instance, that none of Foty's evidence shows that the differences between the two are material to the dispute. But that is the very question to be answered: whether the two are different in ways that render them not the "same invention." That does not mean, however, that presenting the opinions to the trier of fact runs afoul of Eolas . The fundamental flaw of Intel's argument is on display when Intel states that Foty's opinion "contradicts Dr. Auth's unrebutted testimony [Redacted] applied to both the 45 nm test chip and the subsequent commercial processor" while Foty opines that the [Redacted] . Foty Mot. 21. Auth's testimony would only be "unrebutted" if Foty's were excluded—a quintessential factual dispute. The request is DENIED.

4. Reduction to Practice

Foty opines that there is insufficient evidence showing that the 45 nm test chip was reduced to practice by the date Intel asserts. "In order to establish an actual reduction to practice, an inventor's testimony must be corroborated by independent evidence." Cooper v. Goldfarb, 154 F.3d 1321, 1330 (Fed. Cir. 1998). Intel attempts to corroborate its claims of reduction to practice with a press release issued at the time it claims the 45 nm test chip was reduced to practice. Foty opines that the press release does not contain sufficient technical detail to show that the chip had been fabricated and work for its intended purpose. See Foty Mot. Ex. 3 ¶ 472. Intel argues that these opinions are contrary to law because there is no specific form corroboration must take, including that the documents contain a certain level of technical detail. That is true, and to the extent Foty suggests to the trier of fact that there is a legal standard requiring that, his testimony would be improper. But I do not take Foty to say that.

Instead, the opinion Intel objects to is that what was disclosed in the press release, in his view, does not corroborate certain aspects of the invention. While "reduction to practice does not require corroboration for every factual issue contested by the parties," Goldfarb, 154 F.3d at 1330, which facts are and are not corroborated is certainly relevant and helpful. Intel's argument is better aimed at the trier of fact. Whether the trier of fact believes the press release is adequate corroboration is a matter for it. See id. ("The rule requires an evaluation of all pertinent evidence when determining the credibility of an inventor's testimony. In order to corroborate a reduction to practice, it is not necessary to produce an actual over-the-shoulder observer. Rather, sufficient circumstantial evidence of an independent nature can satisfy the corroboration requirement." (internal citations omitted)).

Intel also objects to Foty's opinion regarding Auth's potential testimony. Intel expects Auth to testify that the 45 nm test chip was reduced to practice at the relevant time. See Foty Mot. 22-23. At his deposition, Foty was asked if he could dispute Auth's knowledge of the test chip's alleged reduction to practice. He replied, "I can't dispute it, but as I said, Dr. Auth is testifying not in January 2006 so Dr. Auth's testimony was not available to a person of skill in the art doing this hypothetical analysis in January 2006. That information was not available to that person." See Foty Mot. Ex. 2 at 516:12-518:9. Intel argues that, "[c]ontrary to Dr. Foty's suggestion, the jury is permitted to consider evidence of reduction to practice that was not publicly available at the time, including inventor testimony." Foty Mot. 23. I agree, and to the extent Foty purports to instruct the jury about the legal standard, he may not. But, again, I do not read Foty's statements as Intel does. Nowhere in this statement (which occurred in answer to a deposition question) is he saying it would be improper to consider such statements.

Finally, I agree with Intel that one of the opinions it objects to is inadmissible. Foty states that, "One can imagine that in a chip with more than a billion transistors, one cell not being tested or functioning properly would not prevent a press release claiming successful fabrication." Foty Mot. Ex. 3 ¶ 472. That statement is not plausibly within Foty's expertise as a technical expert; it does not concern the technology at issue or, really, the industry in general. As a lay opinion, it is inadmissible speculation, as Intel argues. See FED. R. EVID. 602; 701(a). Intel's motion is GRANTED IN PART only as to this opinion; it is otherwise DENIED.

v. Conclusion Regarding Foty

Intel's motion to strike is GRANTED IN PART with respect to Foty's opinions about thermal annealing, supplemental report, and reliance on layout files and design abstractions; it is otherwise DENIED. Intel's motion to exclude Foty's testimony is GRANTED IN PART with respect to his opinions regarding enablement being required for abandonment, suppression, and concealment and the single speculative sentence about the press release; it is otherwise DENIED.

B. Intel's Motions to Strike and Exclude Opinions of Mr. Greene

Intel moves to strike portions of the expert report of Jed Greene, Tela's damages expert, because they were not disclosed. See Foty/Greene Mot. And it moves to exclude testimony of his under Daubert. See Intel's Motion to Exclude Certain Damages Opinions of Jed Greene ("Greene Mot.") [Dkt. No. 242].

i. Failure to Disclose

The law on a party's failure to disclose contentions that appear in expert reports is stated above and will not be repeated.

1. U.S. Sales-Based Royalty Rate

Intel contends that Tela's damages contentions were based, in part, on an effective royalty rate agreement between Tela and Qualcomm for Qualcomm's worldwide alleged infringing sales. Greene's expert report, Intel argues, for the first time bases the rate on Qualcomm's U.S. alleged infringing sales, which increases the potential effective royalty rate. See Foty/Greene Mot. 13-14; see also id. Ex. 26 ¶ 171. The Local Patent Rules require that, when a party submits damages contentions, it must identify "its theories of recovery, factual support for those theories, and computations of damages within each category, including . . . reasonable royalty." PATENT L. R. 3-8(a). I agree with Tela that its broad theory of these damages—royalty rates based on Qualcomm's sales—was fairly disclosed and that the data for both theories appeared in the same documents. See Foty/Greene Oppo. 19-20. But the Rule requires "computations of damages." It is not disputed that the manner and result of how Greene now calculates damages was not disclosed.

On this record, Intel was prejudiced. If Tela had disclosed in its damages contentions that it planned to base calculations on U.S. sales, Intel could have taken discovery (including through the deposition of Tela's licensing representative, Neal Carney) on the question that undergirds Greene's analysis: whether Qualcomm and Tela believed that the U.S. sales figures were the more relevant for their negotiations. That is a factual question. Because Greene opines that it is true and goes on to award Tela a higher damages calculation based on it, it would only be fair for it to have been disclosed so that Intel could focus on it. It was not enough that Tela at some point identified the underlying data because that would not reasonably put Intel on notice. Accordingly, I will strike Greene's conclusions to the extent they depend on U.S. sales instead of worldwide sales. I note that Tela itself classifies the U.S. sales-based calculation as an "alternative means of calculating" damages, see Foty/Greene Oppo. 20; it may rely on the worldwide sales-based figures and method it disclosed in its contentions.

2. Per-Unit Calculation

Intel argues that Tela's damages contentions were based, in part, on calculating royalty rates as a percentage of revenue; that is, dividing a licensee's royalty payment by its sales revenue of the licensed products. See Foty/Greene Mot. 16. Greene's report, it asserts, instead calculates royalty rates by taking a fixed royalty payment for each chip (regardless of price). See, e.g., id. Ex. 26 ¶ 175. Tela responds that this latter calculation—the per-unit method—need not have been disclosed because it is a "subsidiary calculation[]" in Greene's report and its previous disclosure of revenue-based calculations was just "exemplary." Foty/Greene Oppo. 21. Tela cannot classify something as "subsidiary" to get away from its failure to disclose. That calculation by Greene, if accepted, would result in a different result on a different basis than the one Tela disclosed.

If there were no prejudice to Intel, Tela's failure to disclose might not result in a successful motion to strike. But here, Intel deposed Carney (Tela's licensing representative) about Tela's revenue-based contentions and, to Intel, elicited answers demonstrating that the revenue-based calculations for royalty rates were litigation-driven, while Tela internally had never regarded the value of its licenses to be based on that calculation. See Foty/Greene Mot. Ex. 30 at 123:24-124:11, 129:2-7. If Tela had timely disclosed its per-unit calculation, Intel could have taken discovery about it—including, at the least, questioning Carney as it did with Tela's disclosed contentions. Additionally, striking these calculations is not unduly prejudicial to Tela as Greene also performs calculations based on revenue and, in fact, Tela itself asserts that the report "primarily relies" on those calculations. Foty/Greene Oppo. 22.¹⁴ I will strike Greene's conclusions regarding per-unit royalty rates.

14 Tela argues vaguely that "Intel questioned Mr. Carney extensively about both percentage of revenue and per-unit royalty rates" and cites, without elaboration, various points in Carney's deposition. Foty/Greene Oppo. 23 (emphasis in original). But that does not mean that Intel was on sufficient notice that Tela would be pursuing this theory and, accordingly, there was no motive to question Carney (or take other evidence) in the same way or to the same degree as it did about the per-unit calculation.

3. Pre-Suit Damages

Intel argues that Tela never disclosed the basis for any pre-suit damages that Greene includes in his report. Foty/Greene Mot. 19-20. Intel at one point served interrogatories that required Tela to "explain all factual and legal bases for any contention that Tela is entitled to pre-suit damages from Intel" including various specific requirements. See Foty/Greene Mot. Ex. 33 at 6. Tela objected on numerous (often boilerplate) grounds, including that it was premature. Id. at 7. But Tela committed to supplementing the request after providing damages contentions. Id. Its supplement only responded (1) "Tela does not contend that its licensees have marked or do mark any products with the Asserted Patents," (2) by incorporating its contentions, (3) and that investigation was ongoing and it still required discovery. Id. Ex. 34 at 8 (emphasis added).

It is revealing that Tela does not simply point to some disclosure that plainly sets out a basis for its liability claim. Tela responds first by citing a January 10, 2019, disclosure that states, "Intel itself alleges that it became aware of an assertion by Tela of . . . the asserted patents . . . in August 2014 when Tela sent Intel a list of certain patents for a licensing discussion." Foty/Greene Oppo. 23; see also id. ("This disclosure, in and of itself, moots Intel's claim."); Foty/Greene Mot. Ex. 34 at 8. While I would not necessarily, as Intel urges, grant this motion to strike based solely on a disclosure appearing in infringement contentions rather than damages contentions, see Foty/Greene Reply 13, I fail to see how this disclosure would put Intel on notice concerning even the basic facts required. Tela does not point to any more explicit such notice than this.¹⁵ Tela also argues that, when I denied Intel's motion to dismiss, I held that Intel had sufficient pre-suit notice of infringement. Even if Tela were wholly correct in interpreting that order—which, it bears mention, was at the motion-to-dismiss stage—it does not mean that Tela disclosed to Intel its theory for why Intel would be subject to pre-suit liability.

15 Tela includes a footnote after the first sentence in its briefing on this point. That sentence states that Intel's argument is "fundamentally misplaced" and the footnote begins, "See, e.g.," and cites, without explanation, roughly two dozen pages of exhibits. See Foty/Greene Oppo. 23 & n.14. It is telling that Tela does not highlight any statement within those exhibits—which are, in any case, witness transcripts, not disclosures—that would put Intel on notice.

This failure to disclose is not simply a technical but harmless violation of Tela's duties. Tela's failure to disclose meant that an entire class of damages discovery could not be taken. The failure to disclose is especially prejudicial in that disputes over pre-suit damages could plausibly depend on substantially different evidence than other types of patent damages. The motion to strike Greene's pre-suit damages opinions is GRANTED. Because I giant the motion on this basis, I do not address the parties' dispute over whether Greene's opinions have adequate factual support.

4. Breach-of-NDA Damages

Intel moves to strike Greene's damages calculation for Tela's breach-of-NDA claim because it is (1) based on "speculative assumptions" that (2) were not adequately disclosed. Foty/Greene Mot. 22.

The assumptions that Greene's opinion depends on could have properly been the subject of discovery. Intel objects to Greene's assumptions that [Redacted] ; that Intel would not have sued without the alleged breach; and that Tela would not have sued absent Intel's suit. Foty/Greene Mot. 22.

Some of these are plainly admissible and did not need to be disclosed during discovery. The statements about suits are simply disputable assumptions that can be subjected to cross-examination and about which the parties likely have sufficient evidence. The fact-dependent assumptions about [Redacted] are more complex. Tela argues that "the basis of the Greene Report's analysis—that Intel's breach inhibited Tela's licensing prospects—was more than sufficiently disclosed." Foty/Greene Oppo 24. But the fact that the general statement that "Intel's breach inhibited Tela's licensing prospects" was disclosed does not mean that any of the fact-bound assumptions Greene relies on were disclosed. If they had been, Intel would have been able to take discovery regarding each of them. Tela does not dispute, for example, that Greene's opinion is that [Redacted] ; without that being disclosed, Intel could not reasonably have anticipated that it would need to conduct discovery about any of the many facts surrounding such a hypothetical [Redacted] . Tela does not dispute that there is no fact in the record related to this hypothetical [Redacted] . Intel's motion is GRANTED IN PART only to the extent it relates to assumptions about the suit frustrating a purported early renewal of the license, not the other assumptions Intel attacks. Greene may testify to the results of his calculation without the assumption that [Redacted] .

ii. Damages Opinions

Intel moves to exclude Greene's testimony under Daubert and related doctrines applicable to expert testimony. "Under settled evidence law, an expert may express an opinion that is based on facts that the expert assumes, but does not know, to be true." Williams v. Illinois, 567 U.S. 50, 57 (2012). Greene's damages opinions are about royalty rate calculations based on, in part, previous licensing negotiations between Tela and Qualcomm, Samsung, and TSMC. Reasonable royalty analyses in particular "necessarily involve[] an element of approximation and uncertainty." Unisplay, S.A. v. Am. Elec. Sign Co., 69 F.3d 512, 517 (Fed. Cir. 1995). These damages will be based on "a hypothetical negotiation between the patentee and the infringer when the infringement began." Id. While no expert can admissibly engage in pure guesswork, courts "have never required absolute precision in this task." Virnetx, Inc. v. Cisco Sys., Inc., 767 F.3d 1308, 1328 (Fed. Cir. 2014).

1. Litigation Risk Adjustment

Greene opines that the royalties paid to three Tela licensees—Qualcomm, Samsung, and TSMC—are lower than what Intel should pay because those other licensees received a [Redacted] "litigation risk" adjustment, leading to a [Redacted] "discount." Intel argues that the assumption that Tela actually provided a discount is speculative and contradicts the testimony of Tela's corporate representative. Greene Mot. 4-9. That argument is for cross-examination, not exclusion under FRE 702 and Daubert.

Intel first asserts that the evidence Greene's "assumption" of the litigation risk adjustment based on shows only that [Redacted] . Although Greene may not rely on "mere speculation or guess," he is permitted to base his opinion on an assumption about what occurred. See, e.g., Williams, 567 U.S. at 57. Intel may seek to argue to the finder of fact—and show through cross-examination—that the assumption is unwarranted, but there is sufficient evidence on this record to support it and render Greene's reliance on it reasonable. Intel also argues that Tela's licensing representative, Carney, contradicted Greene. As an initial matter, this "contradiction" is overstated; it is that the litigation risk adjustment was "one of many factors" that went into the ultimate agreement. See Greene Mot. Ex. 6 at 97:22-98:14. Moreover, Carney testified that a spreadsheet that includes a [Redacted] litigation risk adjustment was used by Tela during negotiations. Greene Oppo. Ex. F at 16:6-20. Intel may probe this issue on cross-examination, but it does not show that Greene used method that is unreliable within the meaning of FRE 702 or Daubert.

Separately, Intel argues that Greene improperly exports the alleged [Redacted] discount regarding Qualcomm to Samsung and TSMC, who also struck licensing deals with Tela that Greene uses as the basis for his opinions. Greene Mot. 7-8. According to Intel, Greene reveals no methodology to guide comparison of Qualcomm's perceived litigation risk with Samsung and TSMC. Id. These opinions are certainly based on a longer string of assumptions than the Qualcomm opinion. Even so, Greene has explained his reasoning, including that many of the same licensing considerations Tela had regarding Qualcomm would cany over to other companies. See Greene Oppo. Ex. C at 212:1-2. Daubert and related doctrines exist to ensure that experts employ reliable methodology; there is sufficient grounding in the record for Greene's opinion to clear that hurdle.

Intel also attacks Greene's comparison of the litigation risk discount to overall litigation success statistics found in a report by PwC. See Greene Mot. 6-7. If Greene were relying on these general statistics to judge the particular litigation risk relevant to this case, that would potentially raise Daubert problems. See Carucel Investments, L.P. v. Novatel Wireless, Inc., No. 16-CV-118-H-KSC, 2017 WL 1215838, at *20-*21 (S.D. Cal. Apr. 3, 2017) (excluding the PwC report as a basis to assess litigation risk in a particular case). Here, however, Greene uses the report to show that Tela and Qualcomm's alleged views of the litigation risk would have been more reasonable. While courts have held that nation- or District-wide statistics should not be taken as the actual predicted rate relevant to the negotiation, see, e.g., Avocent Redmond Corp. v. Rose Elecs., 2013 WL 8844098, at *5 (W.D. Wash. Mar. 11, 2013), Greene uses this case's particular facts to determine the risk adjustment and uses the PwC study to illustrate that the adjustment would have been reasonable to the parties.

2. First Mover Adjustment

Intel objects to Greene's reliance on [Redacted] "first mover" discount that Qualcomm requested. Greene Mot. 9-10. Intel's argument on this point mirrors its litigation-risk argument above: it asserts there is no evidence Qualcomm actually received this discount and Carney testified that it was one factor of many that was considered in the negotiation. For the same reasons discussed above, this is a matter for cross-examination, not evidentiary exclusion.

3. Allocation of Value Among Patents

Because Greene's findings depend on the previous licensing negotiations and those licensing negotiations covered Tela's [Redacted] , Greene's report attempts to allocate value between the [Redacted] and the Asserted Patents in particular. See, e.g., Greene Oppo. Ex. A ¶¶ 205-23. Intel argues that Greene's allocation of value is inadmissible for three reasons. Greene Mot. 11-18.

First, Intel argues that Greene does not employ a reliable methodology (or any methodology) to determine that the Asserted Patents are the "most important" of the "4 Series Family" of which they are a part. In his report, Greene analyzes the relative value of a group of Tela's patents called "4 Series Family" compared to its overall portfolio. See Greene Oppo. Ex. A ¶¶ 222-23. The 4 Series Family is a family of patents within Tela's "layout patents" portfolio. Id. ¶ 207. The 4 Series Family, according to Greene, includes 42% of Tela's U.S. patents. Id. But, in Intel's view, there is no support for jumping from assertions about the value of the 4 Series to the conclusion that the Asserted Patents are the most important of that series. Tela's counterarguments are somewhat wanting in that they largely relate to justifying the value allocated to the 4 Series as a whole. See, e.g., Greene Oppo. 12 ("Tela's patent strategy was to grow the size of a particular patent portfolio that it believed was the strongest and most important technology—i.e., the 4 Series family.").

However, I conclude that this potential flaw goes to weight, not admissibility. Greene asserts that "Even within the 4 Series Family, the Asserted Patents are the most applicable to Intel, and would likely be viewed by Intel as the most valuable. For instance, Intel states that it chose to challenge the '012 Patent and the '966 Patent because they are broader and more relevant to addressing Tela and Intel's dispute." Id. Ex. A ¶ 212. Accordingly, I view this as an underlying assumption, rather than an unsupported guess—even if the underlying evidence is somewhat faltering. Intel objects that even this statement deals with only two patents, but here (1) the patents are closely related and (2) Greene makes clear that he extrapolates from this principle to the Asserted Patents more broadly. Intel may cross-examine about this issue.

Next, Intel objects to Greene's reliance on two papers that address allocation of value among patent portfolios. Greene Mot. 15-16. These papers have been relied on by other experts and courts as one way of guiding allocation of value. See Greene Oppo. 13 (collecting cases). Intel responds that the methodology of the papers is insufficiently tied to the specifics of this case. That is an issue of weight, not exclusion under Daubert. There is evidence that damages experts rely on these papers with some regularity. Although the papers use general data to show how to allocate the value of patents, courts have made clear that patent damages calculations will necessarily include some level of approximation. Here, Greene has chosen a methodology that, even if not a perfect fit, is (on this record) generally regarded as reliable in his field for evaluating allocations. From there, Greene then applies it to this case, as other experts have done in their cases.

Finally, Intel takes issue with Greene's allocation of value among the Asserted Patents. Again, this issue is for cross-examination. Intel's primary concern is that, when Greene apportions the value down to different combinations of specific Asserted Patents, each combination is still given the same value as all of the Asserted Patents combined. Greene Mot. 16-17. The reason, though, is that the patents are largely overlapping, so Greene makes (as his report explains) the relative changes in value to other measures, such as royalty rate and base, but not to the value as a percent of the patent portfolio. Intel is free to question those assumptions, but they have underlying reasoning sufficient to survive this motion.

The motion to exclude these allocation opinions is DENIED.

4. Allegedly Non-Comparable License Agreements

Intel argues that Greene improperly relies on several previous licenses involving Intel with large lump sum payments up to $1.5 billion that are non-comparable to this case. Greene Mot. 18-20. Greene himself classified the licenses as non-comparable but Tela argues that he is not impermissibly using them to inflate the effective royalty rates or unduly influence the jury by placing irrelevant large figures in front of it. Greene Oppo. 16-19. Instead, Tela argues that Greene uses these figures to shed light on Intel's general licensing practices, including its purported willingness to [Redacted] See id. Ex. A. ¶ 261. I agree with Tela that Greene is using these practices to try to demonstrate Intel's "usual" or "historical" approach to licensing. See, e.g., Comcast Cable Commc'ns, LLC v. Sprint Commc'ns Co., LP, 218 F. Supp. 3d 375, 388 (E.D. Pa. 2016). It is unlike Intel's leading case, Lucent Technologies., Inc. v. Gateway, Inc., 580 F.3d 1301 (Fed. Cir. 2009), for several reasons, including that the Federal Circuit there was reviewing a damages award based on sufficiency of the evidence and the non-comparable agreements were proffered as the evidence to support the award. Id. at 1327-32. Here, in contrast, Greene's damages opinions primarily rely on the comparable Qualcomm, Samsung, and TSMC negotiations; the non-comparable agreements are then brought in to further reveal Intel's more general licensing practices.

5. Future Projections

Intel objects to Greene's reliance on projections that Tela created of potential future licensing royalties to calculate damages because Greene did not create them, they were created by Tela and are likely to be self-serving, Greene did not verify or corroborate them, and at least one contradicts Greene's own analysis. See Greene Mot. 22-23. None of Intel's cases require excluding this opinion. They generally disapprove experts being mere conduits for clients' "self-serving" projections without independent analysis. See, e.g., United Energy Trading, LLC v. Pac. Gas & Elec. Co., No. 15-CV02383-RS, 2018 WL 5013580, at *2 (N.D. Cal. Oct. 16, 2018). But here Greene has conducted his own analysis and relies on Tela's projections, in his view, to assess its expectations. Intel is free to cross-examine about this reliance, or about its argument that some of these projections are inconsistent with Greene's own; neither reason requires exclusion today.

iii. Alleged Technical Opinions

Intel also contends that Greene offers improper technical opinions beyond his expertise. Greene Mot. 20-22. To the extent that Greene attempts to offer technical expert opinions as the substantive evidence that they are true at trial, that would be improper. Here, however, Greene merely—as he states in the report—relies on the technical knowledge that has been communicated to him. See, e.g., Greene Oppo. Ex. C at 124:2-23 (discussing the basis of his opinions as including discussions with Foty). Intel responds that some of these opinions are not in Foty's report. But Greene relied on his discussions with Foty, which Intel does not object is improper. Intel may object at trial if any technical opinion is offered as substantive evidence of its truth (as opposed to the underlying understanding of the technology to form damages opinions).

iv. Conclusion Regarding Greene

Intel's motion to strike Greene's testimony for failure to disclose is GRANTED IN PART; it is only DENIED with respect to some breach-of-NDA damages opinions, as described. Intel's motion to exclude DENIED.

C. Tela's Motion to Strike and Exclude Opinions of Dr. Subramanian

Tela moves to strike and exclude various portions of the report and various opinions of Intel's expert, Dr. Vivek Subramanian. See Tela's Motion to Strike and Exclude Regarding Dr. Vivek Subramanian ("Subramanian Mot.") [Dkt. No. 250].

i. Practicing the Prior the Art Defenses

Tela argues that Subramanian's opinions on several topics assert improper "practicing the prior art" defense. In a line of cases, the Federal Circuit has held that accused infringers may not offer a practicing the prior art defense. This defense "typically refers to the situation where an accused infringer compares the accused infringing behavior to the prior art in an attempt to prove that its conduct is either noninfringing or the patent is invalid as anticipated because the accused conduct is simply 'practicing the prior art.'" Cordance Corp. v. Amazon.com, Inc., 658 F.3d 1330, 1337 (Fed. Cir. 2011). The concern is that invalidity must be proven by clear and convincing evidence, while noninfringement need only be proven by a preponderance of the evidence. See id. The difference in burdens "can potentially allow a defendant to skirt evidentiary hurdles and conflate the infringement and invalidity inquiries." 01 Communique Lab., Inc. v. Citrix Sys., Inc., 889 F.3d 735, 742 (Fed. Cir. 2018). "In essence, an accused infringer forsakes any comparison between the asserted claims and the accused product, relying instead upon purported similarities between the accused product and the prior art." Id. "This does not, however, preclude a litigant from arguing that if a claim term must be broadly interpreted to read on an accused device, then this same broad construction will read on the prior art." Id.

1. Invalidity Opinions

Tela's concerns about Subramanian's opinions on invalidity are most easily addressed. The parties, in fact, appear to interpret Subramanian's argument the same way: he argues that if the asserted claims are interpreted broadly enough to apply to the Accused Products, they must also be interpreted broadly enough to apply to the 45 nm products (the alleged prior art). See, e.g., Subramanian Mot. 10; Intel's Opposition to the Subramanian Mot. ("Subramanian Oppo.") [Dkt. No. 250] 4-6. That argument is what the Federal Circuit explicitly approved in Citrix: the prohibition on practicing the prior art defenses "does not . . . preclude a litigant from arguing that if a claim term must be broadly interpreted to read on an accused device, then this same broad construction will read on the prior art." Citrix, 889 F.3d at 742. Indeed, "when an accused product and the prior art are closely aligned, it takes exceptional linguistic dexterity to simultaneously establish infringement and evade invalidity." Id. at 743. None of this removes the burden of proving invalidity by clear and convincing evidence; all that Subramanian and Intel attempt to do with these opinions is show that Tela's interpretation of the asserted claims ropes in prior art.

2. Noninfringement Opinions

Tela argues that several of Subramanian's opinions regarding noninfringement assert practicing the prior art defenses. Tela argues that "Dr. Subramanian compares Intel's 45nm Products to the Accused Products, impermissibly attempting to shift the burden to Tela to prove the validity of its patents along with its infringement proof." Subramanian Mot. 5. Intel responds that Subramanian is merely arguing that certain structures in the 45 nm products are the same as in the Accused Products and those structures do not infringe—apparently because of a lack of direct evidence about the Accused Products. Subramanian Oppo. 7-10. Even if not explicitly a practicing the prior art defense, I agree with Tela that these opinions raise the specter of conflation of burdens of proof that the Federal Circuit has warned of. Subramanian is permitted to testify about the [Redacted] technology at issue. He is not permitted, however, to suggest to the jury that the 45 nm products do not infringe and therefore the Accused Products do not infringe.

Intel claims that many of his opinions are aimed at Tela's argument in the ITC proceedings that the [Redacted] technology is inconsistent with the claim limitations. In other words, Tela argued that its products were not anticipated by the 45 nm products for that reason. It would, Intel says, be inconsistent—because the technologies are purportedly the same—for Tela to argue that there is infringement on this ground. That is a different issue and Tela does not ask me to exclude such evidence. However, I grant Tela's motion to the extent that Subramanian argues that there is no infringement because the Asserted Products share the same structures as the 45 nm products. He may opine about the structures in the Accused Products themselves, but Intel "need not compare the accused products with prior art to make the point that certain qualities or components of a product do not contain all the elements of [the claim]." Opticurrent, LLC v. Power Integrations, Inc., No. 17-CV-03597-WHO, 2018 WL 6727826, at *15 (N.D. Cal. Dec. 21, 2018). My holding may not neatly line up with the paragraphs that Tela seeks to exclude, especially because Tela cites 15 paragraphs without individualized explanations about each. Nevertheless, I trust that, at trial, the parties will be able to ensure that Subramanian stays within the bounds I lay out.

3. Damages Opinions

Subramanian offers certain opinions about the contributions of the art by the Asserted Patents; those opinions are then utilized by Intel's damages expert in forming her opinions. See, e.g., Subramanian Mot. Ex. 2 ¶¶ 730-39. Tela objects that Subramanian makes comparisons between the 45 nm products, the Accused Products, and the Asserted Patents to determine how meaningful the contribution of the Asserted Patents was. To Tela, this comparison again raises the concerns implicit in the practicing the prior art defense. Subramanian Mot. 6-9. Intel responds that Subramanian is merely comparing the prior art (here, allegedly, the 45 nm products) to the asserted claims to determine "how to account for the relative value of the patentee's invention in comparison to the value of the conventional elements recited in the claim, standing alone." AstraZeneca AB v. Apotex Corp., 782 F.3d 1324, 1338 (Fed. Cir. 2015); see Subramanian Oppo. 10-12.

To the extent that comparison is based on the assumption that the Asserted Patents are found valid, see Subramanian Mot. Ex. 2 ¶ 712, it is permissible. But that does not explain the comparison between the 45 nm products and the Accused Products. Subramanian finds, among other things, that "most of the design rules implicated by Tela's infringement accusations were already present in Intel's prior art 45nm process" and "the design rules that were not present in Intel's 45nm process are a natural evolution of the previously existing design rules from the gate layer to metal layers." In this particular context, however, I conclude that this is not an attempt to "forsake[] any comparison between the asserted claims and the accused product, relying instead upon purported similarities between the accused product and the prior art" to show either invalidity or noninfringement. Citrix, 889 F.3d at 742. Instead, Subramanian's analysis accepts, for the sake of argument, that there is infringement and the patents are valid and provides views on (1) the marginal contribution of the Asserted Patents and (2) to the extent to which that contributive, novel aspect of the patents is infringed by the Accused Products.

ii. Enablement

Tela moves to exclude Subramanian's ultimate opinions regarding whether the patents were enabled or not. Subramanian Mot. 12. "Enablement is a question of law based on underlying factual findings." MagSil, 687 F.3d at 1380. And "an expert witness cannot give an opinion as to her legal conclusion, i.e., an opinion on an ultimate issue of law." Hangarter v. Provident Life & Acc. Ins. Co., 373 F.3d 998, 1016 (9th Cir. 2004) (emphasis omitted). Accordingly, Tela argues, Subramanian may not testify to this conclusion. It appears that district courts have come out on both sides of this question (or an analogous one). Compare, e.g., Sonos, Inc. v. D & M Holdings Inc., 297 F. Supp. 3d 501, 513 (D. Del. 2017) (holding experts may opine about ultimate conclusion of enablement), with Erfindergemeinschaft UroPep GbR v. Eli Lilly & Co., No. 2:15-CV-1202-WCB, 2017 U.S. Dist. LEXIS 35052 (E.D. Tex. Mar. 13, 2017) (holding experts may not opine about whether a patent had a sufficient written description).

The proper starting point is FRE 704(a), which provides that "[a]n opinion is not objectionable just because it embraces an ultimate issue." Experts cannot give pure "legal conclusions"—that is, they cannot instruct the jury about the law. See Hangarter, 373 F.3d at 1016; see, e.g., Huawei Techs., Co v. Samsung Elecs. Co., 340 F. Supp. 3d 934, 965 (N.D. Cal. 2018). But enablement is directed toward a person skilled in the art. An expert opining about whether something is enabled might therefore be useful to the jury. It is also not a "legal conclusion" in the sense used in Hangarter because it is not instructing the jury about any legal standard; instead, it is a label for the conclusion that a POSITA would be able to practice the art without undue experimentation from the patents. By the same token, though, an expert merely labelling something enabled is not conclusive. Nor does it introduce a dispute of material fact about enablement because it is a conclusory label, not a genuine underlying fact. The motion is DENIED.

iii. Non-Infringement "Regions" Opinions

In his rebuttal report, Subramanian criticizes Foty's infringement analysis, among other reasons, on the basis of his purported failure to identify any "region" that is fully occupied by gates or includes a circuit—that is, it encompasses a single cell. See, e.g., Subramanian Mot. Ex. 2 ¶ 638. Tela argues that this is effectively a unilateral claim construction and a new noninfringement theory because it has never before been stated that the "region" at issue have these characteristics. Subramanian Mot. at 13-14.

The term has not been construed and the use of "region" in the patents that Tela points to has no qualifications on its face that either supports or undermines Subramanian's theory. On this record, at the motion-to-strike stage, I cannot say that such amorphous terms support Tela's position any more than they support Intel's. Subramanian appears to be arguing just that the regions that Foty identifies do not infringe because they lack the relevant characteristics. See, e.g., Subramanian Mot. Ex. 2 ¶ 636-38. Notably, because there is no claim construction on point, it will be up to the parties to demonstrate whether there is infringement or not independent of any specialized definition. See Asetek Danmark A/S v. CMI USA Inc., 852 F.3d 1352, 1359 (Fed. Cir. 2017). Put another way, the trier of fact will not be instructed that Subramanian's purported requirements are needed by any claim construction, so Tela is free to argue that they are not.¹⁶

16 This portion of the order should not be understood to comment on potential jury instructions, which are not yet ripe for determination.

iv. Supplemental Report

In June 2020, Tela deposed Mashhood Rassam, an Intel employee, about Intel's selection of which patents it brought suit about. Foty submitted a supplemental report responding to that testimony. I permitted Intel leave to then submit a responsive report from Subramanian. See Dkt. No. 233. That report is limited to "directly respond[ing] to Foty's analysis into the technical underpinnings of Intel's reasons for selecting the patents-in-suit rather than other Tela patents." Id. at 3. I further stated that, "To the extent that Foty's follow-up report undermines the technical accuracy of that testimony, it is appropriate for Subramanian to respond to these limited opinions. The same is true regarding the consistency between Intel's response to Interrogatory Number 27 [about patent selection] and Rassam's testimony, provided that Subramanian's opinions address only the technical facts of Intel's proffered criteria for selection." Id. at 2-3.

Tela argues that various parts of Subramanian's report (Dkt. No. 247-31) fall outside of the scope I permitted. I briefly address those paragraphs I will not strike followed by those I will.

The following paragraphs are within the scope I set. Paragraphs 4 and 5 respond to Foty's analysis that some selected patents did not qualify for technical reasons; Subramanian's view is that those technical aspects did in fact qualify under the interrogatory. Paragraph 8 concerns what Subramanian perceives to be an error in Foty's analysis regarding the import of certain technical aspects of the patents. Paragraph 14 disagrees with Foty's conclusion that an alternate patent should have been selected, and gives the technical reason for that conclusion. Paragraphs 29-30 are liminal but, ultimately, within the scope I set. Subramanian does inch up to the line by stating that Rassam's review of the '098 patent was "quick" and giving his own that is more detailed; I conclude that opinion is sufficiently geared toward rebutting Foty's report about Rassam on technical grounds.

The following paragraphs are not within the scope I set and I grant Tela's motion with respect to them. Paragraph 9 does not appear to respond to any particular alleged flaw in Foty's analysis but, instead, advances a new argument; Intel defends it only on the ground that would apply to paragraph 8, but not to it. See Subramanian Oppo. 21. Paragraphs 19, 22, 25, and 27 do not concern any technical opinions, they concern Intel's "buckets." Intel says it selected patents from various "buckets"—those asserted in the ITC proceeding, patents pending as of April 2015, and the remaining patents in the proper family. See Subramanian Mot. Ex. 14 at 13-15. These paragraphs of Subramanian's report simply critique Foty for purported mistakes in determining which bucket a patent came from, which is not "address[ing] only the technical facts of Intel's proffered criteria for selection."

v. Conclusion Regarding Subramanian

The motion to strike is GRANTED IN PART with respect to certain paragraphs of the supplemental report; it is otherwise DENIED. The motion to exclude is GRANTED IN PART to the extent that Subramanian argues that the Accused Products do not infringe because they are identical to the 45 nm products; it is otherwise DENIED.

D. Tela's Motion to Strike and Exclude Opinions of Ms. Davis

Tela moves to strike and exclude various opinions of Intel's damages expert, Julie Davis. See Tela's Motion to Strike and Exclude Regarding Julie L. Davis ("Davis Mot.") [Dkt. No. 249]. The standards for disclosure under the patent rules, general principles of the admissibility of expert opinions, and apportionment law are discussed above.

i. [Redacted] Opinions

As discussed above, the parties base their damages calculations on past licenses and negotiations between Tela and Qualcomm, Samsung, and TSMC. Davis bases some of her damages calculations on, among other things, a [Redacted] provision agreed to by [Redacted] . See, e.g., Intel's Opposition to Davis Mot. ("Davis Oppo.") [Dkt. No. 256] Ex. 2 at 23-24. That provision [Redacted] . Id. Tela objects to Davis's use of the [Redacted] provision. It argues that the [Redacted] provision is not probative of the actual royalty rate between Qualcomrn and Tela. Davis Mot. 7. Intel responds that the provision is probative of royalty rates because it shows what Tela itself thought of the value of the royalty rates it should receive. Davis Oppo. 3-8. Additionally, it effectively argues that the best way of understanding how the lump-sum was calculated is using the rate related to the [Redacted] provision. Id.

I agree with Intel that the use of the [Redacted] provision is admissible. As should be clear from the preceding discussion about Tela's damages expert, the proper way to determine the past agreements' effective royalty rates is a deeply contested issue. Just as Tela's expert argues that several considerations (such as the litigation risk adjustments) are implicitly incorporated into the effective royalty rates, Davis argues that the [Redacted] provision would be. In both cases, the parties disagree over whether there is sufficient evidence tying the rate to that basis. In both cases, I conclude that—whatever the ultimate merits—the experts' views are at least relevant and not unreliable under Daubert. This case is unlike Tela's leading case to the contrary, MLC Intellectual Property., LLC v. Micron Technology, Inc., No. 14-CV-03657-SI, 2019 WL 2863585 (N.D. Cal. July 2, 2019). Here, unlike there, the experts' primary task is to determine what the effective royalty rate should be based on the past lump sums; in MLC, in contrast, the court excluded reliance on a most favored customer provision precisely because it found that the testimony that lump-sum "licenses 'reflect' a 0.25% royalty rate is speculative and not based on the facts of the actual licenses." Id., at *13.

ii. Research and Development Opinions ¹⁷

17 This analysis applies to Tela's argument about R&D spending in general and related to design rules in particular.

Davis bases her determination of how important the technology of the patents is in the Accused Products, in part, on analyzing Intel's investments in research and development ("R&D"). Tela argues that (1) the theory was not adequately disclosed and (2) it should be excluded as irrelevant and unreliable. See Davis Mot. 9-15.

First, this theory was adequately disclosed. Intel's rebuttal damages contentions state: "One approach for allocating value is to use Intel's decisions regarding its relative investments in developing each of these categories of technology (the former of which are completely distinct and separate from alleged infringement) as a measure of the relative significance and value of these categories to Intel's products and customers." Davis Oppo. Ex. 8 at 22. It also stated, "One approach for allocating value (to design rules generally without limitation to alleged infringement) is to use Intel's decisions regarding its relative investments in developing design rules as opposed to other aspects of fabrication process technology as a measure of the relative significance and value of these respective items to Intel's products and customers." Id. at 23. Tela classifies these statements as "vague[]," Tela's Reply in Support of Davis Mot. ("Davis Reply") [Dkt. No. 278], but they disclose that Intel might allocate value based on the relative investment in development on the allegedly infringing technology versus the rest of the technology.

Tela replies that "'[i]nvestments' are not necessarily 'R&D,' and there is no allusion to R&D anywhere in Intel's Rebuttal Damages Contentions." Davis Mot. 10. But the words that follow "investments" are "in developing each of these categories of technology" and "in developing design rules," so Tela's argument is unconvincing. Tela also argues that Intel does not "cite to a single document or specific fact about what these ambiguous 'investments' are." Id. Especially as there is no evidence Tela requested more specificity after receiving these disclosures, it does not make sense to hold a party to that standard. Moreover, if I were to find for Tela on this point, some of its own disclosure arguments, discussed above, would likely fall too. Tela states that if I employ the "lenient" standard Intel seeks, I should do the same for Greene's opinions. Davis Reply 14 n.12. I have employed the same standard across all of the parties' numerous disputes over disclosure. Notably, Intel's disclosure put Tela on adequate notice to conduct discovery into Intel's relative investments; that is the key difference between this disclosure and several of Tela's failures to disclose above.

Second, Tela argues that the analysis is unreliable for several reasons. Tela first argues that "the R&D apportionment methodology arbitrarily assumes that every R&D dollar is equal, and that every R&D dollar imparts equal value to the ultimate product's functions." Davis Mot. 11. I agree that using R&D dollars as a proxy for how much Intel valued a technology has certain inherent drawbacks. Namely, it is of course possible that some high value technologies will nonetheless require low R&D spending and vice versa. But these analyses "necessarily involve[] an element of approximation and uncertainty." Unisplay, 69 F.3d at 517. That is especially true where, as here, experts are attempting to reverse engineer royalty rates from past negotiations that did not use them. Davis has represented that hers is a reasonable method and the record shows there is some alignment between perceived value and R&D. Tela is free to attack this assumption on cross-examination.¹⁸

18 Relatedly, Tela argues that the R&D data itself is unreliable because it is forecasts rather than "actuals." Davis Mot. 12. First, expected (i.e., forecast) R&D spending would likely reveal much about perceived value, because it still requires a company to allocate dollars to technologies. Second, this is an issue for cross-examination; it does not show that the data is unreliable.

Tela also contends that Davis "simply assumes, without any evidence, that Tela and Intel would have negotiated using Intel's R&D expenses." Mot. 13. In Riles v. Shell Exploration and Production Co., the court remanded a reasonable royalty finding because the expert's "models ignored Riles's established licensing practice." 298 F.3d 1302, 1313 (Fed. Cir. 2002). "A reasonable royalty determination for purposes of making a damages evaluation must relate to the time infringement occurred, and not be an after-the-fact assessment." Id. Accordingly, the court relied on and quoted cases holding that the "patentee's prior license agreements should carry considerable weight in calculating a reasonable royalty rate." See id. (internal quotation marks omitted). There, the record "suggest[ed] that Riles's past licenses based the royalty rate on a percentage of any savings that the licensee would realize from use of the patent or upon a fee based on platform depth, and not on a percentage of the cost of the platform or the revenue for first year or both." Id. Here, Tela argues, there is no evidence that Intel has ever relied on R&D allocation in striking past deals. Tela is correct that Intel has pointed to no evidence it ever considered R&D spending in striking licensing deals. However, the difference between this case and Riles is that the expert here is not attempting to divine the entire effective royalty rate on the basis of considerations that the company has never employed. Instead, she uses this evidence to determine the apportionment aspect of the products, which was not relevant to past negotiations.

Tela next argues that Davis effectively asserts a practicing the prior art defense by comparing the Accused Products with the prior art—again the 45 nm products. For identical reasons to those I gave regarding Subramanian's damages opinions, I disagree.

Last, Tela argues that "the Davis Reports' R&D apportionment methodology is simply unsupported and not tied to the facts of the case." Davis Mot. 14. Tela gives several alleged examples: (1) Davis makes an "unsupported statement that 'the accused technology falls within the larger category of manufacturing technology, when presumably the technology relates to a category Ms. Davis excludes, circuit design"; (2) she "rel[ies] on R&D expenses 'for the years 2010 through 2019' with no explanation for this data set"; and (3) her "methodology fails to link the analysis to the technology at issue" by treating all three types of Accused Product as well as the 35 and 45 nm products the same. On the first, Davis appears to have relied on Subramanian's classification, as is proper for a damages expert. See Davis Oppo. Ex. 2 at 33. The second issue goes to weight, because Davis explains that this is the best data that was accessible. Id. at 41 n.199. The third likewise does not appear to be improper because Davis did the calculations for all product types and chose the most favorable to Tela for simplicity; Tela can, again, cross-examine about this.

iii. Miscellaneous Alleged Failures to Disclose

Finally, Tela asserts that Intel moved to strike "subsidiary damages calculations" by Greene for not being adequately disclosed and that, if I agree, "the issue should be treated consistently for both sides" and that "to the extent the Court agrees with Intel that every possible subsidiary calculation and analysis that a damages expert may do must be disclosed, the Court should also strike" a series of Davis's opinions. Davis Mot. 20-21. Tela misses the mark. No matter what label (such as "subsidiary damages calculation") is attached to a theory, damages contentions must be adequately disclosed. Above, I explained why certain of Greene's theories were not sufficient disclosed. I will evaluate Tela's cursory, miscellaneous requests to strike for failure to disclose, but the touchstone is whether each theory needed to be and was disclosed prior to being advanced in an expert report.

First, Davis calculates different effective royalty rates that Intel disclosed in its rebuttal contentions. Tela fails to state what overall theory was not disclosed; it points in a conclusory way only to differences in final figures. Second, Tela argues that Intel failed to disclose any specific royalty structure, while Davis says it is structured as a lump-sum payment. Yet Intel disclosed that it "may present analysis of alleged damages in the form of a lump sum payment amount" and disclosed the formula it would use to do so. See Davis Oppo. Ex. 8 at 17. Third, Tela argues that Davis calculates the maximum profits attributed to the accused technology but Intel never disclosed calculation or apportionment of profits. However, Intel disclosed its approach to apportionment and determining a profit-based royalty rate in its rebuttal damages contentions. See id. at 22-24. Last, Tela argues that Davis calculates the value of patents based on some Intel argues are covered by the CNTS; Tela contends that "Intel's Rebuttal Contentions, however, do not even mention the CNTS." In response, Intel argues that this is because that portion of the Davis report responds to Greene's findings about the CNTS, raised for the first time in his report. In its Reply, Tela does not address this. These were all adequately disclosed.

iv. Conclusion Regarding Davis

Tela's motion to strike and exclude Davis's report and testimony is DENIED.

E. Tela's Motion to Strike and Exclude Opinions of Dr. Bravman

Tela challenges the report and testimony of Dr. John C. Bravman, one of Intel's experts. See Tela's Motion to Strike and Exclude Regarding John C. Bravman, Ph.D. ("Bravman Mot.") [Dkt. No. 248]. Tela argues that Bravman is not an expert in the first place and so should not be permitted to testify; in the alternative, it requests that certain of his opinions be excluded.

i. Expertise

To give expert opinions, a witness must be "qualified as an expert by knowledge, skill, experience, training, or education." FED. R. EVID. 702. Intel has met its burden to show that, on the record before me, Bravman is qualified as an expert in semiconductor layouts. Bravman has a Bachelor of Science in materials science and engineering, a Master of Science, and a Ph.D. in which his dissertation focused on "the nature of silicon." See Intel's Opposition to Bravman Mot. ("Bravman Oppo.") [Dkt. No. 254] Ex. 1 [Dkt. No. 253-3] ¶ 4. He worked at a semiconductor company for approximately five years during graduate school. Id. ¶ 5. He taught at Stanford University in materials science and engineering from 1985 to 2010. Id. He has taught courses in materials science and engineering, including in integrated circuit materials and processing. Id. ¶ 11. He stated at his deposition that he had personally made semiconductor layouts and had supervised graduate students who did so. Bravman Mot. Ex. 2 at 32:5-33:8. Tela, moreover, stipulated in the ITC proceeding that Bravman was an expert "the field of microprocessors and semiconductor process technology." Bravman Oppo. Ex. 3 at 1042:15-19. Tela now argues that "Bravman's testimony in the ITC was limited to Intel's history and the time it would take for Intel's competitors to replace products excluded from the United States." Tela's Reply in Support of the Bravman Mot. ("Bravman Reply") [Dkt. No. 277]. But that does not change the fact that it recognized Bravman as an expert in the field generally, not just for some particular purpose.

Tela offers two reasons that Bravman is not an expert. First, it argues that he "testified that he does not consider himself to be an expert in semiconductor layout design and has not personally made any semiconductor layouts since he was a student in the 1980s and 1990s." Bravman Mot. 7. The portion of his testimony from the ITC, however, is not as clear as Tela portrays it. In response to whether he considered himself an expert "on circuit design," Bravman replied, "I understand much of circuit design, but I'm not—I would not say I'm a circuit designer." Bravman Oppo. Ex. 8 [Dkt. No. 254-19] 31:14-19. Then, in response to whether he considered himself an expert on "semiconductor layouts," Bravman replied, "I would give the same type of answer. I've dealt with and have contributed to semiconductor layouts for noncommercial environments for research purposes, but my expertise is focused as answered previously." Id. at 31:20-32:4. While this testimony may prove useful in a cross-examination, I find it goes more to the weight of Bravman's opinions than their admissibility. These—at best for Tela, muddled—admissions do not overcome the other evidence of Bravman's expertise. Neither is a clear admission that he is not an expert, merely that his work has been "focused" on other things. That does not mean, however, that he is unqualified to offer these expert opinions.

Tela also argues that Bravman was "unable to answer basic questions" about Pileggi's semiconductor layouts. Bravman Mot. 7. It points only to one example, however: his inability to identify which features on a layout were poly and which were metal. See id. Ex. 3 at 72:3-18. This evidence is not as slight as Intel would have it appear because that layout is actually included in Bravman's report, see id. Ex. 1 ¶ 27, potentially raising questions about his analysis. Nonetheless, Tela is only able to point to a lone example in a deposition in which Bravman discussed many relevant topics at length. This single moment goes, at most, to the weight of his testimony.

ii. Inventorship

Tela moves to exclude Bravman's opinions on the "ultimate issue of inventorship." As explained above, "an expert witness cannot give an opinion as to her legal conclusion, i.e., an opinion on an ultimate issue of law." Hangarter, 373 F.3d at 1016 (emphasis omitted). The standard for inventorship is plainly not for an expert to instruct the jury about, and any opinions regarding that are excluded. And inventorship itself is a question of law. Shum v. Intel Corp., 633 F.3d 1067, 1083 (Fed. Cir. 2010). Inventorship is different than enablement (as noted, I find that an expert can testify about that ultimate conclusion) because whether something is enabled is directed at what experts would understand, so they are not improperly treading on the role of the judge or jury. There is no similar role for them here. Intel's cases establish only the principle that experts may offer opinions related to inventorship—as they can about any appropriate issue—not that they may testify to the ultimate inventorship conclusion.

Intel's other counterargument is that Tela's request is overly broad and should, if granted, only apply to the ultimate conclusion of inventorship, not analysis, opinions, or facts underlying that conclusion. I agree, and I interpret Tela to agree as well. Tela, however, appears to have a more expansive view of what constitutes the "ultimate legal conclusion" than I do. Tela's motion is granted only to the extent that it applies to Bravman's opinions or testimony about whether Pileggi should have been named as an inventor on the patents, or qualifies as an inventor under the legal definition. Bravman's opinions and testimony are otherwise proper (on this ground), including with respect to Pileggi's alleged contributions to the technology and how substantial they were. To this narrow extent, Tela's motion is GRANTED.¹⁹

19 Foty also opines about the ultimate inventorship issue. Under this standard, his opinions would be treated the same.

iii. Alleged Non-Expert Opinions

The definitions of lay and expert opinions are mutually exclusive. Lay opinions are, among other things, "not based on scientific, technical, or other specialized knowledge within the scope of Rule 702." FED. R. EVID. 701(c). Expert opinions must be based on "the expert's scientific, technical, or other specialized knowledge." FED. R. EVID. 701(a). Tela argues that much of Bravman's report is not based on scientific, technical, or other specialized knowledge but, instead, purported opinions "on the occurrence or non-occurrence of facts for which he has no personal knowledge and for which no technical expertise could possibly be relevant." Bravman Mot. 10.

While Tela seeks to strike numerous paragraphs, its fundamental reason for doing so is consistent—as is the flaw in it. Tela's basic theory is that Bravman is asserting various facts that should be established by percipient witnesses, not by him. Additionally, Tela is concerned that he accepts these facts as true, and may so assert. If Bravman testifies about these underlying facts he would, as with any such evidence, have to meet the strictures of Rule 703 and Williams v. Illinois, 567 U.S. 50 (2012). The principles embodied in those authorities ensure that experts are not mere conduits for otherwise inadmissible evidence, or factual evidence that should be but was not proven by percipient witnesses. All of Tela's examples in its motion, however, are things fact witnesses have said that form the basis of Bravman's opinion. As an expert, he is permitted to form his opinions based on facts told to him, so long as he otherwise adheres to the rules regarding expert opinions. That he takes so many factual statements from witnesses to form his opinions, as Intel argues, is a matter for cross-examination. See Fed. R. Evid. 705. On this understanding, the motion is DENIED. Tela should object at trial if FRE 702, 703, and 705 or Williams are violated.

iv. Conclusion Regarding Bravman

Tela's motion is GRANTED IN PART with respect to the ultimate issue of inventorship; it is otherwise DENIED.

IV. MOTIONS TO SEAL

The parties have filed more than twenty administrative motions to seal in connection with the motions here.

Courts "start with a strong presumption in favor of access to court records." Foltz v. State Farm Mut. Auto. Ins. Co., 331 F.3d 1122, 1135 (9th Cir. 2003). The public possesses a right to inspect public records, including judicial records. Ctr. for Auto Safety v. Chrysler Grp., LLC, 809 F.3d 1092, 1096 (9th Cir. 2016). Accordingly, when a party seeks to seal judicial records connected to motions—such as the ones at issue here—that are "more than tangentially related to the underlying cause of action," it "must demonstrate that there are 'compelling reasons' to do so." Id. at 1096-99. "When ruling on a motion to seal court records, the district court must balance the competing interests of the public and the party seeking to seal judicial records." In re Midland Nat. Life Ins. Co. Annuity Sales Practices Litig., 686 F.3d 1115, 1119 (9th Cir. 2012). This district's local rules require that requests to seal be "narrowly tailored to seek sealing only of sealable material." CIV. L. R. 79-5(b).

The numerous motions to seal here are overbroad and insufficiently justified. Rather than denying them with prejudice, I will allow the parties an opportunity to protect genuinely confidential information from disclosure. In submitting a renewed motion—the parameters for which are discussed below—the parties must adhere to the following guidance.

First, I remind the parties that, when sealing information under the compelling-reasons standard, the mere fact that the information was subject to protective order is insufficient. See Civ. L. R. 79-5(d)(1)(A). Any renewed motion to seal must identify with specificity the compelling reasons that the information should not be publicly disclosed. Second, the parties' requests must be as narrowly tailored as possible. I will not seal whole documents unless there are specific compelling reasons to do so. Instead, the parties' proposed redactions must apply to as little of the document as possible to prevent damaging disclosures. Third, the required declaration submitted in support of the request cannot come from counsel in this case. The party seeking to seal must instead submit a declaration from someone within the company who can competently assert the harm that would result from disclosure; such a person should be familiar with the confidentiality practices surrounding that information. A declaration that harm will occur must be specific to each piece of information and avoid conclusory assertions. Fourth, I will not seal anything that has already been publicly disclosed, including at the hearing on these motions, or is publicly available.

With those principles in mind, I address several categories of information more specifically. I will not seal information in other documents that is unredacted in this order. This includes correspondence between various people referenced in the inventorship portion of my order and portions of the CNTS. The most plausibly sealable information discussed in this order relates to the technical details of Intel's products. That information, however, is also central to the dispute. To prevent disclosure, Intel would therefore have to make an extraordinarily strong showing of likely harm. Intel's conclusory justifications such as "highly confidential trade secret information" are far from sufficient. See, e.g. Dkt. Nos. 217, 217-1. As discussed above, that assertion is also not specific to each piece of information sought to be sealed. Additionally, it seems possible that certain details about Tela's past business deals with other companies might be sealable, but I have similar concerns to those just explained. Finally, Intel's internal business practices are referenced in the briefs and this order. I will not seal business practices merely because a company does not publicly disclose them; Intel must illustrate what particularized harm would come from disclosure.

The parties must submit a renewed joint motion within 30 days. That joint motion must include a list of the filings the parties no longer seek to maintain under seal or have redacted. If there are any filings that meet the compelling-reasons standard that either party wishes to remain sealed entirely or redacted, the parties shall include a table that identifies (1) the docket entry of the request, (2) the specific redactions requested, (3) the party that originally submitted the document, (4) the general subject matter of the sealing request, and (5) a citation to the declaration providing competent reasons for the request. I recognize that there are numerous filings to address, but the public's resources and authority have been brought to bear to settle this dispute and the public has a weighty interest in understanding judicial decisions to the greatest extent possible.

I am filing this order with some information redacted to give the parties a chance to justify it; those redactions should not be taken as an indication about the merits of sealing. I will issue an entirely unsealed version of this order if no timely renewed joint motion is made. Alternatively, I will issue an updated order once I rule on the renewed motion. If there is anything in this order that a party wishes to remain redacted, it should specifically address it in the motion and support it by competent declaration.

CONCLUSION

Intel's motion for summary judgment is GRANTED. Tela's motion for summary judgment is GRANTED IN PART and DENIED IN PART as described. The motions to strike and exclude expert reports and testimony are resolved as explained above. Accordingly, Intel is entitled to declaratory judgment that (1) its Accused Products do not infringe the asserted claims as laid out in this Order and (2) the '334 and '335 Patents are invalid.

Finally, it appears that this case may proceed to trial on February 25, 2021 as long as COVID-19 conditions in the Bay Area counties allow for it. I will have further information at the Pretrial Conference.

IT IS SO ORDERED. Dated: December 22, 2020

/s/_________

William H. Orrick

United States District Judge