From Casetext: Smarter Legal Research

ICON Health & Fitness, Inc. v. Peloton Interactive, Inc.

Feb 22, 2021
Civil Action No. 20-cv-1386-RGA (D. Del. Feb. 22, 2021)


Civil Action No. 20-cv-1386-RGA




Before me is ICON's motion for a preliminary injunction. (D.I. 7). The motion is fully briefed. (D.I. 9, 43, 73, 89). I heard remote oral argument on this motion on December 21, 2020. (D.I. 97). For the reasons set forth below, ICON's motion is DENIED.


The parties are both players in the at-home fitness equipment market. ICON manufactures a line of remote-controlled fitness equipment under its iFit brand (including exercise bikes and treadmills). (D.I. 9 at 2). Peloton sells a narrower range of products, primarily focusing on the Peloton Bike and the newly released Bike+. (Id. at 4; D.I. 43 at 3). Both companies maintain subscription-based fitness programs in addition to selling fitness equipment. (D.I. 9 at 3).

ICON asserts that Peloton's new Bike+ with Auto Follow infringes its patented remote-control technology recited in U.S. Patent No. 7,166,062 (the "062 Patent"). (Id. at 1). Specifically, ICON alleges that Peloton's Bike+ infringes claim 47, which depends from claim 29. (Id.).

Claim 29. An exercise device configured to enable interaction of a user, the exercise device comprising
(a) an exercise mechanism comprising a movable element for movement in performance of exercise by a user, the exercise mechanism having one or more operating parameters;
(b) at least one user interface device, communicating with the exercise mechanism, the at least one interface device gathering a first signal from the user;
(c) a communicating mechanism, communicating with the user interface device, the communicating mechanism receiving a packetized second signal from a remote source over a network, the packetized second signal including one or more packetized control signals; and
(d) a controller, responsive to the packetized second signal, configured to control the operating parameters of the exercise mechanism.

Claim 47. An exercise device as recited in claim 29, wherein the exercise device further comprises a scaling control, the scaling control being configured to enable a user to select a value representative of the proportional change to be made to the packetized control signal received by the communicating means.
'062 Patent 51:27-43 (claim 29), 52:45-50 (claim 47) (emphasis added).

I presume the "communicating means" of claim 47 is referring to the "communicating mechanism" of claim 29.

ICON requests that this court preliminarily enjoin Peloton from selling the Bike+ with features that infringe on the '062 Patent. (D.I. 9 at 20).


"The decision whether to enter a preliminary injunction is committed to the sound discretion of the trial court." Duraco Prods., Inc. v. Joy Plastic Enters., Ltd., 40 F.3d 1431, 1437 (3d Cir. 1994) (quoting Merchant & Evans, Inc. v. Roosevelt Bldg. Prods. Co., 963 F.2d 628, 633 (3d Cir. 1992)). When seeking a preliminary injunction, the movant must establish: "(A) they are likely to succeed on the merits of their claims, (B) they are likely to suffer irreparable harm without relief, (C) the balance of harms favors them, and (D) relief is in the public interest." Issa v. Sch. Dist. of Lancaster, 847 F.3d 121, 131 (3d Cir. 2017). The Third Circuit has cautioned that a preliminary injunction is "an extraordinary remedy" to be granted "only in limited circumstances." Novartis Consumer Health, Inc. v. Johnson & Johnson-Merck Consumer Pharm. Co., 290 F.3d 578, 586 (3d Cir. 2002) (quoting Instant Air Freight Co. v. C.F. Air Freight, Inc., 882 F.2d 797, 800 (3d Cir. 1989)).


A. Likelihood of Success on the Merits

When seeking a preliminary injunction in an infringement suit, the patentee "must show that it will likely prove infringement, and that it will likely withstand challenges, if any, to the validity of the patent." Titan Tire Corp. v. Case New Holland, Inc., 566 F.3d 1372, 1376 (Fed. Cir. 2009). With respect to infringement, the court adheres to the same two-step analysis used at other stages of the case. See Waters Corp. v. Agilent Techs, Inc., 410 F. Supp. 3d 702, 708 (D. Del. 2019). First, the court determines the asserted claim scope., Inc. v., Inc., 239 F.3d 1343, 1351 (Fed. Cir. 2001). Second, "the properly construed claim is compared with the accused device to determine whether all of the claim limitations are present either literally or by a substantial equivalent." Id . When examining invalidity, the court will compare the asserted claims, as construed, to the prior art. Waters Corp., 410 F. Supp. 3d at 708.

A defendant may succeed in defeating a patentee's motion for a preliminary injunction if it raises a "substantial question" as to infringement or invalidity of the patent-in-suit., Inc., 239 F.3d at 1350. A "substantial question" means that the defendant "asserts an infringement or invalidity defense that the patentee cannot prove 'lacks substantial merit.'" Id . at 1350-51 (quoting Genentech, Inc. v. Novo Nordisk A/S, 108 F.3d 1361, 1364 (Fed. Cir. 1997)).

ICON argues that Peloton's Bike+ literally infringes claim 47 of the '062 Patent. (D.I. 9 at 8-9). However, ICON offers only this conclusory assertion and no analysis; instead, it cites to forty pages containing 102 paragraphs of an expert declaration in support. (Id. at 9).

In other words, ICON effectively extended the page limits for its brief by 200% without any permission. Under the circumstances, I will decide this motion on the merits, but Peloton should feel free to remind me of this act at some later point in the case if it seems then appropriate to do so.

In response, Peloton asserts three noninfringement arguments. (D.I. 43 at 7-8; see also D.I. 89). First, the Bike+ does not infringe claim 29 because it does not receive a "packetized second signal from a remote source over a network." (D.I. 43 at 7; '062 Patent 51:38-39). Second, and closely related to the first argument, the Bike+ cannot infringe claim 29 because it is not remotely controlled. (D.I. 43 at 7; D.I. 89 at 2). Third, the Bike+ does not infringe claim 47 because it lacks the "scaling control" limitation. (D.I. 43 at 8).

The briefing offers little argument as to the scope of the asserted claims. Nevertheless, the crux of the dispute on infringement, at least with respect to Peloton's two claim 29 noninfringement arguments, is whether claim 29 encompasses local control of the exercise device following the download of packetized information from a remote source. ICON argues that claim 29 (and therefore dependent claim 47) do not require the control signal to be sent in "real time" because this is encompassed by dependent claim 39, which contains a "real-time" limitation. (D.I. 73 at 2-3; '062 Patent 52:15-16). ICON's argument appears to equate remote control of the exercise device with "real-time" control. Peloton argues that claim 29 requires remote control and that ICON disclaimed the possibility of local control when amending claim 29 to distinguish the Hickman Patent (U.S. Patent No. 6,749,537). (D.I. 43 at 9; D.I. 97 at 30:20-22).

There is no dispute that the Target Metrics—Peloton's recommended resistance and cadence ranges—are downloaded to the Bike+ from a remote source. (D.I. 43 at 3; D.I. 73 at 2).

Dr. Easttom's initial declaration stated that the Bike+'s resistance is controlled by Peloton's server in real-time. (D.I. 10 at ¶ 83). In his supplemental declaration. Dr. Easttom clarified that while Peloton's video streams in "real-time," Auto Follow occurs in "real-time" only "from the perspective of the user." (D.I. 74 at ¶¶ 24, 68).

On this record, I see no reason to depart from the plain and ordinary meaning of the claim language. See Aventis Pharm. Inc. v. Amino Chem. Ltd., 715 F.3d 1363, 1373 (Fed. Cir. 2013) ("There is a heavy presumption that claim terms are to be given their ordinary and customary meaning."). Claim 29 recites "the communicating mechanism receiving a packetized second signal from a remote source over a network, the packetized second signal including one or more packetized control signals." '062 Patent 51:37-40. A "packetized" signal as required by claim 29 is one that has been transmitted via a network. '062 Patent 16:42-46. The parties do not dispute this definition. (D.I. 10 at ¶ 102; D.I. 43 at 7). A "control signal" is a signal that "control[s] the operating parameters of [the device], such as speed, inclination, resistance, and the like." '062 Patent 7:66-67, 8:1. The parties also do not appear to dispute this understanding. (See, e.g., D.I. 10 at ¶ 37; D.I. 44 at ¶ 54).

I do not believe that adopting these preliminary constructions is inconsistent with the portions of the prosecution history cited by Peloton. The above-cited language in claim 29 was added to overcome a rejection for anticipation by the Hickman Patent. (D.I. 49, Ex. 9 at 116). The accompanying remarks explain that Hickman "neither teaches nor suggests that a packetized control signal is received from a remote source, as claimed, or even that any control signal is packetized." (Id. at 119). Instead, the control signals in Hickman are "the signal from the local computer to the exercise device." (Id.). I agree that a control signal originating on a local computer is not encompassed by the claim language as construed above.

I next consider the accused product. The parties offer very different analyses of how Auto Follow on the Bike+ works. ICON's expert, Dr. Easttom, asserts that when using Auto Follow, the Bike+ receives control signals via the user's network connection that "deliver a packetized control signal for scaling control of the [] Bike+." (D.I. 10 at ¶ 100). Dr. Easttom came to this conclusion by analyzing network traffic to and from the Bike+ during a workout using Auto Follow. (Id. at ¶¶ 90-101). Peloton argues that Dr. Easttom's conclusion is a misinterpretation of encrypted data. (D.I. 43 at 6). Instead, Peloton asserts that the Target Metrics that control the Bike+'s resistance are downloaded before a class starts and that the Auto Follow program is executed locally. (Id. at 6 (citing D.I. 45 at ¶¶ 6-8)). The metrics are downloaded in the form of JSON information. (D.I. 44 at ¶ 72; D.I. 74 at ¶ 28).

Dr. Easttom describes himself as an "independent" expert (D.I. 10 at ¶ 1) but he sounds like anything but independent as he offers the opinion that the testimony of Peloton's expert and employee is "false" and that Peloton's expert "feigned confusion." (D.I. 74 at ¶76).

Peloton states that the network traffic observed by Dr. Easttom was the result of streaming video to the Bike+ and the "exchange of analytics data." (D.I. 43 at 6 n.4).

In reply, ICON changes tack and argues that the JSON information downloaded before a Peloton class begins can satisfy the "packetized control signal" limitation in claim 29. (D.I. 73 at 2). The JSON information consists of resistance ranges and timestamps. (D.I. 74 at ¶ 29). In his supplemental declaration, Dr. Easttom concludes that this information constitutes a packetized control signal. (Id. at ¶¶ 29-32). At oral argument, however, Peloton noted that citing JSON information to satisfy the packetized control signal limitation was inconsistent with Dr. Easttom's conclusion on invalidity that raw data transmitted from a remote source used in a local calculation could not constitute a "packetized control signal" within the meaning of claim 29. (D.I. 97 at 32:15-25, 33:1-3; D.I. 74 at ¶¶ 85-86). Dr. Easttom reached this conclusion because the raw data was unpacketized before it was used "locally to calculate a control signal (e.g., a resistance level)." (D.I. 74 at ¶ 86). I agree with Peloton that Dr. Easttom's invalidity analysis is inconsistent with the conclusion that the transmission of JSON information can satisfy claim 29. Based on ICON's own analysis, the JSON information downloaded in advance of a class cannot satisfy the second packetized control signal limitation.

ICON's most convincing piece of evidence offered on reply is a test run by Dr. Easttom in which he disconnected the Bike+ from the internet while it was running Auto Follow and noted that it did not continue making automatic resistance changes as the class continued offline. (Id. at ¶¶ 73-76). Dr. Easttom asserts that these results indicate information received via the internet (which would be, by definition, packetized) is necessary to run Auto Follow. (Id. at ¶ 76). While this test undermines Peloton's assertion that the Auto Follow feature is run entirely locally, I do not believe that it necessarily demonstrates the existence of a packetized control signal. Dr. Easttom concludes that Auto Follow "clearly requires something from the video and Peloton's remote servers to operate" but does not state that the "something" is a control signal. (Id. at ¶ 77).

This information was provided in Dr. Easttom's supplemental declaration (D.I. 74) and cited in ICON's Reply Brief. Peloton's expert declined to give an explanation of the videotaped test at his deposition, which was the first time he saw it, without being able to conduct additional analysis. (D.I. 75, Ex. H at 181:1-182:8, 194:20-195:23).

Peloton also argues that the red resistance knob on the Bike+ identified by ICON as the "scaling control" does not allow the user to make a proportional change to a control signal consistent with claim 47. (D.I. 43 at 8). Peloton argues that the knob allows the user to change the actual resistance of the bike which is not reflected in a proportional change in the resistance calculated by Auto Follow. (Id.). Peloton's expert provides the following example:

[I]n the scenario where the current range is 30-40 and the user adjusts the resistance to 20%, the Bike+'s resistance is changed automatically to 20 (which is not 20% in the range of 30-40). When the next timestamp hits and a recommended resistance range now becomes 0-10, the resistance is automatically
changed to 0 (since the value "20" selected by the user is below the previous recommended resistance range).
(D.I. 44 at ¶ 91).

ICON's reply on this point is that Peloton admitted that resistance on the Bike+ is based on "relative" position, which it argues is a synonym for proportional. (D.I. 73 at 2 (citing D.I. 47 at ¶¶ 35-38)). Peloton's expert did describe the Auto Follow program as adjusting the rider's resistance to "stay at the same relative position" going into the next resistance range (see, e.g., D.I. 45 at ¶ 10) but clarified that the user cannot scale the resistance ranges with the resistance knob. (Id. at ¶ 11). Rather, the expert states that the rider's use of the knob changes the resistance of the Bike+ as a percentage of the maximum resistance of the bike. (D.I. 45 at ¶ 12).

Dr. Easttom provides data that indicates that the Auto Follow feature maintains the rider's resistance in proportion to the rider's initially selected position within the resistance range. (D.I. 10 at ¶ 82). However, as ICON argues that the JSON information can satisfy the control signal limitation (D.I. 73 at 2), it does not appear that such data demonstrates that a rider can effect a proportional change to the information contained in the JSON file with the control knob. For example, it does not appear that a rider makes a proportional change to the pre-downloaded resistance ranges with the red resistance knob.

In light of ICON's cursory showing on infringement, I cannot conclude that it has carried its burden to demonstrate a likelihood of success on the merits as to infringement of the '062 Patent. Peloton has raised a substantial question as to whether the resistance on the Bike+ when using Auto Follow is controlled via locally stored information, and, assuming the existence of a packetized control signal, whether the Bike+'s red resistance knob is configured to effect a proportional change to that control signal.

B. Irreparable Harm

The movant seeking the preliminary injunction must demonstrate "that it is likely to suffer irreparable harm if the preliminary injunction is not granted and there is a causal nexus between the alleged infringement and the alleged harm." Metalcraft of Mayville, Inc. v. Toro Co., 848 F.3d 1358, 1368 (Fed. Cir. 2017). The alleged harm must not be compensable via monetary damages. Id. (stating that where no monetary damage is calculable, "the harm cannot be adequately compensated and is irreparable").

ICON asserts three grounds of irreparable harm as a result of Peloton's infringement: (1) ICON and Peloton are head-to-head competitors; (2) once Peloton secures customers via its infringing product it retains them for a long period of time; and (3) ICON's iFit platform will be injured during a period of "unprecedented demand," prior to ICON's anticipated IPO and the launch of new products. (D.I. 9 at 11-12).

ICON argues that Peloton is its direct competition in the "at home fitness equipment" market (D.I. 11 at ¶ 17) or, as argued more narrowly in ICON's Reply Brief, the "at-home connected fitness market" (D.I. 73 at 5; see also D.I. 76 at ¶¶ 4-7). In support of its market definition, ICON cites internal Peloton documents referencing Peloton's "core market" and identifying ICON's iFit brand. (D.I. 50, Ex. 39 at 26). At oral argument, Peloton asserted that the connected fitness market could not be considered a two-player market because other companies, like Echelon, Tonal, Nautilus, and Hydrow, compete in the space. (D.I. 97 at 40:16-17, 21-24). However, together ICON and Peloton control approximately 90% of the connected fitness market and the parties' market shares often move more or less inversely. (See D.I. 73 at 6; D.I. 76 at ¶¶ 4-5). The presence of additional competitors in the market is not fatal to a finding of irreparable harm. See Robert Bosch, LLC v. Pylon Mfg. Corp., 659 F.3d 1142, 1151 (Fed. Cir. 2011) (stating that "the absence of a two-supplier market does not weigh against a finding of irreparable harm"). As Peloton has not cited to any evidence to the contrary in its briefing (beyond the existence of other at-home connected fitness companies), I believe ICON's evidence is more than sufficient to demonstrate that Peloton and ICON are direct competitors.

Next, ICON argues that its loss of market share to Peloton is particularly damaging because of the nature of subscriber-based platforms. (D.I. 9 at 12). Peloton has estimated that once it obtains a subscriber, the relationship will last 13 years. (D.I. 11 at ¶ 59). ICON has estimated that it retains its subscribers for 9 years. (Id. at ¶ 59). The subscriber base is an important means through which to sell additional "brand-centric" products and services. (D.I. 73 at 7). Further, ICON asserts that the loss of these downstream sales to its subscribers and the loss of subscribers in advance of its IPO are impossible to quantify fully. (D.I. 9 at 13, 16).

Lastly, ICON advances a price erosion argument. (Id. at 14). Following the release of the Bike+, Peloton dropped the price of its Bike product, which lacks the accused Auto-Follow feature, by $600. (Id.). This price drop has allowed Peloton to seize more market share from ICON via purchasers of the Bike. (Id.). However, ICON offered no economic analysis to support its price erosion theory and, without such evidence, it is speculation to assert that competing with an admittedly non-infringing product will drive down ICON's prices or lead to a loss in market share. Waters Corp., 410 F. Supp. 3d at 715-16 (price erosion argument without "concrete pricing evidence" was "too speculative").

ICON did cite to expert discussion that the price drop in the Bike was an effort to expand Peloton's subscriber base. (D.I. 11 at ¶ 51). But this citation does not adequately connect the marketing strategy for the non-infringing Bike to the sales of the Bike + with Auto-Follow or address any downward pricing pressure on ICON's products attributable to the Auto-Follow feature.

Peloton argues that the harms articulated by ICON are, alternatively, calculable and speculative. (D.I. 43 at 15). First, Peloton notes that all of the lost profits evidence ICON provided concerning Bike+ customers reflected individuals switching from the Peloton Bike to the Bike+. (Id. at 16 (citing D.I. 49, Ex. 7 at 69-72, 82-86)). I agree that this is inadequate to demonstrate that these individuals would have necessarily considered ICON's products before switching to the accused product.

Second, Peloton argues that ICON has produced no evidence of lost convoyed sales and, if convoyed sales are indeed lost, these are calculable. (D.I. 43 at 16). In its opening brief, ICON identified seven categories of sales it expects to make to subscribers over their estimated nine-year lifespan. (D.I. 9 at 4). In response, Peloton cites to internal ICON documents estimating the value per subscriber in each of these categories. (See D.I. 49, Ex. 4). ICON did not address the relevance of this document in its Reply Brief, instead citing Apple, Inc. v. Samsung Elecs. Co., 678 F.3d 1314, 1336-37 (Fed. Cir. 2012) (O'Malley, J., concurring-in-part and dissenting-in-part), in support of the theory that lost downstream sales can support irreparable harm. (D.I. 73 at 7). A high degree of brand loyalty and potential lost customers "may have long term effects that are difficult to calculate and may not be recaptured." Apple v. Samsung, 678 F.3d at 1336-37.. However, since ICON has undertaken to estimate the value of these lost customers, the general difficulties identified by Judge O'Malley may not impact the damages calculation in this case. (See D.I. 49, Ex. 4).

Next, Peloton argues that ICON has not produced any evidence of a loss in market share as a result of the Bike+. (D.I. 43 at 17). In reply, ICON produced evidence that ICON's market share the week before the Bike+ launch (23%) was at its highest since June 2020, while Peloton's market share (64%) was at its lowest. (D.I. 73 at 6 (citing D.I. 48, Att. 47 at 6)). Following the Bike+ launch in September 2020, Peloton's market share jumped to 75% while ICON's dropped to 14%. (D.I. 46, Att. H at 16). However, at oral argument, Peloton noted that subsequent market data indicated that ICON's share had trended back up in October while Peloton's dropped, arguing that ICON has seen consistent subscriber growth since December 2019, while market share has regularly fluctuated. (D.I. 49, Ex. 2 at 12; D.I. 97 at 46: 19-22, 47:14-16).

Lastly, Peloton argues that ICON's licensing activity negates its claims of irreparable harm. (D.I. 43 at 13). It appears that only one of the cited licenses is relevant to this action. In 2014, ICON licensed the '062 Patent to Saris, a company that sells bike trainers. (Id. at 14). ICON argues that this license is not relevant to market exclusivity because Saris is not a direct competitor. (D.I. 73 at 8 (citing D.I. 75, Ex. D at 128:9-130:3; 131:14-132:7)). Neither Peloton nor ICON noted Saris in the market data displayed for the at-home connected fitness market. While it is likely true that ICON's stationary bike products compete with bike trainers for a certain subset of customers, there is no evidence indicating that Saris is a "significant competitor" to ICON. See Nichia Corp. v. Everlight Ams., Inc., 855 F.3d 1328, 1343 (Fed. Cir. 2017) (considering the fact that licenses were granted to "significant competitors who posed major threats to [patentee's] flagship products") (internal quotations omitted). Thus, this license does not weigh significantly against irreparable harm. See Acumed LLC v. Stryker Corp., 551 F.3d 1323, 1328 (Fed. Cir. 2008) (noting factors to weigh when considering the grant of past licenses).

The '062 license to Nautilus excludes the remote-control technology ICON asserts here. (D.I. 50, Ex. 11; D.I. 75 at ¶¶ 13-18). The license to Echelon contains a right of first refusal to ICON's "control patents," not a license to use the asserted technology. (D.I. 50, Ex. 38 at 4).

Bike trainers require that a customer already own a road bike. (D.I. 49, Ex. 7 at 129:16-19). --------

On this record, I believe that ICON has established that it competes directly with Peloton and that a loss of market share, and thus subscribers, to Peloton has the potential to be damaging due to the longevity of Peloton's subscriber base. However, I do not believe that ICON has demonstrated that such losses are imminent as the result of the release of the Bike+ or non-compensable. ICON's assertion that it is planning an IPO and planning to release new remote-control products (D.I. 9 at 12) is speculative and cannot serve as the grounds for finding irreparable harm without additional evidence. See Chestnut Hill Sound Inc. v. Apple Inc., 2015 WL 6870037, at *5 (D. Del. Nov. 6, 2015).

In view of my conclusions on the likelihood of success on the merits and on irreparable harm, I need not make any findings concerning the third and fourth factors. Id. at *6.


As I have concluded that ICON has failed to demonstrate a likelihood of success on the merits or irreparable harm, I will not address the remaining factors. ICON's motion for a preliminary injunction is DENIED.

IT IS SO ORDERED this 22 day of February 2021.

/s/ Richard G. Andrews

United States District Judge

Summaries of

ICON Health & Fitness, Inc. v. Peloton Interactive, Inc.

Feb 22, 2021
Civil Action No. 20-cv-1386-RGA (D. Del. Feb. 22, 2021)
Case details for

ICON Health & Fitness, Inc. v. Peloton Interactive, Inc.

Case Details



Date published: Feb 22, 2021


Civil Action No. 20-cv-1386-RGA (D. Del. Feb. 22, 2021)