ericsson, inc. v. d link systems, inc. 1201 · ericsson, inc. v. d link systems, inc. 1201 cite as...

1201ERICSSON, INC. v. D–LINK SYSTEMS, INC.Cite as 773 F.3d 1201 (Fed. Cir. 2014)

grounds for invalidity, such as enablement,if necessary.8

VACATED AND REMANDED

,

ERICSSON, INC., TelefonaktiebolagetLM Ericsson, and Wi–Fi One, LLC,

Plaintiffs–Appellees,

v.

D–LINK SYSTEMS, INC., Netgear, Inc.,Acer, Inc., Acer America Corporation,and Gateway, Inc., Defendants–Appel-lants,

and

Dell, Inc., Defendant–Appellant,

and

Toshiba America Information Systems,Inc. and Toshiba Corporation,

Defendants–Appellants,

and

Intel Corporation, Intervenor–Appellant,

and

Belkin International, Inc., Defendant.

Nos. 2013–1625, 2013–1631,2013–1632, 2013–1633.

United States Court of Appeals,Federal Circuit.

Dec. 4, 2014.

Background: Patentee commenced actionagainst competitors, alleging infringementof patents generally relating to Wi–Fitechnology employed by electronic devices

to wirelessly access the Internet. TheUnited States District Court for the East-ern District of Texas, Leonard Davis, J.,2013 WL 4046225, denied competitor’spost-judgment motions after a jury verdictin the patentee’s favor and upheld thejury’s infringement and validity findingsand refused to grant a new trial. Competi-tors appealed.

Holdings: The Court of Appeals, O’Mal-ley, Circuit Judge, held that:

(1) substantial evidence supported jury’sfindings that accused product infringedlimitation over ‘‘service type identifierwhich identifies a type of payload in-formation’’;

(2) term, ‘‘responsive to the receiving step,constructing a message field for a sec-ond data unit, said message field in-cluding a type identifier field’’ meantresponsive to the receiving step, gener-ating a message field including a fieldthat identifies the message type of thefeedback response message from anumber of different message types;

(3) accused devices met ‘‘responsive to thereceiving step, constructing a messagefield for a second data unit, said mes-sage field including a type identifierfield’’ step;

(4) competitor induced indirect infringe-ment of method claim;

(5) accused products did not infringe claimrequiring that ‘‘a transmitter command[ ] a receiver to a) receive at least onepacket and b) release any expectationof receiving outstanding packets’’;

(6) jury was entitled to credit patentee’sexpert testimony over competitor’s ex-pert testimony to conclude that priorart publication did not anticipate pat-ent;

8. The pending motion to dissolve the injunc- tion pending appeal is denied as moot.

1202 773 FEDERAL REPORTER, 3d SERIES

(7) competitor waived argument that thedistrict court prejudicially erred by al-lowing patentee’s counsel to referencetotal cost of laptop when discussingrequested royalty rate; and

(8) on issue of first impression, districtcourt could not include all 15 Georgia–Pacific factors in its damages instruc-tion without considering their rele-vance to record created at trial andwhether they were misleading.

Affirmed in part, reversed in part, vacatedin part, and remanded.

Taranto, Circuit Judge, filed opinion dis-senting-in-part.

1. Courts O96(7)In a patent case, a trial court’s deci-

sion on a motion for judgment as a matterof law (JMOL) is reviewed under the lawof the regional circuit.

2. Patents O1970(13)Patent claim construction is an issue

of law that is reviewed de novo.

3. Patents O1970(5, 15)Patent infringement and anticipation

are issues of fact reviewed for substantialevidence.

4. Patents O1828(2)Substantial evidence supported jury’s

findings that accused product infringedlimitation over ‘‘service type identifierwhich identifies a type of payload informa-tion’’ in patent generally relating to Wi–Fitechnology employed by electronic devicesto wirelessly access the Internet; althoughevidence showed that traffic identifier(TID) field, relied upon by patentee tomeet service type identifier limitation, didnot always identify its payload type, atleast one video calling program used theinvention, accused device was reasonablycapable of arranging information for trans-

mission that identified type of payload in-formation without significant alterations,and compliant chip producer instructed de-velopers to use TID field in infringingmanner.

5. Patents O1555To prove literal patent infringement,

the patentee must show that the accuseddevice contains each and every limitationof the asserted claims.

6. Patents O1395Phrase ‘‘responsive to the receiving

step, constructing a message field for asecond data unit, said message field includ-ing a type identifier field,’’ in patent gener-ally relating to Wi–Fi technology employedby electronic devices to wirelessly accessthe Internet, meant responsive to the re-ceiving step, generating a message fieldincluding a field that identifies the mes-sage type of the feedback response mes-sage from a number of different messagetypes.

7. Patents O1341Generally, patent claim terms should

be given their ordinary and customarymeaning to a person having ordinary skillin the art at the time of the effective dateof the patent application.

8. Patents O1313, 1338(1), 1345To ascertain the scope and meaning of

the asserted patent claims, a court looks tothe words of the claims themselves, thespecification, the prosecution history, andany relevant extrinsic evidence; this inqui-ry typically begins and ends with the in-trinsic evidence.

9. Patents O1328When ascertaining the scope and

meaning of the asserted patent claims, thespecification is the single best guide to themeaning of the claim terms, and it is usual-ly dispositive.


10. Patents O1329Although patent claims must be read

in light of the specification, it is importantthat a court avoids importing limitationsfrom the specification into the claims.

11. Patents O1655Accused devices met ‘‘responsive to

the receiving step, constructing a messagefield for a second data unit, said messagefield including a type identifier field’’ stepin asserted claims of patent generally re-lating to Wi–Fi technology employed byelectronic devices to wirelessly access theInternet, as required for literal infringe-ment; although accused devices used onlyone type of feedback response, multipledifferent feedback response types did nothave to be actually used.

12. Patents O1564A method patent claim is directly in-

fringed when someone practices every stepof the patented method.

13. Patents O1600In order to prove induced patent in-

fringement, the patentee must show thatthe alleged infringer performs, or inducesanother party to perform, every singlestep in the method.

14. Patents O1600Inducement requires that the alleged

patent infringer knowingly induced in-fringement and possessed specific intent toencourage another’s infringement.

15. Patents O1661Competitor induced indirect infringe-

ment of method claim in patent generallyrelating to Wi–Fi technology employed byelectronic devices to wirelessly access theInternet, where competitor knew about thepatent, it knew that the patent potentiallywas essential to wireless chip standard towhich it intentionally complied, and allsteps of method claim were performed on

end product that was controlled by third-party.

16. Patents O1655

Accused products did not infringeclaim requiring that ‘‘a transmitter com-mand [ ] a receiver to a) receive at leastone packet and b) release any expectationof receiving outstanding packets,’’ in pat-ent generally relating to Wi–Fi technologyemployed by electronic devices to wireless-ly access the Internet, where receiver au-tomatically handled out-of-order packets.

17. Patents O603

Jury was entitled to credit patentee’sexpert testimony over competitor’s experttestimony to conclude that prior art publi-cation did not anticipate patent generallyrelating to Wi–Fi technology employed byelectronic devices to wirelessly access theInternet.

18. Patents O489(2)

A claim is anticipated only if each andevery limitation is found either expresslyor inherently in a single prior art refer-ence.

19. Patents O577

Because patents are presumed valid,anticipation must be proven by clear andconvincing evidence.

20. Courts O96(7)

In a patent case, decisions on motionsfor a new trial and the admission of experttestimony are reviewed under the law ofthe regional circuit.

21. Federal Courts O3606(1)

The Fifth Circuit reviews the denial ofa new trial motion for abuse of discretion,reversing only if there is an absolute ab-sence of evidence to support the jury’sverdict.


22. Federal Courts O3600The Fifth Circuit reviews the trial

court’s admission or exclusion of experttestimony for an abuse of discretion.

23. Patents O1967The legal sufficiency of a jury instruc-

tion on an issue of patent law is reviewedde novo.

24. Federal Courts O3703(1)A jury verdict will be set aside only if

the jury instructions were legally errone-ous and the errors had prejudicial effect.

25. Patents O1903In a patent infringement case, appor-

tionment is required even for non-royaltyforms of damages: a jury must ultimatelyapportion the defendant’s profits and thepatentee’s damages between the patentedfeature and the unpatented features usingreliable and tangible evidence. 35U.S.C.A. § 284.

26. Patents O1904, 1914In a patent infringement case, an

economist can measure the value of pat-ented and unpatented features for a rea-sonable royalty award by careful selectionof the royalty base to reflect the valueadded by the patented feature, where thatdifferentiation is possible, by adjustmentof the royalty rate so as to discount thevalue of a product’s non-patented features,or by a combination thereof; the essentialrequirement is that the ultimate reason-able royalty award must be based on theincremental value that the patented inven-tion adds to the end product. 35 U.S.C.A.§ 284.

27. Patents O1904When addressing a reasonable royalty

award in a patent infringement case,where a multi-component product is at is-sue and the patented feature is not theitem that imbues the combination of the

other features with value, care must betaken to avoid misleading the jury by plac-ing undue emphasis on the value of theentire product; reliance on the entire mar-ket value might mislead the jury, who maybe less equipped to understand the extentto which the royalty rate would need to dothe work in such instances. 35 U.S.C.A.§ 284.

28. Patents O1903Where the entire value of a machine

as a marketable article is properly andlegally attributable to the patented fea-ture, the damages owed to the patentee forinfringement may be calculated by refer-ence to that value; however, where it isnot, a court must insist on a more realisticstarting point for the royalty calculationsby juries which often is the smallest sala-ble unit and, at times, even less. 35U.S.C.A. § 284.


award in a patent infringement case, thefact that a license is not perfectly analo-gous generally goes to the weight of theevidence, not its admissibility; in each case,district courts must assess the extent towhich the proffered testimony, evidence,and arguments would skew unfairly thejury’s ability to apportion the damages toaccount only for the value attributable tothe infringing features. 35 U.S.C.A.§ 284.

30. Federal Courts O3701(9)In a patent infringement case, where

expert testimony explains the need to thejury when considering a reasonable royaltyaward to discount reliance on a given li-cense to account only for the value attrib-uted to the licensed technology, the merefact that licenses predicated on the valueof a multi-component product are refer-enced in that analysis, and the districtcourt exercises its discretion not to exclude


such evidence, is not reversible error. 35U.S.C.A. § 284; Fed.Rules Evid.Rule 403,28 U.S.C.A.

31. Patents O1849

In a patent infringement case, whenlicenses based on the value of a multi-component product are admitted as evi-dence of a reasonable royalty award, oreven referenced in expert testimony, acourt should give a cautionary instructionregarding the limited purposes for whichsuch testimony is proffered if the accusedinfringer requests the instruction; thecourt also should ensure that the instruc-tions fully explain the need to apportionthe ultimate royalty award to the incre-mental value of the patented feature fromthe overall product. 35 U.S.C.A. § 284.

32. Patents O1957

Competitor waived argument that thedistrict court prejudicially erred by allow-ing patentee’s counsel to reference totalcost of laptop when discussing requestedroyalty rate; although competitor had con-tinuing objection to patentee’s expert’s ref-erence to prior licenses, competitor did notobject to counsel’s reference to marketvalue of laptop at trial, competitor referredto value of its own end products on cross-examination, and it did not raise issue inits post-trial motions. 35 U.S.C.A. § 284.

33. Patents O1849

District court could not include all 15Georgia–Pacific factors in its damages in-struction in patent infringement case with-out considering their relevance to recordcreated at trial and whether they weremisleading; court had to consider facts ofrecord when instructing jury on reasonableroyalty and had to avoid rote reference toany particular damages formula. 35U.S.C.A. § 284.

34. Patents O1849

A trial court must carefully considerthe evidence presented when crafting anappropriate jury instruction for a reason-able royalty award for patent infringe-ment. 35 U.S.C.A. § 284.

35. Patents O1849

A trial court should consider the pat-entee’s actual reasonable and nondiscrimi-natory (RAND) commitment in craftingthe jury instruction regarding a reasonableroyalty award. 35 U.S.C.A. § 284.

36. Patents O1904

As with all patents, the royalty ratefor standard essential patents (SEPs) mustbe apportioned to the value of the patentedinvention. 35 U.S.C.A. § 284.

37. Patents O1904

When dealing with standard essentialpatents (SEPs), to ensure that the royaltyaward is based on the incremental valuethat the patented invention adds to theproduct, not any value added by the stan-dardization of that technology, the patent-ed feature must be apportioned from all ofthe unpatented features reflected in thestandard, and the patentee’s royalty mustbe premised on the value of the patentedfeature, not any value added by the stan-dard’s adoption of the patented technology.35 U.S.C.A. § 284.

38. Patents O1904

A royalty award for a standard essen-tial patent (SEP) must be apportioned tothe value of the patented invention, or atleast to the approximate value thereof, notthe value of the standard as a whole, anda jury must be instructed accordingly;however, if a patentee can show that hisinvention makes up the entire value of thestandard, an apportionment instructionprobably would not be appropriate. 35U.S.C.A. § 284.


39. Patents O1914When considering a reasonable royal-

ty award, a patent holder should only becompensated for the approximate incre-mental benefit derived from his invention.35 U.S.C.A. § 284.

40. Patents O1914The royalty for standard essential pat-

ents (SEPs) should reflect the approximatevalue of that technological contribution,not the value of its widespread adoptiondue to standardization. 35 U.S.C.A. § 284.

41. Federal Civil Procedure O2174A court should not instruct a jury on a

proposition of law about which there is nocompetent evidence.


award in a infringement case, if an accusedinfringer wants an instruction on patenthold-up and royalty stacking, it must pro-vide evidence on the record of patent hold-up and royalty stacking in relation to boththe reasonable and nondiscriminatory(RAND) commitment at issue and the spe-cific technology referenced therein. 35U.S.C.A. § 284.

43. Patents O1485Competitor did not have license to

practice patents at issue under masterpurchase agreement (MPA) with subsid-iary of corporate parent, where corporateparent was patent owner and it was not asignatory to MPA. 35 U.S.C.A. § 281.

44. Principal and Agent O24Under New York law, the existence of

an agency relationship, where one partyhas legal authority to act for another, is amixed question of law and fact.

45. Principal and Agent O1In order to establish an agency rela-

tionship under New York law, the factsmust show that: (1) the principal manifest-

ed intent to grant authority to the agent,and (2) the agent agreed or consented tothe agency relationship.

46. Principal and Agent O1

An agency relationship under NewYork law requires the principal to retaincontrol and direction over key aspects ofthe agent’s actions.

47. Principal and Agent O50

Under New York law, a principal can-not grant authority to an agent if theprincipal does not itself possess the powergranted.

Patents O2091

6,424,625. Cited.

Patents O2091

6,466,568, 6,772,215. Infringed.

Douglas A. Cawley, McKool Smith, P.C.,of Dallas, TX, argued for plaintiffs-appel-lees Ericsson Inc., et al. With him on thebrief were Theodore Stevenson, III andWarren Lipschitz, and John B. Campbelland Kathy H. Li, of Austin, TX. Of counselon the brief was John M. Whealan, ofChevy Chase, MD.

William F. Lee, Wilmer Cutler Picker-ing Hale and Dorr LLP, of Boston, MA,argued for defendants-appellants and in-tervenor-appellant. With him on the brieffor intervenor-appellant Intel Corporationwere Joseph J. Mueller, Mark C. Fleming,and Lauren B. Fletcher, of Boston, MA;and James L. Quarles, III, of Washington,DC. Of counsel on the brief were GregArovas, Kirkland & Ellis LLP, of NewYork, New York, Adam R. Alper, of SanFrancisco, CA, and John C. O’Quinn, ofWashington, DC. On the brief for defen-dants-appellants D–Link Systems, Inc., et


al., were Robert A. Van Nest, Steven A.Hirsch, Eugene M. Paige and Matan Shac-ham, Keker & Van Nest LLP, of SanFrancisco, CA; Christine M. Morgan,Doyle B. Johnson, Jonah D. Mitchell, ScottD. Baker, Reed Smith LLP, of San Fran-cisco, CA; and James C. Martin, of Pitts-burgh, PA. On the brief for defendants-appellants Toshiba Corporation, et al.,were John J. Feldhaus and Pavan K.Agarwal, Foley & Lardner LLP, of Wash-ington, DC.

Michael J. Newton, Alston & Bird LLP,of Dallas, TX, argued for defendant-appel-lant, Dell, Inc. With him on the brief wereDwayne C. Norton and Shaun W. Hassett;and Frank G. Smith, III, of Atlanta, GA.

Michael A. Lindsay, Dorsey & WhitneyLLP, of Minneapolis, MN, for amicus curi-ae The Institute of Electrical and Elec-tronics Engineers, Incorporated. Of coun-sel on the brief was Eileen M. Lach,IEEE, General Counsel and Chief Compli-ance Officer, of New York, N.Y.

Richard M. Brunell, for amicus curiaeAmerican Antitrust Institute, of Washing-ton, DC.

Jeffrey Blumenfeld, Lowenstein SandlerLLP, of New York, NY, for amici curiae,Cisco Systems, Inc., et al. Of counsel onthe brief was Marta Beckwith, Cisco Sys-tems, Inc., San Jose, CA. On the brief foramicus curiae Hewlett–Packard Companywas Barry K. Shelton, Bracewell & Giulia-ni, of Austin, TX.

T. Andrew Culbert, Microsoft Corpora-tion, of Redmond, VA, for amicus curiaeMicrosoft Corporation. With him on thebrief was David E. Killough.

Dan L. Bagatell, Perkins Coie LLP, ofPhoenix, AZ, for amici Broadcom Corpora-tion. With him on the brief was AmandaTessar, of Denver, CO. On the brief forMarvell Semiconductor, Inc. was DonaldM. Falk, Mayer Brown LLP, of Palo Alto,

CA. On the brief for Media Tek Inc. wasSteven C. Holtzman, Boies, Schiller &Flexner LLP, of Oakland, CA.

Richard S. Taffet, Bingham McCutchenLLP, of New York, N.Y., for amicus curiaeDolby Laboratories, Inc. On the brief wasPatrick Strawbridge, of Boston, MA.

Roger G. Brooks, Cravath, Swaine &Moore LLP, of New York, NY, for amicuscuriae Qualcomm Incorporated.

Daryl L. Joseffer, King & SpaldingLLP, of Washington, DC, for amici curiaeNokia Corporation, et al. With him on thebrief was Ethan P. Davis.

Before O’MALLEY, TARANTO, andHUGHES, Circuit Judges.

O’MALLEY, Circuit Judge.

Ericsson, Inc. & TelefonaktiebolagetLM Ericsson (collectively, ‘‘Ericsson’’)brought suit against D–Link Systems, Inc.;Netgear, Inc.; Acer, Inc.; Acer AmericaCorp.; Gateway, Inc.; Dell, Inc.; ToshibaAmerica Information Systems, Inc.; andToshiba Corp., with Intel Corp. interven-ing (collectively, ‘‘D–Link’’), in the UnitedStates District Court for the Eastern Dis-trict of Texas, alleging infringement of,inter alia, certain claims from U.S. PatentNos. 6,424,625 (‘‘the 8625 patent’’); 6,466,-568 (‘‘the 8568 patent’’); and 6,772,215(‘‘the 8215 patent’’). All of the patents atissue generally relate to Wi–Fi technologyemployed by electronic devices to wireless-ly access the Internet. Ericsson allegedthat all of the patents at issue were essen-tial to the Wi–Fi standard, which wouldmean that all Wi–Fi–capable devices in-fringe Ericsson’s patents.

The case progressed to a jury trial,where the jury found that D–Link infring-ed the asserted claims of the three patentsand assigned roughly $10 million in dam-ages—approximately 15 cents per infring-


ing device. After post-trial motions, thedistrict court upheld the jury’s infringe-ment and validity findings and refused togrant a new trial based on an allegedviolation of the ‘‘entire market value rule’’(‘‘EMVR’’) and allegedly deficient jury in-structions regarding the standard-settingcontext and Ericsson’s ‘‘reasonable andnon-discriminatory’’ licensing obligationsderived from that context. For the rea-sons explained below, we affirm-in-part,reverse-in-part, vacate-in-part, and re-mand.

I. BACKGROUND

A. Technology and StandardsBackground

Interoperability is an essential require-ment for many electronic devices. Forexample, if a user brings her laptop to alocal coffee shop, she expects that herlaptop will charge when she plugs it in andthat she will be able to access the Internetwhen she connects to the coffee shop’swireless network. For the user to be ableto charge her laptop, the plug must be inthe correct shape and the laptop chargermust be able to accept the voltage outputof the outlet. For the user to be able toconnect to the Internet, her laptop mustknow, inter alia, what frequency to searchfor the wireless signal, what messages tosend to the network to set up a connection,and how to interpret the messages sentfrom the network. Though most userstake for granted that their electronic de-vices will be able to charge and connect tothe wireless Internet anywhere, interoper-

ability does not happen automatically. Be-cause of the multitude of devices, devicedesigners, and manufacturers, there mustbe an established standard mode of opera-tion to ensure compatibility among all ofthese different devices.

Standards development organizations(‘‘SDOs’’) publish standards, which arelists of technical requirements. Compli-ance with these technical requirements en-sures interoperability among compliant de-vices. Of course, at least a critical mass ofdevice developers must adopt the standardin order to ensure mass interoperability.1

Relevant to this case, the Institute of Elec-trical and Electronics Engineers, Inc.(‘‘IEEE’’) publishes the 802.11 standards,more commonly known as ‘‘Wi–Fi.’’ Br. ofAmici Curiae Institute of Electrical andElectronics Engineers, Inc. (‘‘IEEE Br.’’)at 1–2. The 802.11 standard is the prevail-ing wireless internet standard and has al-ready reached widespread adoption. A de-vice is considered 802.11–compliant if itadheres to the IEEE’s technological re-quirements stated in the 802.11 standard.Requiring all 802.11–compliant devices tooperate in a certain way ensures that ev-ery compliant device can communicatewith all other 802.11–compliant devices.2

For example, an 802.11–compliant laptopwill be able to establish a connection withan 802.11–compliant router. The 802.11standards also govern how subsequentdata is passed between the laptop and therouter once that connection is established.This includes, inter alia, data formatting,

1. A single standard will often emerge even ifinitially there are competing standards. Forexample, the Blu-ray standard won out overthe HD DVD standard in the high-definitionoptical disc war. Martin Fackler, ToshibaAcknowledges Defeat as Blu-ray Wins FormatBattle, N.Y. TIMES, February 20, 2008, at C1.

2. The 802.11 standard is not a static set ofrequirements. Indeed, it has been updatedmultiple times. When the 802.11 standard isupdated, it is given a letter to indicate theversion of the standard, e.g., 802.11(g). De-vices are typically compliant with a specificversion of the 802.11 standard, indicated bythe letter. This case specifically addresses the802.11(n) standard.


prioritization, error handling, and flow con-trol.

Importantly for this case, data files arenot sent between a router and a laptop in asingle transmission. For example, if a lap-top user wants to download a video, therouter does not send the entire file in asingle huge transmission. Instead, eachdata file is broken into ‘‘packets,’’ whereeach packet is sent in a different transmis-sion. Small files may only require a singlepacket, whereas large files, e.g., video andsound, may require thousands of packets.The receiving device then reassembles thefile out of the packets. The data from thefile in the packet is called the ‘‘payload.’’Because packets may be lost or arrive outof order, the 802.11 standard providesways to handle these errors. For exam-ple, each packet has a ‘‘header’’ that is sentto the receiving device with the packet.The header contains, inter alia, a sequencenumber so the receiving device knows theorder in which to reassemble the payloadof the packets.

Creating some standards, like IEEE’s802.11 standard, is a complicated processthat involves the collaboration and can in-volve cooperation of a number of interest-ed parties. IEEE Br. 4–12. Due to thecollaborative nature of this process, thechosen standard may include technologydeveloped by a number of different par-ties. Sometimes that technology is cov-ered by patents. Because the standardrequires that devices utilize specific tech-nology, compliant devices necessarily in-fringe certain claims in patents that covertechnology incorporated into the standard.These patents are called ‘‘standard essen-tial patents’’ (‘‘SEPs’’). IEEE Br. 13–14.

SEPs pose two potential problems thatcould inhibit widespread adoption of thestandard: patent hold-up and royaltystacking. Patent hold-up exists when theholder of a SEP demands excessive royal-

ties after companies are locked into usinga standard. Royalty stacking can arisewhen a standard implicates numerous pat-ents, perhaps hundreds, if not thousands.If companies are forced to pay royalties toall SEP holders, the royalties will ‘‘stack’’on top of each other and may becomeexcessive in the aggregate. To help allevi-ate these potential concerns, SDOs oftenseek assurances from patent owners be-fore publishing the standard. IEEE, forexample, asks SEP owners to pledge thatthey will grant licenses to an unrestrictednumber of applicants on ‘‘reasonable, andnondiscriminatory’’ (‘‘RAND’’) terms.IEEE Br. at 16–18.

B. Ericsson’s SEPs

Ericsson has asserted that all of thepatents at issue are SEPs for IEEE’s802.11(n) standard. Ericsson promised tooffer licenses for all of its 802.11(n) SEPsat a RAND rate via letters of assurance tothe IEEE. In its letters, Ericsson pledgedto ‘‘grant a license under reasonable ratesto an unrestricted number of applicants ona worldwide basis with reasonable termsand conditions that are demonstrably freeof unfair discrimination.’’ Joint Appendix(‘‘J.A.’’) 17253. The parties agree that thiscommitment is binding on Ericsson. Seealso IEEE Br. 19–20.

1. The 8568 Patent

The 8568 patent, titled ‘‘Multi–Rate Ra-diocommunication Systems and Termi-nals,’’ describes prioritizing packets basedon the type of payload in the packet. Theprioritization of packets is important be-cause networks all have a bandwidth limi-tation. Bandwidth refers to the amount ofdata that can be sent across the networkat one time. When a network receivesmultiple requests at the same time, it mustbe able to respond to all of the requests ina timely fashion. Due to the network’s


bandwidth limitation, however, the mes-sages cannot be sent all at once. Thoughnetworks can deal with the bandwidth limi-tation problem in different ways, eachmethod involves dividing the availablebandwidth among the outstanding re-quests.

The 8568 patent explains that networkstransmit a variety of different types ofpayloads, including ‘‘voice, video, anddata.’’ According to the 8568 patent, net-works in the prior art did not have theability to prioritize certain types of dataover others. Because certain types oftransmissions are less preferable when de-layed—e.g., voice calling—the 8568 patentdiscloses transmitting the type of trans-mission as part of the header. This wouldallow the network to dedicate more band-width to the higher priority transmissiontypes, thereby sending those packets morequickly.

Claims 1 and 5 are at issue in thisappeal. Claim 1 is representative:

1. A communications station compris-ing:a processor for arranging informationfor transmission including providing atleast one first field in which payloadinformation is disposed and providing atleast one second field, separate fromsaid first field, which includes a servicetype identifier which identifies a type ofpayload information provided in said atleast one first field; anda transmitter for transmitting informa-tion received from said processor includ-ing said at least one first field and saidat least one second field.

8568 patent col. 13 ll. 11–21 (emphasis add-ed).

2. The 8215 Patent

As described above, files are broken intopackets, which are sent to the receivingdevice with sequence numbers so the re-

ceiving device can reassemble the payloadin the correct order. Packets, however,are often lost or corrupted during trans-mission. To ensure that the receiver re-ceives the payload in those lost or corrupt-ed packets, the transmitter will have toresend those packets. For the transmitterto know which packets need to be resent,the receiving device must tell the transmit-ting device which packets it did not receiveor are corrupted. This may be done usingan ‘‘Automatic Repeat Request’’ (‘‘ARQ’’)protocol. In an ARQ protocol, the receiv-ing device will send a ‘‘feedback response’’to the transmitting device. Though feed-back response messages can be in differ-ent formats, the feedback response willgenerally indicate which packets, if any,are missing or corrupted. The transmit-ting device will then retransmit thosemissing packets.

Although ARQ protocols existed in theprior art, the 8215 patent, titled ‘‘Methodfor Minimizing Feedback Responses inARQ Protocols,’’ asserts that those priorart ARQ protocols wasted bandwidth be-cause they were ‘‘static’’ and not adapta-ble. By making the feedback responsetype dynamic, the 8215 patent disclosesthat the response could be formatted inthe most efficient response type. For ex-ample, if 1 packet out of 100 is missing,just the missing packet number could besent. Conversely, if 50 out of 100 packetsare missing, the response could be a bit-map with a bit set to one to indicate themissing packets, instead of a list of all 50missing packet numbers. To solve thisalleged deficiency in the prior art, the 8215patent discloses adding a ‘‘type identifierfield’’ (‘‘TIF’’) to the feedback responsethat identifies the format of that feedbackresponse. This would allow the receiver tochoose dynamically between differenttypes of feedback responses based on


which response would be most efficient,e.g., a list of packet numbers or a bitmap.

Claim 1 is the independent claim at is-sue:

1. A method for minimizing feedbackresponses in an ARQ protocol, compris-ing the steps of:

sending a plurality of first data unitsover a communication link;

receiving said plurality of first dataunits; and

responsive to the receiving step, con-structing a message field for a seconddata unit, said message field includinga type identifier field and at least one ofa sequence number field, a length field,and a content field.

8215 patent col. 10 ll. 19–28 (emphasis add-ed).

3. The 8625 Patent

Due to technical limitations, prior artreceiving devices used a limited ‘‘receptionwindow’’ for keeping track of which pack-ets it had received. Because this receptionwindow was finite, if the receiving devicereceived a packet outside of the window, itwould not accept the packet. The windowwould not move forward until it receivedall of the packets in the current window.This process ensures that the receiver willreceive all of the missing packets. Forcertain ‘‘delay sensitive applications,’’ how-ever, a 0% packet loss rate is not requiredand significantly delayed packets provideno benefit—e.g., ‘‘telephony, video confer-encing, and delay sensitive control sys-tems.’’ 8625 patent col. 3 ll. 51–53.

According to the 8625 patent, prior arttransmitting devices had no way to tell thereceiving device to ignore unnecessary,missing packets and shift the receivingwindow forward. The 8625 patent dis-closes adding a way for the transmittingdevice to force the receiving device to ac-

cept packets that may be out of its recep-tion window. This will also shift the re-ception window forward and the receivingdevice will forget about the delayed or lostpackets that would no longer provide anybenefit.

Claim 1 is at issue in this appeal:1. A method for discarding packets in adata network employing a packet trans-fer protocol including an automatic re-peat request scheme, comprising thesteps of:a transmitter in the data network com-manding a receiver in the data networkto a) receive at least one packet having asequence number that is not consecutivewith a sequence number of a previouslyreceived packet and b) release any ex-pectation of receiving outstanding pack-ets having sequence numbers prior tothe at least one packet; andthe transmitter discarding all packetsfor which acknowledgment has not beenreceived, and which have sequence num-bers prior to the at least one packet.

8625 patent col. 10 ll. 13–26 (emphasesadded).

C. The Accused Products

The accused infringers in this case pro-duce a variety of electronic devices, includ-ing laptop computers and routers (‘‘theend products’’), which incorporate802.11(n) wireless chips made by Intel.Because all of these end products incorpo-rate 802.11(n)–compliant chips, they mustbe capable of the functionality mandatedby the 802.11(n) standard.

For example, the standard requires that,in the header of an 802.11(n)–compliantpacket, there must be a traffic identifier(‘‘TID’’) field that indicates the priority ofthe data. This TID field has a value from0–7, which indicates its priority. Packageswith higher priority typically will be sentmore quickly or given more bandwidththan those with lower priority. And thestandard explains that one use of this TIDidentifier is to associate particular values


with particular types of information insidepackages—to correspond to a particular

kind of payload—as indicated in the follow-ing table giving an example:

J.A. 15781. In this table, each priority isgiven an informative ‘‘Designation.’’ Forexample, priority levels 4 and 5 are giventhe designation ‘‘Video.’’

For further examples of requiredfunctionality, the 802.11(n) standard also

requires feedback response headers toinclude information about the type offeedback response in the BlockAck field:

J.A. 16778. The 802.11(n) standard allowsthree different types of feedback respons-es: Basic BlockAck, Compressed Block-Ack, and Multi–TID BlockAck. Sendingthis information as part of the header ismandatory for interoperability betweendevices.

All 802.11(n)–compliant devices, more-over, must be capable of accepting any

packets they receive. In other words,802.11(n)–compliant devices do not use alimited reception window. The receiver isjust programmed to receive automaticallyall packets, regardless of the packet’s se-quence number.

D. The Dell–Ericsson AB AgreementDell argues on appeal that it has a li-

cense to practice the patents at issue


based on its prior agreement with Erics-son AB. Ericsson AB is a Swedish manu-facturing and development subsidiary ofLM Ericsson. LM Ericsson is the Swed-ish parent corporation of Ericsson AB andowns the patents-in-suit. LM Ericssonand its North American subsidiary, Erics-son, Inc., are the two plaintiffs-appellees inthe suit.

On February 13, 2008, Ericsson AB andDell executed a Master Purchase Agree-ment (‘‘MPA’’), under which Ericsson ABwould provide Dell with mobile broadbandproducts for three years from the date ofexecution of the MPA. Ericsson AB is theonly named ‘‘supplier’’ listed in the MPA,as well as the only signatory to the MPAaside from Dell. The MPA also separatelydefined Ericsson AB’s ‘‘Affiliates.’’ At is-sue in this appeal, Section 12.1 of theMPA, entitled ‘‘Dispute Resolution,’’ statedthat ‘‘[s]upplier will not commence anylawsuit or seek any judicial order affectingDell or add Dell as a party to any pendinglegal or administrative proceeding that isnot directly related to Dell’s purchase ofProducts or that may prevent Dell fromshipping any Dell or third-party products.’’J.A. 6348 (emphasis added).

E. Procedural History

On September 14, 2010, Ericsson filedsuit in the United States District Court forthe Eastern District of Texas, accusing D–Link of infringing nine patents that, ac-cording to Ericsson, were essential to the802.11(n) wireless standard. Intel, thewireless internet chip supplier for the ac-cused products, intervened. On March 8,2013, the magistrate judge issued a claim

construction order, which the district courtjudge adopted. Ericsson Inc. v. D–LinkCorp. (‘‘Claim Construction Order ’’), No.6:10–cv–473, 2013 WL 949378 (E.D.Tex.Mar. 8, 2013). Shortly before trial, thetrial judge denied D–Link’s motion to ex-clude the testimony of Ericsson’s damagesexpert, over D–Link’s argument that thetestimony violated the EMVR. Prior totrial, the court also granted summaryjudgment against Dell, rejecting its argu-ment that it had a license based on theMPA.

Although the parties were forced to nar-row the case for trial, Ericsson still ac-cused D–Link of infringing 5 different pat-ents at trial. On June 13, 2013, after a 7–day jury trial, the jury found that D–Linkinfringed the asserted claims in three ofEricsson’s patents—the 8568, 8215, and8625 patents. The jury also found that the8625 patent was valid over a prior artpublication (‘‘the Petras reference’’). Aspast damages for that infringement, thejury awarded Ericsson approximately $10million—roughly 15 cents per infringingdevice. After the jury trial, the trial courtconducted a separate bench trial regardingseveral RAND issues.3

Following the bench trial, D–Link filed amotion for judgment as a matter of law(‘‘JMOL’’) and a new trial, arguing thatthe jury’s findings of infringement and noinvalidity, as well as its damages award,were not supported by substantial evi-dence. D–Link further contended thatEricsson’s expert violated the EMVR byrelying on licenses that were based on thevalue of the end products. D–Link assert-ed, moreover, that the jury was inade-

3. At the bench trial, D–Link asked the districtcourt to: (1) determine an appropriate RANDrate, (2) find that Ericsson breached its RANDagreement by refusing to license Intel, and (3)find that Ericsson is not entitled to an injunc-tion. D–Link concedes that it proffered evi-dence to the trial court regarding Ericsson’sRAND obligations that it did not offer to thejury, despite its argument to the jury that anyroyalty rate chosen must be reflective of thoseRAND obligations. It is unclear why D–Link

made this choice, particularly because D–Linkrefused to be bound by any court-determinedroyalty rate. Once the jury had set the RANDrate, the judge rejected D–Link’s invitation toseparately determine the award at the benchtrial. D–Link does not appeal that ruling,arguing only that the jury was not adequatelyinstructed about Ericsson’s RAND obli-gations, not that the court should have madethat decision.


quately instructed regarding Ericsson’sRAND obligation.

The trial court denied D–Link’s post-trial motions, finding that substantial evi-dence supported: (1) the jury’s findings ofinfringement; (2) the validity of the 8625patent; and (3) the jury’s $10 millionaward. The judge also concluded that Er-icsson’s damages testimony was not incon-sistent with the EMVR and that the juryinstruction regarding Ericsson’s RANDobligations was adequate. The judge fur-ther found that, based on the jury’s award,15 cents per product was an appropriateongoing RAND rate for the three infring-ed patents. Ericsson Inc. v. D–Link Corp.(‘‘JMOL Order ’’), No. 6:10–cv–473, 2013WL 4046225 (E.D.Tex. Aug. 6, 2013). Ac-cording to the trial court, moreover, Erics-son did not violate its RAND obligationsby offering Intel a license at the rate of 50cents per unit. In fact, the court conclud-ed that it was Intel that violated its obli-gation to negotiate a royalty rate in goodfaith. Id. at *16.

D–Link timely appealed to this court.We have jurisdiction under 28 U.S.C.§ 1295(a)(1) (2012).

II. DISCUSSION

On appeal, D–Link raises a number ofissues: (1) whether the jury had substan-tial evidence to find that D–Link infringedclaims 1 and 5 of the 8568 patent; (2)whether the district court properly con-strued the term ‘‘responsive to the receiv-ing step, constructing a message field for asecond data unit, said message field includ-ing a type identifier field’’ in the 8215patent, and, if the district court correctlyconstrued that term, whether the jury hadsubstantial evidence to find that D–Linkinfringed claims 1 and 2 of the 8215 patent;(3) whether the jury had substantial evi-dence to find that D–Link infringed claim1 of the 8625 patent and that the Petrasreference did not anticipate the 8625 pat-ent; (4) whether Ericsson’s damages theo-

ry was presented in violation of theEMVR; (5) whether the jury was instruct-ed properly regarding Ericsson’s RANDobligations; and (6) whether Dell had alicense to practice the patents at issuebased on its agreement with Ericsson AB.We address each issue in turn.

A. Infringement

[1–3] We first address D–Link’s chal-lenges to the infringement findings. Wereview the trial court’s decision on a mo-tion for JMOL under the law of the re-gional circuit, in this case, the Fifth Cir-cuit. Verizon Servs. Corp. v. Cox FibernetVa., Inc., 602 F.3d 1325, 1331 (Fed.Cir.2010). The Fifth Circuit reviews the deni-al of a motion for JMOL de novo, but the‘‘jury’s verdict can only be overturned ifthere is no legally sufficient evidentiarybasis for a reasonable jury to find as thejury did.’’ Miller v. Raytheon Co., 716F.3d 138, 144 (5th Cir.2013). We reviewissues of patent law applying this court’scase law. Claim construction is an issue oflaw reviewed de novo. Lighting BallastControl LLC v. Philips Elecs. N. Am.Corp., 744 F.3d 1272, 1276–77 (Fed.Cir.2014) (en banc). Infringement and antici-pation are issues of fact reviewed for sub-stantial evidence. 01 Communique Lab.,Inc. v. Log–MeIn, Inc., 687 F.3d 1292,1296 (Fed.Cir.2012); In re Montgomery,677 F.3d 1375, 1379 (Fed.Cir.2012).

1. The 8568 Patent

[4] D–Link contends that the jury didnot have substantial evidence to find in-fringement of the asserted claims of the8568 patent, specifically with respect to the‘‘service type identifier which identifies atype of payload information’’ limitation.Neither party challenges the districtcourt’s construction of that limitation as‘‘an identifier that identifies the type ofinformation conveyed in the payload. Ex-amples of types of information include, butare not limited to, video, voice, data, and


multimedia.’’ Claim Construction Order,2013 WL 949378, at *11.

[5] To prove literal infringement, thepatentee must show that the accused de-vice contains each and every limitation ofthe asserted claims. Presidio Compo-nents, Inc. v. Am. Tech. Ceramics, Corp.,702 F.3d 1351, 1358 (Fed.Cir.2012).

Before the jury, Ericsson relied on theTID field value in the 802.11(n) standardto meet the service type identifier limita-tion. Ericsson’s expert testified that eachTID field contains an integer that estab-lishes a priority of service. According toEricsson’s expert, each TID value may beassigned an ‘‘informative’’ designation:Background, Best Effort, Video, or Voice.Ericsson’s expert further stated that de-vices need to implement the TID field inorder to be compliant with the 802.11(n)standard. Ericsson’s expert pointed toseveral programs that ‘‘take advantage’’ ofthe ‘‘informative’’ use of the TID capabilityto assign TID values to payloads of aparticular type. J.A. 1395–96 (testifyingthat CSipSimple, Skype, Ekiga, and Win-dows Media take advantage of the TIDcapability). Ericsson also presented anIntel document, recommending that devel-opers utilize the TID field based on theinformative designation categories de-scribed by the 802.11(n) standard.

D–Link’s expert, on the other side,countered that the TID designations areused to prioritize packets, but do not iden-tify the type of information contained inthe payload’s packet, e.g., a text email canbe sent using the Video designation. D–Link’s expert testified that the he ran testson video and voice programs that did notassign different TID numbers for videoand voice data. In other words, the pro-gram assigned a TID designation of zero,regardless of the type of data in the pay-load. On cross examination, however, D–Link’s expert admitted that he also ob-

served some traffic in which the TID des-ignation did correspond to the content ofthe payload. Indeed, when presented withhis own expert report, D–Link’s expertadmitted that at least one program—Eki-ga, a video conferencing program—usedthe video TID designation for its packetswith a video payload. In fact, D–Link’sown expert testified that Ekiga was ‘‘usingthe [8568] invention.’’ J.A. 1568 (‘‘A. Ekigais using the invention, you said? Q. Yes.That’s what you tested, right? A. Yes.’’).

The jury, using the district court’s con-struction for service type identifier, foundthat D–Link infringed claims 1 and 5 ofthe 8568 patent. The district court deniedD–Link’s subsequent JMOL motion, ex-plaining that D–Link’s evidence that theTID designation does not always corre-spond to the payload is, ‘‘[a]t best, TTT

evidence show[ing] [its] products can beconfigured in a non-infringing manner.’’JMOL Order, 2013 WL 4046225, at *6.

On appeal, D–Link first argues that,under the district court’s proper construc-tion, the service type identifier must ‘‘iden-tif[y] the type of information conveyed inthe payload.’’ Claim Construction Order,2013 WL 949378, at *11 (emphasis added).D–Link insists that the TID field relatesonly to priority and does not identify thepayload of the packet. D–Link assertsthat Ericsson’s infringement contentionsfor the 8568 patent are thus premised onthe mere ‘‘capability of infringement.’’ Ac-cording to D–Link, it was an error for thejudge to instruct the jury that ‘‘[a]n ac-cused system or product directly infringesa claim if it is reasonably capable of satis-fying the claim elements even though itmay also be capable of non-infringingmodes of operation.’’ Appellants’ Br. 35.D–Link argues that mere capability doesnot constitute infringement unless theclaim language is ‘‘drawn to capability.’’Appellants’ Br. 36 (citing Finjan, Inc. v.


Secure Computing, Corp., 626 F.3d 1197,1204 (Fed.Cir.2010); Ball Aerosol & Spe-cialty Container, Inc. v. Limited Brands,Inc., 555 F.3d 984, 994 (Fed.Cir.2009)).D–Link asserts, moreover, that Ericssonfailed to show any programs that actuallyassign TID values according to their infor-mative designations, e.g., voice and video.

Ericsson responds that the jury’s find-ing of infringement of the asserted claimsof the 8568 patent was supported by sub-stantial evidence, including testimony byEricsson’s expert and the chart showingthe informative designations presented tothe jury. Ericsson further contends thatD-link’s own technical expert’s tests re-vealed traffic that corresponded to the cor-rect TID designations. Ericsson alsopoints to Intel’s instruction manual thaturges developers to use the correct TIDdesignations for the various types of data.According to Ericsson, moreover, thiscourt has repeatedly held that a productinfringes if it is reasonably capable of sat-isfying the claimed elements. For exam-ple, Ericsson argues that, in Finjan, thiscourt ‘‘held that where an apparatus claimis styled as a component ‘for’ performingsome function, the claim is drawn to capa-bility and the reasonable capability testapplies.’’ Appellees’ Br. 33 (citing Finjan,626 F.3d at 1204–05).

We are unpersuaded by D–Link’s argu-ment that the jury did not have substantialevidence to find infringement of claims 1and 5 of the 8568 patent. We recognizethat the evidence showed that the TIDfield does not always identify its payloadtype. Indeed, D–Link’s ex-pert’s testimo-ny that many programs did not utilize theTID field according to the informative des-ignations was unchallenged. D–Link’s ex-pert, however, also admitted that at leastone video calling program ‘‘us[ed] the in-vention.’’ J.A. 1568. Furthermore, Erics-son’s expert testified as to several exam-

ples of programs running on the accuseddevices where the TID field indicated thetype of payload.

We understand that the TID field maybe inherently only a priority field. Butthat field necessarily has the capability tobe used to identify the payload type, asshown by the informative example in thestandard and by the proof that it was infact so used by some device users. Cru-cially, this capability satisfies the patentclaim language here, which means all ac-cused devices could be found by the jury toinfringe.

In Fantasy Sports Properties, Inc. v.Sportsline.com, Inc., 287 F.3d 1108 (Fed.Cir.2002), we held that software for play-ing fantasy football could infringe a claimcovering a ‘‘computer for playing football.’’287 F.3d at 1118. Though a user mustinstall and activate functions in the soft-ware to infringe the claims, the FantasySports opinion explained that the user isonly activating means that are alreadypresent in the underlying software. Id. InBall Aerosol, on the other hand, the patentclaimed an apparatus arranged in a certainmanner. We reversed a grant of summaryjudgment of infringement because theclaims were not drawn to capability andthere was no evidence that the accuseddevice ‘‘was ever placed in the infringingconfiguration.’’ 555 F.3d at 995.

In Finjan, we found that the systemclaims at issue described ‘‘capabilities,’’without describing any software compo-nents that must be ‘‘active’’ or ‘‘enabled.’’For example, the claim language in Finjanrequired ‘‘a logical engine for preventingexecution’’ and ‘‘a communications enginefor obtaining a Downloadable.’’ Finjan,626 F.3d at 1204–05 (emphases added inopinion). In Finjan, we found that, inorder for the accused system to infringe,the logical engine only needed to be rea-sonably capable of ‘‘preventing execution’’


and the communications engine only need-ed to be reasonably capable of ‘‘obtaining aDownloadable.’’ Id. Our court explainedfurther in Versata, that, ‘‘[w]hile ‘a devicedoes not infringe simply because it is pos-sible to alter it in a way that would satisfyall the limitations of a patent claim,’ TTT anaccused product ‘may be found to infringeif it is reasonably capable of satisfying theclaim limitation.’ ’’ Versata Software, Inc.v. SAP Am., Inc., 717 F.3d 1255, 1262(Fed.Cir.2013) (quoting Finjan, 626 F.3dat 1204). In Versata, we found that thepatentee presented evidence that, if a userfollowed the accused infringer’s own in-structions, the system would operate in aninfringing manner. Id. at 1263. In sum,when the asserted claims recite capability,our case law supports finding infringementby a ‘‘reasonably capable’’ accused deviceon a case-by-case basis particularly where,as here, there is evidence that the accuseddevice is actually used in an infringingmanner and can be so used without signifi-cant alterations. See J.A. 1568 (‘‘A. Ekigais using the invention, you said? Q. Yes.That’s what you tested, right? A. Yes.’’);see also Ball Aerosol, 555 F.3d at 995.

The asserted claims of the 8568 patentare most similar to the claim at issue inFinjan. Both use language reciting capa-bility, as opposed to actual operation.Compare Finjan, 626 F.3d at 1204–05(system claims reciting ‘‘a logical enginefor preventing execution’’ and ‘‘a communi-cations engine for obtaining a Downloada-ble’’ (emphases added)), with 8568 patentcol. 13 ll. 12–18 (‘‘a processor for arrang-ing information for transmission TTT

which identifies a type of payload informa-tionTTTT’’ (emphasis added)). Accordingly,just as the accused system in Finjan onlyneeded to have components that are rea-sonably capable of ‘‘preventing execution’’and ‘‘obtaining a Downloadable’’ to in-fringe, Finjan, 626 F.3d at 1204–05, D–Link’s products only need to have a com-

ponent that is reasonably capable of ‘‘ar-ranging information for transmission TTT

which identifies a type of payload informa-tionTTTT’’ 8568 patent col. 13 ll. 12–18.Furthermore, similar to the evidence thepatentee presented in Versata, Ericssonpresented evidence that Intel—the 802.11–compliant chip producer—instructed devel-opers to use the TID field in an infringingmanner. See Versata, 717 F.3d at 1263.We therefore find that the jury could prop-erly base its infringement finding on thereasonable capability of the unmodified ac-cused devices.

For the foregoing reasons, we hold thatsubstantial evidence supports the jury’sfinding that D–Link infringed claims 1 and5 of the 8568 patent.

2. The 8215 Patent

[6] D–Link challenges the districtcourt’s construction of the term ‘‘respon-sive to the receiving step, constructing amessage field for a second data unit, saidmessage field including a type identifierfield’’ in the 8215 patent. Even if we af-firm the district court’s construction ofthat term, according to D–Link, the jurydid not have substantial evidence to findinfringement of claims 1 and 2 of the 8215patent. We first consider the properscope of the claims before comparing theconstrued claims to the accused devices.See Absolute Software, Inc. v. Stealth Sig-nal, Inc., 659 F.3d 1121, 1129 (Fed.Cir.2011).

a. Claim Construction

[7–9] Generally, claim terms should begiven their ordinary and customary mean-ing to a person having ordinary skill in theart at the time of the effective date of thepatent application. Phillips v. AWHCorp., 415 F.3d 1303, 1312–13 (Fed.Cir.2005) (en banc). To ascertain the scope


and meaning of the asserted claims, welook to the words of the claims themselves,the specification, the prosecution history,and any relevant extrinsic evidence. Id. at1315–17. This inquiry typically begins andends with the intrinsic evidence. In fact,the specification is the single best guide tothe meaning of the claim terms; it is usu-ally dispositive. Id. at 1318 (‘‘[T]he speci-fication ‘is always highly relevant to theclaim construction analysis. Usually, it isdispositiveTTTT’ ’’ (internal citations omit-ted)).

D–Link challenges the district court’sconstruction of ‘‘responsive to the receiv-ing step, constructing a message field for asecond data unit, said message field includ-ing a type identifier field’’ as ‘‘responsiveto the receiving step, generating a mes-sage field including a field that identifiesthe message type of the feedback responsemessage from a number of different mes-sage types.’’ Claim Construction Order,2013 WL 949378, at *4–6. D–Link arguesthat the district court should have adoptedits proposed construction: ‘‘responsive tothe receiving step, generating a messagefield including a field identifying the typeof feedback response that is selected frommultiple available feedback responses inorder to minimize the size or number offeedback responses.’’ Id. at *4 (emphasisadded). In essence, D–Link contends thatthe ‘‘type identifier field’’ must be used to‘‘select[ ] from multiple available feedbackresponses’’ and ‘‘minimize the size or num-ber of feedback responses.’’ Id.

D–Link asserts that, because the entirespecification of the 8215 patent emphasizesthat the point of the invention is to selectthe feedback response that minimizes thesize or number of feedback responses, wemust limit the scope of the claims to cap-ture the scope of the actual invention. Er-icsson responds that the district courtproperly excluded the two extraneous limi-

tations—selecting and minimizing—in itsconstruction. Because none of the lan-guage cited by D–Link amounts to a cleardisavowal of the claim scope, Ericsson con-tends that we should not read limitationsfrom the specification into the claims.

[10] We agree with Ericsson that D–Link’s proposed construction improperlyreads limitations from the specificationinto the claims. Although the claims mustbe read in light of the specification, it isimportant that we ‘‘avoid importing limita-tions from the specification into theclaims.’’ Phillips, 415 F.3d at 1323. Werecognize that there is a fine distinctionbetween these two concepts, but we must,as always, draw this distinction from thepoint of a view of a person of ordinary skillin the art. Id. Although the 8215 patentenvisions that the type identifier field beused to select the most efficient format offeedback response, the specification neverrequires the selection of the feedback re-sponse type that minimizes the size ornumber of feedback responses. See id. at1326–27 (‘‘The fact that the written de-scription of the [ ] patent sets forth multi-ple objectives to be served by the bafflesrecited in the claims confirms that theterm ‘baffles’ should not be read restric-tively to require that the baffles in eachcase serve all of the recited functions.’’).The type identifier field actually servesanother purpose—one distinct from theproposed selecting and minimizing func-tions—it must identify the type of feed-back response. This purpose is encom-passed by the district court’s construction.Although the type identifier field may beused to select and minimize, a person ofordinary skill would not read those limita-tions into the claims when the field hasanother purpose as well.

D–Link relies on Metabolite Laborato-ries, Inc. v. Laboratory Corp. of AmericaHoldings, 370 F.3d 1354 (Fed.Cir.2004) to


argue that we should use the preamble ofthe asserted claim to limit our construc-tion. In Metabolite, we stated that ‘‘[a]preamble may provide context for claimconstruction, particularly, where as here,that preamble’s statement of intended useforms the basis for distinguishing the priorart in the patent’s prosecution history.’’370 F.3d at 1362 (emphasis added). D–Link omits the italicized part of the Meta-bolite quote in its briefs, likely because theselection and minimization requirementswere not used as the basis for distinguish-ing the prior art in the prosecution historyof the 8215 patent. Indeed, the addition ofthe type identifier field, which identifiesthe type of feedback response, is a suffi-cient basis for distinguishing the prior art.This is reflected in the district court’s cor-rect construction because the messagefield must ‘‘includ[e] a field that identifiesthe message type of the feedback responsemessage from a number of different mes-sage types.’’ Claim Construction Order,2013 WL 949378, at *4 (emphasis added).

We therefore hold that the district courtcorrectly did not read the additional limita-tions D–Link identifies into the ‘‘type iden-tifier field’’ term of the 8215 patent; weadopt the district court’s construction.

b. Infringement

[11] D–Link argues that, even underthe district court’s construction, the jurydid not have substantial evidence to findthat the accused devices met the ‘‘respon-sive to the receiving step, constructing amessage field for a second data unit, saidmessage field including a type identifierfield’’ step in the asserted claims of the8215 patent.

[12–14] A method claim is directly in-fringed when someone practices every stepof the patented method. Cardiac Pace-makers, Inc. v. St. Jude Med., Inc., 576F.3d 1348, 1359 (Fed.Cir.2009). In order

to prove induced infringement, the paten-tee must show that the alleged infringerperforms, or induces another party to per-form, every single step in the method.Limelight Networks, Inc. v. AkamaiTechs., Inc., 572 U.S. ––––, 134 S.Ct. 2111,2117, 189 L.Ed.2d 52 (2014). Inducementrequires that the alleged infringer ‘‘know-ingly induced infringement and possessedspecific intent to encourage another’s in-fringement.’’ DSU Med. Corp. v. JMSCo., 471 F.3d 1293, 1306 (Fed.Cir.2006).

At trial, Ericsson presented evidence inthe form of expert testimony that all802.11(n)–compliant products, includingthe accused products, must send an appro-priate response in the BlockAck field.Furthermore, Ericsson presented evidencethat the BlockAck field must indicate oneof three different feedback response types.D–Link, to the contrary, presented evi-dence that, although the accused devicessend messages that contained the Block-Ack field, the accused products only useone type of feedback response type. D–Link argued that, because its productsonly use a single feedback response, theaccused product could not satisfy the‘‘from a number of different messagetypes’’ limitation.

The jury found that D–Link infringedclaims 1 and 2 of the 8215 patent. Indenying D–Link’s subsequent JMOL mo-tion, the district court explained thatwhether the use of a single feedback re-sponse meets the ‘‘from a number of dif-ferent message types’’ limitation is precise-ly the type of factual question that is to beresolved by a jury. The district courtclarified that the jury was authorized tofind direct infringement of a method claimby D–Link if its products automaticallyperform the disputed steps without usermodification. JMOL Order, 2013 WL4046225, at *9 (citing SiRF Tech., Inc. v.Int’l Trade Comm’n, 601 F.3d 1319, 1331


(Fed.Cir.2010)). The district court rea-soned that, because the accused productsperformed the claimed method when oper-ated by D–Link’s customers without anymodification, a finding of direct infringe-ment was justified. Regarding indirect in-fringement, the district court found thatEricsson had presented substantial evi-dence that D–Link possessed the requisiteintent because they continued to sell802.11(n)–compliant devices even after re-ceiving notice of the patents.

On appeal, D–Link argues that, underthe district court’s construction, the feed-back response message type must be gen-erated ‘‘from a number of different mes-sage types.’’ Claim Construction Order,2013 WL 949378, at *4 (emphasis added).According to D–Link, because the accusedproducts always send a single type of feed-back response, a jury did not have sub-stantial evidence to find infringement. D–Link further asserts that, under FederalCircuit law, a party that sells a productcontaining instructions to perform a pat-ented method does not directly infringethe method. Appellants’ Br. 46–47 (citingAristocrat Techs. Australia Pty Ltd. v.Int’l Game Tech., 709 F.3d 1348, 1362(Fed.Cir.2013); Ricoh Co. v. QuantaComp. Inc., 550 F.3d 1325 (Fed.Cir.2008)).Regarding induced infringement, more-over, D–Link contends that Ericsson didnot present sufficient facts from which ajury could conclude that it knew the in-duced acts constituted infringement. Ac-cording to D–Link, knowledge of the pat-ents plus advertising compliance with802.11(n) is not enough evidence uponwhich to base a finding of induced infringe-ment.

Ericsson responds that the claim lan-guage does not require that the accusedproducts send multiple types of feedbackresponses, only that they have a field that‘‘identifies the message type.’’ According

to Ericsson, because the accused productsmust utilize the BlockAck field to be com-pliant with the 802.11(n) standard, the juryhad substantial evidence to find infringe-ment. According to Ericsson, moreover,the Ricoh case relied on by D–Link appliesonly to software, not to ‘‘hard-wired’’ de-vices that will automatically perform theinfringing steps. Ericsson asserts that,like the accused products in SiRF, D–Linkdesigns the accused products to performthe method steps automatically wheneverthe products are used. According to Er-icsson, substantial evidence supports thefinding of induced infringement, includingevidence that D–Link advertises 802.11(n)compliance and submits its products forinteroperability testing and certification.

We must address two different issues:(1) whether the jury had substantial evi-dence to find that the BlockAck field wasselected ‘‘from a number of different mes-sage types’’ and (2) whether the jury hadsubstantial evidence to find direct or indi-rect infringement by D–Link. First, weagree with the district court and Ericssonthat the jury had substantial evidence tofind that using the BlockAck field meetsthe ‘‘from a number of different messagetypes’’ step. Claim Construction Order,2013 WL 949378, at *4. Nothing in thecourt’s proper construction requires thatmultiple different feedback response typesactually be used. Indeed, we have alreadyrejected D–Link’s proposed constructionthat would have added this requirement.At trial, Ericsson presented evidence thatthe 802.11(n) standard allows multipletypes of feedback responses. It is undis-puted, moreover, that the accused devicessend the code identifying the feedback re-sponse type in the BlockAck field, as re-quired by the 802.11(n) standard. Al-though the accused devices only use onetype of feedback response, the jury hadsubstantial evidence to find that the ac-cused devices can perform the method


claimed in the 8215 patent. Indeed, this isprecisely the type of factual dispute that ajury should be resolving.

[15] Because the asserted claim is amethod claim, however, the accused de-vices must also actually perform thatmethod. See Cardiac Pacemakers, 576F.3d at 1359. Although the jury was in-structed on both direct and indirect in-fringement, the verdict form only indicatedthat the jury found that D–Link infringedclaims 1 and 2 of the 8215 patent. In otherwords, the verdict did not distinguish be-tween direct and indirect infringement.For the reasons explained below, we mustaddress both D–Link’s direct and indirectinfringement arguments.

The district court relied on SiRF toconclude that the jury properly found thatD–Link directly infringed the methodclaim by selling the accused products. InSiRF, this court affirmed the InternationalTrade Commission’s finding that the man-ufacturer of GPS systems directly infring-ed the asserted method claims. SiRF, 601F.3d at 1331. The method claims at issuein SiRF required some, but not all, of thesteps of the claim to be executed by asatellite, which was controlled by the ac-cused infringers. The remaining stepswere then automatically performed by theaccused GPS products, which were in pos-session of the end users. Id. at 1329–30.In SiRF, accordingly, we concluded that,on these facts, it was the accused infring-ers that performed all the steps requiredfor direct infringement, not the customerswho possessed the GPS products. Id. at1331.

In Ricoh, on the other hand, we held‘‘that a party that sells or offers to sellsoftware containing instructions to per-form a patented method does not infringethe patent under § 271(a).’’ Ricoh, 550F.3d at 1335. The Ricoh court explainedthat there is a difference between the in-

structions contained in software and theprocess within the meaning of § 271(a).Our other decisions echo the idea fromRicoh that the direct infringer must actu-ally perform the steps in the methodclaim. See, e.g., Aristocrat Techs., 709F.3d at 1362 (noting that to prove directinfringement the patentee must show‘‘each and every step of the method orprocess was performed’’ by either the ac-cused infringer personally or ‘‘through an-other acting under [the accused infring-er’s] direction or control’’); AkamaiTechs., Inc. v. Limelight Networks, Inc.,692 F.3d 1301, 1307 (Fed.Cir.2012) (‘‘[F]ora party to be liable for direct patent in-fringement under 35 U.S.C. § 271(a), thatparty must commit all the acts necessaryto infringe the patent, either personally orvicariously.’’), overruled on other grounds,134 S.Ct. 2111; Travel Sentry, Inc. v.Tropp, 497 Fed.Appx. 958, 965 (Fed.Cir.2012) (holding that a party is liable fordirect infringement of a method claim onlyif that party exercises ‘‘control or di-rection’’ over the performance of each stepof the claim, including those the party doesnot itself perform); Lucent Techs., Inc. v.Gateway, Inc., 580 F.3d 1301, 1317 (Fed.Cir.2009) (finding sale of software alonedoes not directly infringe method claims ofpatent and seller can only be liable forinfringement as contributor and/or induc-er); Ormco Corp. v. Align Tech., Inc., 463F.3d 1299, 1311 (Fed.Cir.2006) (‘‘Methodclaims are only infringed when the claimedprocess is performed, not by the sale of anapparatus that is capable of infringinguse.’’).

Contrary to Ericsson’s assertions, ourdecision in SiRF did not create direct in-fringement liability whenever an allegedinfringer sells a product that is capable ofexecuting the infringing method. Our de-cision in SiRF is not applicable here be-cause all of the steps of the method in


claims 1 and 2 of the 8215 patent areperformed on the end product, which iscontrolled by a third party. See SiRF, 601F.3d at 1331. Unlike the method in SiRF,there are no steps automatically per-formed by equipment controlled by D–Link. In fact, none of our decisions havefound direct infringement of a methodclaim by sales of an end user productwhich performs the entire method, and wedecline to do so here. Because Ericssoncannot point to any evidence in the recordthat D–Link performed the infringingsteps, or that any of its customers wereunder its direction or control, the jury didnot have substantial evidence to find directinfringement of claims 1 and 2 of the 8215patent.

Importantly, however, the district courtdid not instruct the jury that D–Link coulddirectly infringe a method claim if theaccused products were used to execute thepatented steps. D–Link does not disputethe legal propriety of the direct infringe-ment instructions. This means that, if thejury found direct infringement, it was afactual error, not a legal error. Althoughwe think that it would have been a factualerror for the jury to find direct infringe-ment of the method claims by D–Linkitself, that error is not enough to set asidethe jury verdict because the jury’s findingalso could have been premised on indirectinfringement. See i4i Ltd. P’ship v. Mi-crosoft Corp., 598 F.3d 831, 849 (Fed.Cir.2010) (stating that a general verdict willnot be set aside ‘‘ ‘simply because the jurymight have decided on a ground that wassupported by insufficient evidence,’ ’’ butrather jury verdict should be upheld ifthere is sufficient evidence to support anyof the plaintiff’s alternative factual theo-ries) (quoting Walther v. Lone Star GasCo., 952 F.2d 119, 126 (5th Cir.1992)); cf.Griffin v. United States, 502 U.S. 46, 112S.Ct. 466, 116 L.Ed.2d 371 (1991) (holdingthat a general verdict should be invalidat-

ed when one of the possible bases wasbased on legal error).

We agree with the district court andEricsson that the jury had substantial evi-dence to find that D–Link induced in-fringement of claims 1 and 2 of the 8215patent. Ericsson presented evidence thatD–Link knew about the patents and knewthat the patents potentially were essentialto the 802.11(n) standard—a standard withwhich D–Link intentionally complied. D–Link countered by presenting evidencethat it did not think its actions constitutedinfringement of any of the claims of the8215 patent. Making findings of fact byweighing evidence—such as the evidencepresented by the parties regarding in-duced infringement—is the role of thejury. Questions of intent are quintessen-tial jury questions. See Allen Organ Co. v.Kimball Int’l, Inc., 839 F.2d 1556, 1567(Fed.Cir.1988) (‘‘Intent is a factual deter-mination particularly within the provinceof the trier of fact.’’). We cannot say thatthe jury did not have substantial evidenceto find induced infringement and we de-cline to supplant the jury’s factual findingswith our own. E.g., Lucent, 580 F.3d at1323 (‘‘Having perused the evidence, weagree with Microsoft that the evidence isnot strong, but we are not persuaded thatthe jury was unreasonable in finding thatMicrosoft possessed the requisite intent toinduce at least one user of its products toinfringe the claimed methods.’’).

For these reasons, we affirm the jury’sfinding of infringement of claims 1 and 2 ofthe 8215 patent.

3. The 8625 Patent

[16] D–Link argues on appeal that thejury did not have substantial evidence tofind that the accused devices infringedclaim 1 of the 8625 patent or find thepatent valid over the Petras prior art ref-


erence. We examine each of the jury’sfindings in turn.

a. Infringement

At trial, Ericsson argued that everytransmitted data packet acts as its own‘‘command to receive’’ from the transmit-ter. Ericsson’s expert testified that all802.11(n)–compliant receivers automatical-ly accept all packets, even if those packetsare out of order. Ericsson contended thatthis was just like an embodiment of the8625 patent where every packet is com-manding the receiver to accept an out-of-order packet. D–Link countered that, be-cause the normal operation of 802.11(n)–compliant receivers is to accept all pack-ets, there is no ‘‘command to receive’’ fromthe transmitter. D–Link insisted that Er-icsson failed to present any evidence that anormal packet will release the receiverfrom expectations of receiving outstandingpackets.

The jury found that D–Link infringedclaim 1 of the 8625 patent. The districtcourt refused to grant D–Link’s requestfor a JMOL, explaining that the jury wasentitled to credit Ericsson’s expert overD–Link’s expert.

D–Link argues on appeal that the ac-cused products are already programmed toaccept all valid data packets and do notneed to be commanded by the transmitterto accept out-of-order packets. D–Linkcontends that this means the jury’s findingis not supported by substantial evidence.According to D–Link, Ericsson’s expertadmitted that regular packet transmissionscould not act as a command to receivebecause the receivers already had the abil-ity to receive the packets. D–Link ex-plains that, because the receiver alreadyhad the ability to receive, the transmitterdid not command the receiver to do any-thing.

Ericsson responds that its expert testi-fied at trial that the accused products meteach and every limitation of the claims inthe 8625 patent. According to Ericsson,the fact that a receiver must receive apacket from the transmitter, even if out oforder, qualifies the message as a ‘‘com-mand to receive.’’ Ericsson emphasizesthat the patent conceived of an embodi-ment where the ‘‘command to receive’’ issent with every single message.

We agree with D–Link that the jury didnot have substantial evidence to find thatthe accused products infringe claim 1 ofthe 8625 patent. The asserted claim of the8625 patent requires that ‘‘a transmitterTTT command [ ] a receiver TTT to a) re-ceive at least one packet TTT and b) releaseany expectation of receiving outstandingpacketsTTTT’’ 8625 patent col. 10 ll. 16–21(emphases added). But there is no evi-dence in the record that it is the transmit-ter in the accused devices that commandsthe receiver to receive the out-of-orderpackets and release expectations of receiv-ing earlier packets. Instead, all of Erics-son’s evidence confirms that the receiverautomatically handles out-of-order packets.In other words, the transmitter does notcommand the receiver to do anything.The receiver just operates as programmedto handle out-of-order packets, regardlessof the messages that the transmittersends. Indeed, Ericsson’s expert admittedthat the receivers already have the abilityto accept all packets. J.A. 1412–13(‘‘Question: If a receiver could receive apacket that a transmitter was sending toit, is it correct that you would not need thecommand to receive in the 8625 patent tocommand or force the receiver to receivethe packet? Answer: I mean, that almostseems like a tautology. If it could receive,then would you need to insist that it re-ceives it? No, because it could alreadyreceive it.’’). In other words, the transmit-


ter does not command the receiver to ac-cept the packets.

Because no reasonable jury could havefound that the accused products meet eachand every limitation of claim 1 of the 8625patent, we reverse the district court’s re-fusal to grant JMOL of no infringement asto that patent.

b. Invalidity

[17–19] D–Link argued at trial thatthe Petras reference, a prior art publica-tion, anticipated claim 1 of the 8625 patent.A claim is anticipated only if each and

every limitation is found either expresslyor inherently in a single prior art refer-ence. Whitserve, LLC v. Computer Pack-ages, Inc., 694 F.3d 10, 21 (Fed.Cir.2012).Because patents are presumed valid, antic-ipation must be proven by clear and con-vincing evidence. Id.

The Petras reference discloses one typeof ARQ protocol where the transmitter cansend a ‘‘discard message’’ to the receiver.The discard message informs the receiverthat a message will not be resent. Thefollowing figure from Petras is illustrative:

J.A. 15041. In this figure, the transmitterfirst sends packets 0–3 to the receiverbefore the receiver sends back its firstfeedback responses. Due to delay andpacket loss, when the transmitter is firstnotified that the receiver did not receivepacket number 2—the second I(X)–SREJ(2) message—the transmitter has al-ready deleted packet 2. The transmitterwill then inform the receiver that it canignore packet number 2 by sending thediscard packet message—the I(5)–DIS-CARD(2) message.

At trial, D–Link argued that a discardmessage acts as the ‘‘command to receive’’from the transmitter. Ericsson’s experttestified against this theory, explainingthat the discard messages were not com-mands to receive; discard messages weremerely notifications that the transmitterhad discarded a packet. The jury foundthat Petras did not anticipate the 8625patent, and the district court refused tooverturn that finding.

On appeal, D–Link argues that Petrasdiscloses sending a discard message, which‘‘commands’’ the receiver to stop waitingfor delayed packets and shift its receptionwindow forward to receive later packets.Ericsson responds that substantial evi-dence supports the jury’s finding of noanticipation, pointing to its expert’s testi-mony that packet discard messages werewell known in the prior art. Ericsson alsoasserts that D–Link’s expert testimonywas inconsistent, so the jury was correct todiscredit his testimony.

We agree with the district court andEricsson that substantial evidence sup-ports the jury’s finding that Petras did notanticipate the 8625 patent. D–Link’s argu-ments on appeal essentially ask us to cred-it its position on discard messages overEricsson’s. We decline to do so. Bothparties presented expert testimony re-garding whether Petras anticipates the8625 patent, and we see no reason why the


jury was not entitled to credit Ericsson’sevidence over D–Link’s evidence. Wetherefore see no reason to disturb thejury’s verdict.

B. Damages

Having affirmed infringement with re-spect to two patents, we must also addressthe damages issues raised by D–Link. Asexplained below, we vacate the jury’s dam-ages award and remand for further pro-ceedings consistent with this opinion.

[20–22] We review decisions on mo-tions for a new trial and the admission ofexpert testimony under the law of theregional circuit. Verizon, 602 F.3d at1331. The Fifth Circuit reviews the denialof a new trial motion for abuse of discre-tion, reversing only if there is an ‘‘an abso-lute absence of evidence to support thejury’s verdict.’’ Duff v. Werner Enters.,Inc., 489 F.3d 727, 729 (5th Cir.2007). TheFifth Circuit reviews the trial court’s ad-mission or exclusion of expert testimonyfor an abuse of discretion. Snap–Drape,Inc. v. Comm’r, 98 F.3d 194, 197 (5thCir.1996).

[23, 24] We review de novo the legalsufficiency of a jury instruction on an issueof patent law. Sulzer Textil A.G. v. Pica-nol N.V., 358 F.3d 1356, 1363 (Fed.Cir.2004). A jury verdict will be set aside onlyif the jury instructions were ‘‘legally erro-neous’’ and the ‘‘errors had prejudicial ef-fect.’’ Id.

1. Admissibility of License Evidence

Before trial, D–Link moved to excludecertain testimony by Ericsson’s damagesexpert, arguing that it violated the EMVR.Specifically, D–Link argued that, becausethe damages calculations were, in part,based on licenses which were themselvestied to the entire value of the licensedproducts, even though the technology be-

ing licensed related to only a component ofthose products, the testimony was imper-missible as a matter of law. In denyingthat motion, the district court explainedthat Ericsson’s expert’s reference to thoseprior licenses was not improper becausethe expert properly apportioned any dam-ages calculations based on those licenses toaccount for the value of the patents atissue. D–Link noted its objection to thisline of testimony at trial by entering acontinuing objection to Ericsson’s expert’stestimony to the extent it was predicatedon or made any reference to these licenses.J.A. 1437–38 at 4:37–5:13. At trial, bothEricsson and D–Link then referred to thevalue of laptops generally. J.A. 1325 at11:24–12:21; J.A. 1332 at 37:22–38:11. Af-ter the jury found infringement andawarded Ericsson 15 cents per infringingdevice, D–Link moved for JMOL and for anew trial, arguing that the admission ofthis expert testimony violated the EMVR.The district court denied both motions.

On appeal, D–Link argues that the dis-trict court prejudicially erred by: (1) notexcluding Ericsson’s damages expert’s tes-timony on the challenged licenses, and (2)allowing Ericsson’s counsel to compare thecost of the end product to the requestedroyalty at trial. According to D–Link, be-cause Ericsson did not dispute that theasserted claims are practiced entirely bythe Wi–Fi chips—not by other componentsof the accused end products—Ericssonshould never have been allowed to base itsdamages award or its arguments at trialon the price of the end products.

Ericsson responds that the jury awardof 15 cents per infringing product is con-sistent with comparable Ericsson licenses,insisting that our court has found compa-rable licenses to be the best evidence of areasonable royalty rate. Ericsson furtherargues that the jury award is consistentwith ‘‘industry norms’’ and in accord with


its damages expert’s testimony. Accord-ing to Ericsson, its expert conducted arigorous analysis, which separated the val-ue of the patents at issue from any otherpatents covered by the licenses he refer-enced. Because of this apportionment,Ericsson asserts that neither its damagescalculation nor its expert’s reference toactual industry licenses was improper, un-der the EMVR or otherwise. With re-spect to counsel’s reference to the cost oflaptops at trial, Ericsson argues that D–Link never objected to these references,and made similar references itself.

We conclude that the district court prop-erly admitted evidence of the licenses towhich D–Link objects and that any objec-tion to counsel’s references to the cost ofitems incorporating the allegedly infring-ing chips was waived. While a number ofour cases have referred to the concept ofan entire market value ‘‘rule,’’ the legalstandard actually has two parts, which aredifferent in character. There is one sub-stantive legal rule, and there is a separateevidentiary principle; the latter assistingin reliably implementing the rule when—ina case involving a per-unit royalty—thejury is asked to choose a royalty base asthe starting point for calculating a reason-able royalty award.

As we explained recently in VirnetX,Inc. v. Cisco Systems, Inc., 767 F.3d 1308(Fed.Cir.2014), where multi-componentproducts are involved, the governing ruleis that the ultimate combination of royaltybase and royalty rate must reflect thevalue attributable to the infringing fea-tures of the product, and no more. 767F.3d at 1326 (citing Garretson v. Clark,111 U.S. 120, 121, 4 S.Ct. 291, 28 L.Ed. 371(1884)). As a substantive matter, it is the‘‘value of what was taken’’ that measures a‘‘reasonable royalty’’ under 35 U.S.C.§ 284. Dowagiac Mfg. Co. v. Minn. Mo-line Plow Co., 235 U.S. 641, 648, 35 S.Ct.

221, 59 L.Ed. 398 (1915). What is takenfrom the owner of a utility patent (forpurposes of assessing damages under§ 284) is only the patented technology, andso the value to be measured is only thevalue of the infringing features of an ac-cused product.

[25, 26] When the accused infringingproducts have both patented and unpatent-ed features, measuring this value requiresa determination of the value added by suchfeatures. Indeed, apportionment is re-quired even for non-royalty forms of dam-ages: a jury must ultimately ‘‘apportionthe defendant’s profits and the patentee’sdamages between the patented feature andthe unpatented features’’ using ‘‘reliableand tangible’’ evidence. Garretson, 111U.S. at 121, 4 S.Ct. 291. Logically, aneconomist could do this in various ways—by careful selection of the royalty base toreflect the value added by the patentedfeature, where that differentiation is possi-ble; by adjustment of the royalty rate soas to discount the value of a product’s non-patented features; or by a combinationthereof. The essential requirement is thatthe ultimate reasonable royalty awardmust be based on the incremental valuethat the patented invention adds to the endproduct.

[27, 28] Our cases have added to thatgoverning legal rule an important eviden-tiary principle. The point of the evidentia-ry principle is to help our jury systemreliably implement the substantive statuto-ry requirement of apportionment of royal-ty damages to the invention’s value. Theprinciple, applicable specifically to thechoice of a royalty base, is that, where amulti-component product is at issue andthe patented feature is not the item whichimbues the combination of the other fea-tures with value, care must be taken toavoid misleading the jury by placing undueemphasis on the value of the entire prod-


uct. It is not that an appropriately appor-tioned royalty award could never be fash-ioned by starting with the entire marketvalue of a multi-component product—by,for instance, dramatically reducing theroyalty rate to be applied in those cases—it is that reliance on the entire marketvalue might mislead the jury, who may beless equipped to understand the extent towhich the royalty rate would need to dothe work in such instances. See LaserDy-namics, Inc. v. Quanta Computer, Inc.,694 F.3d 51, 67, 68 (Fed.Cir.2012) (barringthe use of too high a royalty base—even ifmathematically offset by a ‘‘ ‘low enoughroyalty rate’ ’’—because such a base ‘‘car-ries a considerable risk’’ of misleading ajury into overcompensating, stating thatsuch a base ‘‘ ‘cannot help but skew thedamages horizon for the jury’ ’’ and ‘‘makea patentee’s proffered damages amountappear modest by comparison’’ (quotingUniloc USA, Inc. v. Microsoft Corp., 632F.3d 1292, 1320 (Fed.Cir.2011))). Thus,where the entire value of a machine as amarketable article is ‘‘properly and legallyattributable to the patented feature,’’ thedamages owed to the patentee may becalculated by reference to that value. Id.Where it is not, however, courts must in-sist on a more realistic starting point forthe royalty calculations by juries—often,the smallest salable unit and, at times,even less. VirnetX, 767 F.3d at 1327–28.

We apply these concepts to a challengeto expert testimony regarding licenses inwhich royalties were set by reference tothe value of an end product. We concludethat the expert testimony about which D–Link complains violated neither the rulefrom Garretson regarding apportionment,nor the evidentiary principle demanding anappropriate balance between the probativevalue of admittedly relevant damages evi-dence and the prejudicial impact of suchevidence caused by the potential to mis-lead the jury into awarding an unduly high

royalty. We find, accordingly, that thedistrict court did not err by failing toexercise its discretion under Federal Ruleof Evidence 403 to exclude the licensetestimony at issue here. Uniloc, 632 F.3dat 1320; see LaserDynamics, 694 F.3d at77–78 (finding that the district courtabused its discretion by failing to exclude alicense under Federal Rule of Evidence403).

[29] This court has recognized that li-censes may be presented to the jury tohelp the jury decide an appropriate royaltyaward. See, e.g., Monsanto Co. v. McFar-ling, 488 F.3d 973, 978 (Fed.Cir.2007) (‘‘Anestablished royalty is usually the bestmeasure of a ‘reasonable’ royalty for agiven use of an inventionTTTT’’); Georgia–Pacific Corp. v. U.S. Plywood Corp., 318F.Supp. 1116, 1120 (S.D.N.Y.1970) (findingthat ‘‘royalties received by the patentee forthe licensing of the patent in suit’’ is arelevant factor for the jury to consider).Prior licenses, however, are almost neverperfectly analogous to the infringement ac-tion. VirnetX, 767 F.3d at 1330. Forexample, allegedly comparable licensesmay cover more patents than are at issuein the action, include cross-licensing terms,cover foreign intellectual property rights,or, as here, be calculated as some percent-age of the value of a multi-componentproduct. Testimony relying on licensesmust account for such distinguishing factswhen invoking them to value the patentedinvention. Recognizing that constraint,however, the fact that a license is notperfectly analogous generally goes to theweight of the evidence, not its admissibili-ty. See Apple Inc. v. Motorola, Inc., 757F.3d 1286, 1326 (Fed.Cir.2014) (‘‘Here,whether these licenses are sufficientlycomparable such that Motorola’s calcula-tion is a reasonable royalty goes to theweight of the evidence, not its admissibili-ty.’’); accord ActiveVideo Networks, Inc. v.


Verizon Commc’ns, Inc., 694 F.3d 1312,1333 (Fed.Cir.2012) (‘‘Although we maynot have decided these evidentiary issuesthe same way had we presided over thetrial, the district court did not abuse itsdiscretion.’’). In each case, district courtsmust assess the extent to which the prof-fered testimony, evidence, and argumentswould skew unfairly the jury’s ability toapportion the damages to account only forthe value attributable to the infringing fea-tures.

[30] As the testimony at trial estab-lished, licenses are generally negotiatedwithout consideration of the EMVR, andthis was specifically true with respect tothe Ericsson licenses relating to the tech-nology at issue. Making real world, rele-vant licenses inadmissible on the groundsD–Link urges would often make it impossi-ble for a patentee to resort to license-based evidence. Such evidence is relevantand reliable, however, where the damagestestimony regarding those licenses takesinto account the very types of apportion-ment principles contemplated in Garret-son. In short, where expert testimonyexplains to the jury the need to discountreliance on a given license to account onlyfor the value attributed to the licensedtechnology, as it did here, the mere factthat licenses predicated on the value of amulti-component product are referenced inthat analysis—and the district court exer-cises its discretion not to exclude suchevidence-is not reversible error.4

[31] We do conclude, however, that,when licenses based on the value of amulti-component product are admitted, oreven referenced in expert testimony, the

court should give a cautionary instructionregarding the limited purposes for whichsuch testimony is proffered if the accusedinfringer requests the instruction. Thecourt should also ensure that the instruc-tions fully explain the need to apportionthe ultimate royalty award to the incre-mental value of the patented feature fromthe overall product. As to the first, whileD–Link did ask for a generic instructionon the EMVR, it did not ask for an in-struction specifically referencing the li-censes or the testimony relating theretoabout which it objected. On the second,while the court told the jury about theGeorgia–Pacific factors—which do takethe concepts of apportionment into accountto some extent—it did not separately cau-tion the jury about the importance of ap-portionment.5 As explained in Section B.2below, we need not determine whether D–Link preserved its objections to these in-structions or, if it did, whether it wasprejudiced by the instructions actually giv-en on these issues, because we vacate thedamages award for other reasons.

[32] As noted, D–Link also argues thatthe district court prejudicially erred byallowing Ericsson’s counsel to referencethe total cost of a laptop when discussingthe requested royalty rate. We find thatD–Link waived this argument. D–Link’scontinuing objection only applied to Erics-son’s expert’s reference to the prior licens-es. See J.A. 1437–38 at 4:37–5:13. No-where in the record does D–Link object tocounsel’s reference to the market value ofa laptop at trial. In fact, D–Link actuallyreferred to the value of its own end prod-ucts on crossexamination. And D–Link

4. Because D–Link does not challenge themethodology used by Ericsson’s damages ex-pert, we need not consider the propriety ofhis apportionment analysis. See Daubert v.Merrell Dow Pharm., Inc., 509 U.S. 579, 597,113 S.Ct. 2786, 125 L.Ed.2d 469 (1993).

5. While factors 9 and 13 of the Georgia–Pacif-ic factors allude to apportionment concepts,we believe a separate instruction culled fromGarretson would be preferable in future cases.


failed to raise this issue in its post-trialmotions. See J.A. 1332 at 37:22–38:11.We see no prejudice or injustice thatwould require us to address this issue forthe first time on appeal and we thereforedecline to do so. See Novo Nordisk A/S v.Becton Dickinson & Co., 304 F.3d 1216,1220 (Fed.Cir.2012) (‘‘Although appellatetribunals are not prohibited from takingremedial action when it is apparent thatprejudice or unfairness entered the trialand the interest of justice requires, ‘coun-sel for the defense cannot as a rule remainsilent, interpose no objections, and after averdict has been returned seize for thefirst time on the point that the commentsto the jury were prejudicial.’ ’’ (quotingUnited States v. Socony–Vacuum Oil Co.,310 U.S. 150, 238–39, 60 S.Ct. 811, 84L.Ed. 1129 (1940))).

2. The District Court’s RANDJury Instruction

[33] Because Ericsson was obligated tolicense the patents at issue on RANDterms, D–Link asked the district court toinstruct the jury on that RAND obligation.Among other things, D–Link requestedthat the district court instruct the juryregarding the dangers of patent hold-upand royalty stacking in RAND-related con-texts. Rather than adopt the language D–Link proposed, the district court grantedD–Link’s request only in part. The courtadded a sixteenth factor to the 15 Geor-gia–Pacific factors on which it instructedthe jury, telling the jury that it ‘‘mayconsider TTT Ericsson’s obligation to li-cense its technology on RAND terms.’’J.A. 226. After the jury returned its in-fringement verdict and assigned damages,the district court denied D–Link’s motionsfor JMOL and a new trial based on its

failure to provide more detailed instruc-tions on RAND issues.

On appeal, D-link argues that enforcingRAND commitments is critical to preserv-ing the benefits of standards and must beconsidered in any damages award. Ac-cording to D–Link, the district court re-versibly erred by giving the jury the cus-tomary Georgia–Pacific factors becausemany of those either are not applicable, ormay be misleading, in the RAND context.D–Link further contends that the districtcourt erred by refusing to instruct the juryto consider patent hold-up and royaltystacking.

Ericsson responds that the district courtdid not err in refusing to instruct the juryabout patent hold-up and royalty stackingbecause the Georgia–Pacific factors al-ready encompassed these concerns, and tothe extent they did not, the inclusion of a‘‘sixteenth’’ factor referring to Ericsson’sRAND obligations was sufficient. Accord-ing to Ericsson, a jury instruction regard-ing patent hold-up or royalty stackingwould have been inappropriate because D–Link failed to present any evidence re-garding either patent hold-up or royaltystacking to the jury. We agree with bothD–Link and Ericsson, to some extent.

This is an issue of first impression forus. To our knowledge, only three othercourts have considered the issue of appro-priate RAND royalty rates—all districtcourts. See Realtek Semiconductor, Corp.v. LSI Corp., No. C–12–3451, 2014 WL2738216, at *5–6 (N.D.Cal. June 16, 2014);In re Innovatio IP Ventures, LLC PatentLitig., No. 11 C 9308, 2013 WL 5593609(N.D.Ill. Oct. 3, 2013); Microsoft Corp. v.Motorola, Inc., No. C10–1823JLR, 2013WL 2111217 (W.D.Wash. Apr. 25, 2013).6

6. The issue in Microsoft was whether Motoro-la had breached its contractual obligation tooffer a RAND license to Microsoft. Because

the jury needed to determine whether Moto-rola’s offer letters were in fact RAND offers,the Microsoft court analyzed the facts to find

U0152867

Highlight


a. The district court’s use of theGeorgia–Pacific factors

Although we have never described theGeorgia–Pacific factors as a talisman forroyalty rate calculations, district courtsregularly turn to this 15–factor list whenfashioning their jury instructions. Indeed,courts often parrot all 15 factors to thejury, even if some of those factors clearlyare not relevant to the case at hand. And,often, damages experts resort to the fac-tors to justify urging an increase or adecrease in a royalty calculation, with littleexplanation as to why they do so, and littlereference to the facts of record. See Whit-Serve, LLC v. Computer Packages, Inc.,694 F.3d 10, 31–32 (Fed.Cir.2012) (‘‘We donot require that witnesses use any or all ofthe Georgia–Pacific factors when testify-ing about damages in patent cases. Ifthey choose to use them, however, recitingeach factor and making a conclusory re-mark about its impact on the damagescalculation before moving on does no morethan tell the jury what factors a damagesanalysis could take into consideration.’’).

In this case, the district court included all15 Georgia–Pacific factors in its damagesinstruction—over objection—without con-sidering their relevance to the record cre-ated at trial.

In a case involving RAND-encumberedpatents, many of the Georgia–Pacific fac-tors simply are not relevant; many areeven contrary to RAND principles. SeeBr. of Amici Curiae American AntitrustInstitute (‘‘AAI Br.’’) 11–20 (arguing thatthe Georgia–Pacific factors are not appro-priate for determining RAND royalties).For example, factor 4 is ‘‘[t]he licensor’sestablished policy and marketing programto maintain his patent monopoly by notlicensing others to use the invention or bygranting licenses under special conditionsdesigned to preserve that monopoly.’’Georgia–Pacific, 318 F.Supp. at 1120. Be-cause of Ericsson’s RAND commitment,however, it cannot have that kind of policyfor maintaining a patent monopoly. SeeMicrosoft, 2013 WL 2111217, at *18.Likewise, factor 5—‘‘[t]he commercial rela-

a range of appropriate values. In doing so,the court created a modified set of Georgia–Pacific factors to consider when the patentsare encumbered by RAND contracts, notingthat a number of the unmodified factors donot adequately address the RAND situation.Microsoft, 2013 WL 2111217, at *18–20.

In Innovatio, the parties asked the districtcourt to calculate the appropriate RAND roy-alty rate for a group of 802.11 SEPs. TheInnovatio court largely adopted the methodol-ogy used in Microsoft, but made a few modifi-cations. Innovatio, 2013 WL 5593609, at *6–7. The Innovatio court presented the con-cepts relevant to a RAND rate, including pat-ent hold-up, royalty stacking, and incentiviz-ing inventors to participate in the standard-setting process. The court then explainedthat, based on these considerations, the royal-ty rate: (1) must distinguish between the in-trinsic value of the technology and the valueof the standardization of that technology; (2)take into account what part of the standardthe patent actually covers; and (3) must behigh enough to ensure that innovators have

appropriate incentive to invest in future de-velopments and contribute their inventions tothe standard-setting process. Id. at *8–12.

In Realtek, the district court upheld thejury’s patent infringement damages award ina RAND patent case, explaining that the jury’saward was supported by substantial evidence.Realtek, 2014 WL 2738216, at *5–6. TheRealtek court’s jury instruction informed thejury it ‘‘should not consider LSI’s advantageresulting from the standard’s adoption, if any.However, you may consider any advantageresulting from the technology’s superiority.’’Realtek Semiconductor, Corp. v. LSI Corp.(‘‘Realtek Jury Instruction ’’), No. C–12–3451,ECF No. 267, 21 (N.D.Cal. February 10,2014). The Realtek court further instructedthe jury to use a two-step approach to deter-mine the RAND royalty rate: (1) ‘‘compar[e]the technical contribution of the two LSI pat-ents to the technical contributions of otherpatents essential to the standard’’ and (2)‘‘consider the contribution of the standard asa whole to the market value of Realtek’s prod-ucts utilizing the standard.’’ Id. at 23.


tionship between the licensor and licen-see’’—is irrelevant because Ericsson mustoffer licenses at a non-discriminatoryrate. Georgia–Pacific, 318 F.Supp. at1120; see Microsoft, 2013 WL 2111217, at*18.

[34] Several other Georgia–Pacific fac-tors would at least need to be adjusted forRAND-encumbered patents—indeed, forSEP patents generally. For example, fac-tor 8 accounts for an invention’s ‘‘currentpopularity,’’ which is likely inflated be-cause a standard requires the use of thetechnology. Georgia–Pacific, 318 F.Supp.at 1120. Factor 9—‘‘utility and advan-tages of the patented invention over theold modes or devices,’’ J.A. 225—is alsoskewed for SEPs because the technologyis used because it is essential, not neces-sarily because it is an improvement overthe prior art. Factor 10, moreover, con-siders the commercial embodiment of thelicensor, which is also irrelevant as thestandard requires the use of the technolo-gy. Other factors may also need to beadapted on a case-by-case basis dependingon the technology at issue. Consequently,the trial court must carefully consider theevidence presented in the case when craft-ing an appropriate jury instruction. Inthis case, the district court erred by in-structing the jury on multiple Georgia–Pacific factors that are not relevant, or aremisleading, on the record before it, includ-ing, at least, factors 4, 5, 8, 9, and 10 of theGeorgia–Pacific factors.7

[35] Trial courts should also considerthe patentee’s actual RAND commitmentin crafting the jury instruction. Ericssonagrees that it is under a binding obligationto license the patents at issue on theRAND terms it pledged to the IEEE. The

district court should have turned to theactual RAND commitment at issue to de-termine how to instruct the jury. In thiscase, Ericsson promised that it would‘‘grant a license under reasonable rates toan unrestricted number of applicants on aworldwide basis with reasonable terms andconditions that are demonstrably free ofunfair discrimination.’’ J.A. 17253. Rath-er than instruct the jury to consider ‘‘Er-icsson’s obligation to license its technologyon RAND terms,’’ J.A. 226, the trial courtshould have instructed the jury about Er-icsson’s actual RAND promises. ‘‘RANDterms’’ vary from case to case. A RANDcommitment limits the market value to(what the patent owner can reasonablycharge for use of) the patented technology.The court therefore must inform the jurywhat commitments have been made and ofits obligation (not just option) to take thosecommitments into account when determin-ing a royalty award.

To be clear, we do not hold that there isa modified version of the Georgia–Pacificfactors that should be used for all RAND-encumbered patents. Indeed, to the ex-tent D–Link argues that the trial courtwas required to give instructions that mir-rored the analysis in Innovatio or Micro-soft, we specifically reject that argument.See Oral Argument at 16:16, Ericsson,Inc. v. D–Link Sys., Inc., 2013–1625,available at http://oralarguments.cafc.uscourts.gov/default.aspx?fl=2013–1625.mp3 (‘‘Our argument was the following onRAND and it doesn’t rely upon any of theevidence that went in during the two hourjury wave portion. It relies upon the re-quest for instructions, basically buildingon the Innovatio decision by Judge Hold-erman and the Microsoft decision by

7. Reference to irrelevant Georgia–Pacific fac-tors would not—in most instances—be suffi-ciently prejudicial to warrant reversal. Here,

however, we find the combination of errors inthe jury instructions merit the remand weorder.


Judge RobartTTTT’’).8 We believe it un-wise to create a new set of Georgia–Pacif-ic–like factors for all cases involvingRAND-encumbered patents. Although werecognize the desire for bright line rulesand the need for district courts to startsomewhere, courts must consider the factsof record when instructing the jury andshould avoid rote reference to any particu-lar damages formula.

b. Apportionment analysis for SEPs

[36, 37] As with all patents, the royaltyrate for SEPs must be apportioned to thevalue of the patented invention. Garret-son, 111 U.S. at 121, 4 S.Ct. 291; see alsoWestinghouse Elec. & Mfg. Co. v. WagnerElec. & Mfg. Co., 225 U.S. 604, 617, 32S.Ct. 691, 56 L.Ed. 1222 (1912) (‘‘[Plaintiff]was only entitled to recover such part ofthe commingled profits as was attributableto the use of its invention.’’). When deal-ing with SEPs, there are two special ap-portionment issues that arise. First, thepatented feature must be apportioned fromall of the unpatented features reflected inthe standard. Second, the patentee’s roy-alty must be premised on the value of thepatented feature, not any value added bythe standard’s adoption of the patentedtechnology. These steps are necessary toensure that the royalty award is based onthe incremental value that the patentedinvention adds to the product, not anyvalue added by the standardization of thattechnology.9

Just like modern electronic devices,technological standards include multipletechnologies. We know that patents oftenclaim only small portions of multi-compo-nent products and we have precedentwhich covers apportionment of damages inthose situations. See, e.g., Garretson, 111U.S. at 121, 4 S.Ct. 291; Uniloc, 632 F.3dat 1318; Lucent, 580 F.3d at 1336. Simi-larly, SEPs can, and, often do, claim onlylimited aspects of the overall standard.

For example, the 802.11 standard en-compasses numerous technologies to en-able devices to communicate with each oth-er via wireless network connection. Thisincludes, among many other things, tech-nologies on link establishment, securityprotocols, error control, and flow control.By way of example, the 8568 patent, atbest, only covers the ability of the systemto prioritize time-sensitive payloads by in-forming the system what type of data is ineach transmission. This is only a smallaspect of the 802.11(n) standard. Indeed,based on the record in this case, it isundisputed that some programs do noteven take advantage of this 802.11(n) stan-dard capability. The 8215 patent, more-over, at best covers the ability to senddifferent feedback response types. Again,based on the undisputed record, some802.11(n) standard products do not usemore than one type of feedback message.

[38] Just as we apportion damages fora patent that covers a small part of adevice, we must also apportion damagesfor SEPs that cover only a small part of a

8. We express no opinion on the methodolo-gies employed in these district court cases—which may yet come before this court—or ontheir applications to the facts at issue there.The facts in those cases, and the decision-makers involved, differ from those at issuehere. We address only the record before usand what a jury must be instructed whenRAND-encumbered patents are at issue andthe jury is asked to set a RAND royalty rate.

9. As we recognized in VirnetX, these tasks arenot always easy and would be difficult to dowith precision. We accept the fact that thejury should be told of its obligation to approx-imate the value added by the patented inven-tion and that a degree of uncertainty in set-ting that value is permissible. VirnetX, 767F.3d at 1328 (citing Unisplay, S.A. v. Am.Elec. Sign Co., 69 F.3d 512, 517 (Fed.Cir.1995)).


standard. In other words, a royalty awardfor a SEP must be apportioned to thevalue of the patented invention (or at leastto the approximate value thereof), not thevalue of the standard as a whole. A jurymust be instructed accordingly. Our deci-sion does not suggest that all SEPs makeup only a small part of the technology inthe standard. Indeed, if a patentee canshow that his invention makes up ‘‘theentire value of the’’ standard, an appor-tionment instruction probably would not beappropriate. Garretson, 111 U.S. at 121, 4S.Ct. 291.

[39] Turning to the value of a patent’sstandardization, we conclude that SupremeCourt precedent also requires apportion-ment of the value of the patented technolo-gy from the value of its standardization.In Garretson, the Supreme Court madeclear that, ‘‘[w]hen a patent is for an im-provement, and not for an entirely newmachine or contrivance, the patentee mustshow in what particulars his improvementhas added to the usefulness of the machineor contrivance. He must separate its re-sults distinctly from those of the otherparts, so that the benefits derived from itmay be distinctly seen and appreciated.’’Garretson, 111 U.S. at 121, 4 S.Ct. 291(emphases added). In other words, thepatent holder should only be compensatedfor the approximate incremental benefitderived from his invention.

[40] This is particularly true for SEPs.When a technology is incorporated into astandard, it is typically chosen from amongdifferent options. Once incorporated andwidely adopted, that technology is not al-ways used because it is the best or theonly option; it is used because its use isnecessary to comply with the standard. Inother words, widespread adoption of stan-dard essential technology is not entirelyindicative of the added usefulness of aninnovation over the prior art. Id. This is

not meant to imply that SEPs never claimvaluable technological contributions. Wemerely hold that the royalty for SEPsshould reflect the approximate value ofthat technological contribution, not the val-ue of its widespread adoption due to stan-dardization.

Because SEP holders should only becompensated for the added benefit of theirinventions, the jury must be told to differ-entiate the added benefit from any valuethe innovation gains because it has becomestandard essential. Although the jury, asthe fact finder, should determine the ap-propriate value for that added benefit andmay do so with some level of imprecision,we conclude that they must be told toconsider the difference between the addedvalue of the technological invention andthe added value of that invention’s stan-dardization. Indeed, Ericsson admitted atoral argument that the value of standard-ization should not be incorporated into theroyalty award. Oral Argument at 55:25(‘‘Q: You agree that it is error to allow[the jury] to include the value from thestandardization? A: In the rate, not inthe baseTTTT The rate must be attribut-able to the value of the invention.’’). Byway of example, the Realtek court instruct-ed the jury that it ‘‘should not considerLSI’s advantage resulting from the stan-dard’s adoption, if any. However, youmay consider any advantage resultingfrom the technology’s superiority.’’ Real-tek Jury Instruction, ECF No. 267, 21.

c. Instructions on patent hold-up and royalty stacking

D–Link argues that the jury should havebeen instructed on the concepts of patenthold-up and royalty stacking because itargues that the jury should know the mis-chief that can occur if RAND royalty ratesare set too high. Many of the amicusbriefs echo D–Link’s concerns. See AAI


Br. 4–9; Br. of Amici Curiae Cisco Sys.,Inc., et al. (‘‘Cisco Br.’’) 14–19; Br. ofAmici Curiae Broadcom Corp., et al.(‘‘Broadcom Br.’’) 10–14.

[41] In deciding whether to instructthe jury on patent hold-up and royaltystacking, again, we emphasize that the dis-trict court must consider the evidence onthe record before it. The district courtneed not instruct the jury on hold-up orstacking unless the accused infringer pres-ents actual evidence of hold-up or stacking.Certainly something more than a generalargument that these phenomena are possi-bilities is necessary. Indeed, ‘‘a courtshould not instruct on a proposition of lawabout which there is no competent evi-dence.’’ See Nestier Corp. v. MenashaCorp.–Lewisystems Div., 739 F.2d 1576,1579–80 (Fed.Cir.1984); see also Br. ofAmici Curiae Nokia Corp., et al. (‘‘NokiaBr.’’) 9–12. Depending on the record, ref-erence to such potential dangers may beneither necessary nor appropriate.

In this case, we agree with the districtcourt that D–Link failed to provide evi-dence of patent hold-up and royalty stack-ing sufficient to warrant a jury instruction.JMOL Order, 2013 WL 4046225, at *25–26(‘‘Defendants failed to present any evi-dence of actual hold-up or royalty stack-ing.’’ (emphasis in original)). If D–Linkhad provided evidence that Ericsson start-ed requesting higher royalty rates afterthe adoption of the 802.11(n) standard, thecourt could have addressed it by instruct-ing the jury on patent hold-up or, perhaps,setting the hypothetical negotiation datebefore the adoption of the standard.10 D–

Link, however, failed to provide any suchevidence. Absent evidence that Ericssonused its SEPs to demand higher royaltiesfrom standard-compliant companies, wesee no error in the district court’s refusalto instruct the jury on patent hold-up or toadjust the instructions expressly to takepatent hold-up into account. Indeed, asnoted above, the court found that Ericssoncomplied with its RAND obligations anddid not demand an unreasonable royaltyfor use of its technology.

A jury, moreover, need not be instructedregarding royalty stacking unless there isactual evidence of stacking. The mere factthat thousands of patents are declared tobe essential to a standard does not meanthat a standard-compliant company willnecessarily have to pay a royalty to eachSEP holder. In this case, D–Link’s expert‘‘never even attempted to determine theactual amount of royalties Defendants arecurrently paying for 802.11 patents.’’JMOL Order, 2013 WL 4046225, at *18.In other words, D–Link failed to comeforward with any evidence of other licens-es it has taken on Wi–Fi essential patentsor royalty demands on its Wi–Fi enabledproducts. Because D–Link failed to pro-vide any evidence of actual royalty stack-ing, the district court properly refused toinstruct the jury on royalty stacking.

We therefore hold that the district courtdid not err by refusing to instruct the juryon the general concepts of patent hold-upand royalty stacking.

* * *

10. One amicus suggests that the jury alwaysshould be told to place the date of the hypo-thetical negotiation as of the date of the adop-tion of the standard (if that date predates theinfringement) so as to discount any value add-ed by the standardization. See, e.g., AAI Br.13–16; see also Microsoft, 2013 WL 2111217,at *19 (‘‘[T]he parties to a hypothetical negoti-

ation under a RAND commitment would con-sider alternatives that could have been writ-ten into the standard instead of the patentedtechnology.’’). D–Link did not request anysuch instruction, however. Accordingly, wedo not address whether shifting the timing ofthe hypothetical negotiation is either appro-priate or necessary.


[42] In sum, we hold that, in all cases,a district court must instruct the jury onlyon factors that are relevant to the specificcase at issue. There is no Georgia–Pacif-ic–like list of factors that district courtscan parrot for every case involving RAND-encumbered patents. The court should in-struct the jury on the actual RAND com-mitment at issue and must be cautious notto instruct the jury on any factors that arenot relevant to the record developed attrial. We further hold that district courtsmust make clear to the jury that anyroyalty award must be based on the incre-mental value of the invention, not the valueof the standard as a whole or any in-creased value the patented feature gainsfrom its inclusion in the standard. Wealso conclude that, if an accused infringerwants an instruction on patent hold-up androyalty stacking, it must provide evidenceon the record of patent hold-up and royal-ty stacking in relation to both the RANDcommitment at issue and the specific tech-nology referenced therein.

As explained above, in this case, we findthat the district court committed legal er-ror in its jury instruction by: (1) failing toinstruct the jury adequately regarding Er-icsson’s actual RAND commitment; (2)failing to instruct the jury that any royaltyfor the patented technology must be ap-portioned from the value of the standardas a whole; and (3) failing to instruct thejury that the RAND royalty rate must bebased on the value of the invention, notany value added by the standardization ofthat invention—while instructing the juryto consider irrelevant Georgia–Pacific fac-tors. We think that these errors collec-tively constitute prejudicial error. SeeEviron Prods., Inc. v. Furon Co., 215 F.3d1261, 1265 (Fed.Cir.2000) (‘‘Prejudicial er-ror is an error that, in the words of theFederal Rules of Civil Procedure, ‘appearsto the court inconsistent with substantialjustice.’ ’’ (quoting Fed.R.Civ.P. 61)). We

therefore vacate the jury’s damages awardand remand for further proceedings con-sistent with this opinion. On remand, thecourt should also be careful to assure thatthe jury is properly instructed on the ap-portionment principles laid out in Garret-son and on the proper evidentiary value oflicenses tied to the entire value of a multi-component product. Because we vacatethe jury’s damages award, moreover, wealso vacate the court’s ongoing royaltyaward.

C. The Dell Agency Issue

[43] Finally, Dell argues that the dis-trict court erred by granting summaryjudgment on Dell’s claim that it was li-censed under the MPA to practice theclaims asserted against it. Because theMPA says it is governed by New York law,both Ericsson and Dell agree that NewYork agency law governs. The partiesfurther agreed that, in order for Dell tohave a license to practice the patents atissue based on the MPA, LM Ericsson, theparent company, must have been acting asan agent of its subsidiary, Ericsson AB,when it filed this lawsuit. The districtcourt granted summary judgment becauseit found that this agency relationship didnot exist as a matter of law.

The Fifth Circuit reviews summaryjudgment decisions de novo. UnitedStates v. Caremark, Inc., 634 F.3d 808, 814(5th Cir.2011). Summary judgment is ap-propriate if, in viewing the evidence in alight most favorable to the non-movingparty, the court finds that ‘‘there is nogenuine dispute as to any material fact andthe movant is entitled to judgment as amatter of law.’’ Fed.R.Civ.P. 56(a).

[44–47] The existence of an agency re-lationship—where one party has legal au-thority to act for another—is a mixedquestion of law and fact. See Cabrera v.


Jakabovitz, 24 F.3d 372, 385–86 (2d Cir.1994) (applying New York law). In orderto establish an agency relationship, thefacts must show that: (1) the principalmanifested intent to grant authority to theagent, and (2) the agent agreed or consent-ed to the agency relationship. See Com-mercial Union Ins. Co. v. Alitalia Air-lines, S.p.A., 347 F.3d 448, 462 (2d Cir.2003) (citations omitted) (applying NewYork law). Further, the principal mustretain control and direction over key as-pects of the agent’s actions. See In reShulman Transport Enters., Inc., 744 F.2d293, 295 (2d Cir.1984) (applying New Yorklaw); Meese v. Miller, 79 A.D.2d 237, 436N.Y.S.2d 496, 499 (N.Y.App.Div.1981). Aprincipal cannot, moreover, grant authorityto an agent if the principal does not itselfpossess the power granted. See MouawadNat’l Co. v. Lazare Kaplan Int’l Inc., 476F.Supp.2d 414, 423 (S.D.N.Y.2007) (apply-ing New York Law).

On appeal, Dell argues that it presentedsufficient evidence that Ericsson AB di-rected LM Ericsson to sue Dell, pointingto the fact that, even though the patentswere assigned to LM Ericsson, the inven-tors of the patents were primarily Erics-son AB employees. Dell further contendsthat it presented evidence that EricssonAB maintained control over key aspects ofthis litigation. Finally, Dell asserts thatthe district court improperly made factualinferences regarding the strength of Dell’sarguments in favor of Ericsson, pointing toan alleged decision to sue Dell made by anEricsson AB employee.

Ericsson responds that there is no mate-rial dispute of fact that LM Ericsson—theparent company and patent owner—is notan agent of Ericsson AB—the subsidiarycompany and signatory to the MPA. Erics-son asserts that Ericsson AB never hadauthority to sue Dell for infringement ofthese patents on its own. Ericsson insists

that this means the alleged principal neverhad the authority it allegedly granted.According to Ericsson, even if an EricssonAB employee helped make the decision tofile the law suit, there is no evidence thathe had the authority to file those suits.

We agree with Ericsson that Dell hasfailed to raise genuine issues of materialfact regarding whether LM Ericsson is anagent of Ericsson AB. It is undisputed thatLM Ericsson is the owner of the patents insuit. As the owner, it is LM Ericsson thathas the authority to sue for infringement;no other entity need grant it the authorityto sue. See 35 U.S.C. § 281 (‘‘A patenteeshall have remedy by civil action for in-fringement of his patent.’’). Even assum-ing that an Ericsson AB employee sug-gested suing Dell for infringement, LMEricsson indisputably had that authorityprior to any suggestion. See Mouawad,476 F.Supp.2d at 423 (‘‘[T]he principal it-self must possess the power that it isattempting to confer on the agent.’’ (citing3 Am.Jur.2d Agency § 9 (2014))). BecauseLM Ericsson is not a signatory to theMPA, any license Dell might have thereun-der does not excuse any acts of infringe-ment involving the patents in suit.

For the foregoing reasons, we affirm thedistrict court’s conclusion that Dell doesnot, as a matter of law, have a license topractice the patents at issue under theMPA.

III. CONCLUSION

For the foregoing reasons, we affirm theinfringement findings relating to the 8568and 8215 patents, but reverse the infringe-ment finding with respect to the 8625 pat-ent. We also affirm the jury’s finding thatthe 8625 patent was not invalid over thePetras reference. We vacate the jury’sdamages award and the ongoing royaltyaward and remand for proceedings consis-tent with this decision. With respect to


Dell’s appeal, we affirm the district court’sgrant of summary judgment to Ericssonon Dell’s license defense.

Accordingly, we affirm-in-part, reverse-in-part, vacate-in-part, and remand for fur-ther proceedings.

AFFIRMED IN PART, REVERSEDIN PART, VACATED IN PART, ANDREMANDED

Opinion dissenting in part filed byCircuit Judge TARANTO.

TARANTO, Circuit Judge, dissenting-in-part.

I join all of the court’s opinion exceptpart II.A.2, which upholds the judgment ofinfringement of the 8215 patent. I con-clude that the district court incorrectlyconstrued the 8215 patent’s claim languagepertaining to the invention’s message field.And there is no infringement under theconstruction that I think is correct. Onthis one issue, I respectfully dissent.

The dispute over the proper construc-tion of claim 1 of the 8215 patent involvestwo related issues. The first is whetherthe phrase ‘‘responsive to the receivingstep, constructing a message field for asecond data unit, said message field includ-ing a type identifier field’’ requires thatthe device ‘‘select’’ a message type depend-ing on the received data—which requiresthat it have at least two message-typeoptions it can select from. The second iswhether the message field must be con-structed specifically to minimize either thesize or the quantity of feedback responses.It is undisputed that if the claims are readto require either selection or minimization,the accused devices do not infringe.

Selecting. I begin with the claim lan-guage. See Phillips v. AWH Corp., 415F.3d 1303, 1312 (Fed.Cir.2005) (en banc).Claim 1 requires the act of ‘‘constructing amessage field’’ to be ‘‘responsive to the

receiving step’’—that is, responsive to ‘‘re-ceiving said plurality of first data units.’’8215 Patent, col. 10, lines 23–26. At aminimum, the ‘‘responsive’’ language al-lows, indeed suggests, selection or choiceabout what to include in the constructedmessage field, based on potentially varyingcharacteristics of the data received. Andthe specification and Ericsson’s own asser-tions during litigation not only confirm therequirement of choice but make clear whatthe choice is: it is a choice from a range ofpossible message types.

Claims must be interpreted in the con-text of the patent as a whole, and thespecification is central to performing theinterpretive task. See Phillips, 415 F.3dat 1321 (‘‘[T]he specification is the singlebest guide to the meaning of a disputedterm TTTT’’) (internal quotation omitted).The specification of the 8215 patent perva-sively describes what Ericsson had invent-ed as optimizing feedback responsesthrough the receiver’s choice among feed-back-response message types—specificallyto minimize the size of the responses andthereby increase efficiency. Even withoutregard to the more specific objective ofminimizing, the receiver’s optimizing taskrequires selecting.

Thus, the abstract describes construct-ing feedback responses ‘‘so as to optimizeperformance in accordance with certaincriteria,’’ with those criteria consisting ofminimizing size and maximizing efficiencyof the feedback responses. 8215 Patent,abstract. The summary of the inventionsimilarly teaches that the receiver con-structs feedback response data units ‘‘so asto optimize system performance,’’ with theoptimization consisting of ‘‘minimizing [ ]size’’ while ‘‘maximizing the number of [se-quence numbers]’’ included in a smaller-sized data unit. Id. col. 4, lines 48–53.Neither of the two stated optimization cri-teria may be met unless the receiver


makes choices among a plurality of mes-sage types. Further, the specification de-scribes the patent’s advance over the priorart as reducing the ‘‘waste of bandwidth’’and ‘‘unnecessary overhead’’ resultingfrom Automatic Repeat Request (ARQ)protocols that are ‘‘static in construction,’’i.e., not varying as a function of the incom-ing data. Id. col. 3, lines 46–47. And thefour technical advantages specifically enu-merated—saving bandwidth, minimizingoverhead, increasing system capacity, andminimizing the number of feedback re-sponses—derive from the invention’s op-timizing of feedback responses. Id. col. 4,lines 54–62. The specification nowherediscloses any contrary embodiment.

The claim language thus suggests a re-quirement of selecting, and the specifica-tion pervasively indicates that what is re-quired is selecting among message types.Construing the claim to require receiverselecting is ‘‘the correct construction’’ be-cause it ‘‘stays true to the claim languageand most naturally aligns with the patent’sdescription of the invention.’’ RenishawPLC v. Marposs Societa’ per Azioni, 158F.3d 1243, 1250 (Fed.Cir.1998), adopted byPhillips, 415 F.3d at 1316.

Ericsson’s own statements about theinvention confirm that the receiver’schoosing among message types, in re-sponse to incoming data, is essential.During claim construction, in an appar-ent effort to prevent the additional ‘‘min-imizing’’ requirement from being readinto the claim, Ericsson conceded—in-deed, repeatedly insisted in clearterms—that selecting among messagetypes by the receiver was a required ele-ment of the invention. See J.A. 6473(‘‘[T]he invention is to build in choice atthe receiver side of a type of feedbackresponse.’’); J.A. 6475 (Given ‘‘the advan-tage TTT gained from incorporating thismessage field that allows a choice, does

the advantage of minimizing the size ornumber of feedback responses necessari-ly have to be read in when it doesn’tappear in the claim element[?]’’); J.A.6478 (‘‘[T]he invention is, as expressed inthe claims, giving the receiver a choiceand constructing a message field thathas a type identifier so it can expresswhat it has chosen to use as a formatfor communicating the packets that havebeen dropped.’’).

For those reasons, I conclude, the meth-od of claim 1 of the 8215 patent claimrequires that the receiver engage in selec-tion among message types in response toreceipt of data. It is undisputed that theaccused devices do not do so. Ericssondid not argue otherwise in its brief, and itconceded the point at oral argument, stat-ing: ‘‘[I]f the court concludes that thereceiver must make a choice, [the 8215patent is] not infringed.’’ Oral Argumentat 34:45–35:00, Ericsson, Inc. v. D–LinkSys., Inc., No. 2013–1625.

Minimizing. Claim 1 of the 8215 patentalso requires ‘‘minimizing feedback re-sponses,’’ as expressly stated in the pream-ble. 8215 Patent, col. 10, line 19. Thepresence of that language indirectly rein-forces the requirement of selection: tominimize, choices must be made. But italso adds to the selection requirement andindependently requires a judgment of non-infringement.

The ‘‘minimizing’’ language is a claimlimitation unless the fact that it appears inthe preamble makes it non-limiting, assometimes is true for preamble language.See Catalina Mktg. Int’l, Inc. v. Coolsav-ings.com, Inc., 289 F.3d 801, 808 (Fed.Cir.2002). For claim 1, however, the preamblemust be limiting, as only the preamblegives content to what the constructed mes-sage field is. See Pitney Bowes, Inc. v.Hewlett–Packard Co., 182 F.3d 1298, 1306(Fed.Cir.1999) (preamble is limiting when

1239KOEHN v. SECRETARY OF HEALTH AND HUMAN SERVICESCite as 773 F.3d 1239 (Fed. Cir. 2014)

necessary to ‘‘give life, meaning, and vitali-ty’’ to the claim) (internal quotation andcitation omitted). The claim recites con-structing a ‘‘message field,’’ but only in thepreamble does the claim indicate that themessage field is to be understood as a‘‘feedback response[ ] in an ARQ protocol.’’8215 Patent, col. 10, lines 19–20, 24–27.The phrase ‘‘feedback response,’’ appear-ing nowhere but the preamble, is crucial tomaking sense of the rest of the claim. SeeGriffin v. Bertina, 285 F.3d 1029, 1033(Fed.Cir.2002) (holding as limiting a meth-od claim’s preamble reciting ‘‘diagnosingan increased risk for thrombosis,’’ whereclaim covering nucleic-acid tests performedon a ‘‘test subject’’ was meaningless with-out understanding that ‘‘[d]iagnosis is TTT

the essence of [the] invention’’).

The construction proposed by Ericssonand adopted by the district court implicitlyrecognizes that the preamble is limiting.The adopted construction of the ‘‘respon-sive’’ step requires ‘‘including a field thatidentifies the message type of the feedbackresponse message.’’ Ericsson, Inc. v. D–Link Corp., No. 6:10–cv–473, 2013 WL949378, at *5 (E.D.Tex. Mar. 8, 2013) (em-phasis added). The phrase ‘‘feedback re-sponse,’’ in the adopted claim construction,comes from nowhere in the claim exceptthe preamble. Where the district courterred was in parsing the preamble to in-clude only one phrase as limiting—‘‘feed-back responses’’—while excluding theword ‘‘minimizing’’ that appears immedi-ately before the phrase ‘‘feedback respons-es.’’ I see no sound basis for that distinc-tion.

Indeed, the specification, as describedabove, shows that the invention is centrallyabout minimizing such responses-whichcan be done either by minimizing the sizeof individual response messages (statusprotocol data units) or by minimizing thenumber of such messages by packing more

information, i.e., sequence numbers, intomessages of fixed size. 8215 Patent, col. 4,lines 49–54 (summary of the inventionidentifying precisely those two ways of‘‘optimiz[ing] system performance’’). Er-icsson’s only specific argument against the‘‘minimizing’’ construction is to note thetwo ways of minimizing the specificationidentifies. Ericsson Br. at 39. But that isnot an argument against D–Link’s pro-posed construction—which covers both ofthose ways of minimizing. J.A. 5084 (‘‘se-lected from multiple available feedback re-sponses in order to minimize the size ornumber of feedback responses’’). I con-clude, therefore, that it was error for thedistrict court to reject D–Link’s ‘‘minimiz-ing’’ claim construction.

It is undisputed that if claim 1 requires‘‘minimizing,’’ the accused devices do notinfringe. Ericsson never argues otherwisein its brief. Ericsson Br. at 38–40. Thisis an additional ground for reversing thejudgment of infringement of the 8215 pat-ent, and requiring judgment of non-in-fringement instead.

,

Cheryl KOEHN, as Mother and NextFriend of, Vanessia Koehn,

Petitioners–Appellants,

v.

SECRETARY OF HEALTH ANDHUMAN SERVICES, Re-

spondent–Appellee.

No. 2014–5054.

United States Court of Appeals,Federal Circuit.

Dec. 4, 2014.

Background: Claimant who alleged thathuman papillomavirus (HPV) vaccination

ericsson, inc. v. d link systems, inc. 1201 · ericsson, inc. v. d link systems, inc. 1201 cite as...

Documents