Analysis, Breaking, Graphics, Hardware, Mobile Computing, VR World

Nvidia Sues Samsung and Qualcomm For Alleged Patent Infringement

Nvidia GPU Logo Nvidia GPU Logo

In a blog, Nvidia (NASDAQ:NVDAstated that it has filed patent infringement complaints with the US District Court in Delaware (where most US companies are incorporated) as well as with the International Trade Comission regarding Samsung (KRX:005935) and Qualcomm’s (NASDAQ:QCOM) infringement of its patents (You can view the complaints here (ITC) and here (Delaware)) .

Nvidia is alleging that Samsung’s devices that use Qualcomm’s chips are infringing upon Nvidia’s own technologies that have been patented. Not just that, but by filing a complaint with the ITC, Nvidia is seeking that such devices that infringe upon these patents be banned from importation and sale within the United States.

The devices that Nvidia claims infringe upon their patents include the Galaxy Note 4, Galaxy Note Edge, Galaxy S5, Galaxy Note 3, and Galaxy S4 as well as the Galaxy Tab S and Galaxy Note Pro. Nvidia claims seven different patents were violated by Samsung’s Exynos processors as well as Qualcomm’s Snapdragon processors. The Qualcomm Snapdragon processors that are claimed to infringe upon these patents include the Snapdragon S4 (using the Adreno 225), Snapdragon 400 (using the Adreno 305), Snapdragon 600 (using the Adreno 320), Snapdragon 800 and 801 (using the Adreno 330), and Snapdragon 805 (using the Adreno 420).

A careful reading Nvidia’s patent infringement complaint shows that Nvidia has made a clear decision here to go after Samsung and Qualcomm. Obviously, this hurts Nvidia’s chances of ever landing a design win inside of Samsung, but by going after Samsung and Qualcomm, the company goes after the biggest seller of Android smartphones and the biggest seller of Android smartphone SoCs. Additionally, the complaint states that Nvidia is claiming that Samsung infringes on their GPU patents by saying Samsung uses Mali and PowerVR graphics, which aren’t actually theirs but rather graphics they license from ARM and Imagination technologies, however Nvidia has chosen not to include them in the suit.

Qualcomm and Samsung are not GPU pioneers or innovators in graphics technology. Qualcomm dominates the global market for smartphone applications processors, with a market share exceeding 50%, and is also a leader in tablet application processors. Samsung leads the global market in sales of smartphones, selling about twice as many as its nearest competitor, and is also a global leader in the sales of tablet computers, nearly tripling its market share over the past two years. Many of Samsung’s smartphones and tablet computers are powered by mobile processors supplied by Qualcomm, which use GPUs commercially known as
“Adreno.” Other smartphones and tablets sold by Samsung use GPUs commercially known as “Mali” or “PowerVR.” All of these products infringe the Asserted Patents. The market success of Qualcomm and Samsung in these areas is built on the back of NVIDIA’s pioneering graphics technology, and Qualcomm and Samsung continue to release new products using NVIDIA’s technology.

The patents in question,  6,198,4886,992,667,  7,038,685, 7,015,9136,697,0637,209,140 and 6,690,372 all pertain in one way or another to graphics patents and it appears quite clear that Nvidia is attacking the two biggest players in Tegra’s own mobile SoC space.

The claims Nvidia makes are all utility patents and they go into deep technical detail about certain graphics processes that Nvidia believes Qualcomm and Samsung are infringing upon. The reality of the situation is that Nvidia could theoretically sue Apple, Imagination Technologies, ARM and virtually any smartphone or SoC manufacturer in the world for patent infringement based upon these claims. But Nvidia is choosing not to instead going after Samsung and Qualcomm, and really it seems like Samsung is just collateral damage here, but it could also be that Nvidia wants to weaken Samsung in order to make room for other vendors that might use their chips like Tegra K1. It could also be that Nvidia may just want to generate IP licensing revenue from all of Samsung’s billions of dollars in sales of Galaxy devices and that could be good enough for them. Until we get some comment from Qualcomm and this thing actually goes to court, we probably won’t know the outcome of this suit. Once again, the tech industry continues to sue one another for patent infringement.

  • Roberto Tomás

    it honestly looks like american companies are silently partnering/working together to try to overcome Samsung, as if it is because Samsung’s position in both electronics and appliances is so strong that it is the only shot they have at being around 5 years from now.

    • force

      No offense berto, but what country do you think nvidia comes from? Or was that just some bad joke?

      • 32Devlin1

        Thing is Nvidia was not did not develop the first GPU if so why are they not sueing AMD or Intel for that matter,there was GPUs before Nvidia came along.

        • nobodyspecial

          They already sued Intel and won…LOL. Nuff said. Intel pays NV every year (well 66mil quarterly really) until they likely re-up again in 2016/2017. It’s either that or another suit probably, ending with another Intel loss. AMD/NV have gobbled up a LOT of companies either directly or indirectly over the years and own a lions share of gpu patents (likely keeping them from going after each other, or it’s just simply massive cross-lics that cover them from directly suing each other).

          Nvidia or AMD have created most of the gpu patents running everything on your pc/mobile device in the last decade or two. That is the problem everyone else will have going forward (with NV having the most I’d say at 7000 gpu patents and directly affecting current mobile/Intel stuff). Now that everyone is trying to do desktop gpu type stuff on mobile etc, they’ll have to pay up in some way shape or form or get sued and lose like Intel. I don’t believe you can make a gpu today without crossing NV/AMD lines, it’s just a matter of how many patents you infringe while doing it.

          If AMD has any worthy patents I think we’ll see a suit from them after NV proves it works and lays all the groundwork for an AMD case. Just like NV is now quoting the Intel case in the new case against mobile makers. When Intel quit letting NV make chipsets, their gpu agreement went down the drain, thus the lawsuit and Intel loss (1.5B).

          The larger problem here is everyone is aware they’ve been infringing (so willful infringement also!) and this is larger than just what Intel did via welching on chipsets which caused their violation of the gpu agreement. I’m fairly certain this is why NV WISELY chose a JURY not a judge to decide their case. They will use the 1.5B Intel case and say, umm these guys in mobile did it ON PURPOSE for YEARS and are still doing it now, so please award us even more massive damages. We know how jury members love to stick it to big companies, especially a foreign one like samsung making ~30B+ while screwing an american company. Once Samsung/Qcom lose there will be no need to sue anyone else, as they’ll know the outcome and license gpu ip and make some agreement to pay for the past (or get sued I guess).

          Apple will fold in some way shape or form once these two lose, and the rest will follow while only ARM themselves will escape probably due to smaller profits and lic usage of ARM IP by NV. NV will win something here, it’s just a matter of how much and from who, and in what form (usage of X number of NV chips ir IP lic for years, or cash, or a combo, or ??? who knows). I find it comic you’re not aware Intel pays NV for exactly this tech in their cpus and that suit is already over and the suit isn’t even that old (2011, and is in every NV yearly financial report). It’s a 6 year agreement.

          http://www.nvidia.fr/object/nvidia-intel-licensing-fees-20110111-fr.html

    • Daniel

      Samsung phones aren’t even that good anymore. Sure, the specs are, but the quality isn’t. Looking at Nvidia’s patents, they’re extremely specific… not like Apple’s “Rounded corner” patent. I don’t really see this as a plot by American companies, but more as an aggressive approach by Nvidia.. Nvidia came out of no where in the mobile space and has progressed very quickly, but hasn’t been able to get inside many phones. If anything, I feel like Samsung and Qualcomm had a plot to keep them out of the arena.. Nvidia is a fairly big threat.

    • nobodyspecial

      What you’re not understanding here is mobile is now trying to run desktop style graphics in games etc (never before needed this type of stuff). They don’t own any of the patents that have been made over the last 15yrs on desktop gpus. NV/AMD own the vast majority of those and you’re going to have to pay one or the other to play in this new GAMING ON MOBILE world that is essentially stealing from the desktop world of the last decade or two. Note the patent dates on these. Mobile just started to do the stuff of desktops/notebooks/consoles in the last 2yrs, so time to pay up for what the owners of this tech created & patented YEARS ago, as tech for powerhouse pc’s & gaming platforms progressed.

      Just like qcom gets paid for modem patents, NV/AMD will be paid (or sued) for GPU patents. Intel is paying 1.5B for a reason as we speak and they actually WERE in the desktop world at points over the years (and are now integrated etc). They don’t own many gpu patents either obviously.

      The only question here is how many billions NV will end up with over the next 3 years or so of the suit (or settlements), and if they end up with license fees for life after it all from everyone else in the picture NOT yet named in litigation but easily can be added. I’m guessing they took two top moneymakers and hope to get the rest to settle up for lics etc (apple, img.l, arm etc). Why pick fights with everyone when you can prove the point with two that will have big ticket payoffs showing the rest with a simple nod/wink to the rest meaning something like, “umm, you get the point right guys?”… 😉

      I’m guessing the stock doubles in 3yrs tops, and more as desktop gpus take over the mobile world providing either lic fees or just winning designs period because if you don’t license, NV will beat you (amd too, if they get in at some point, both KNOW GPU gaming graphics inside out, the rest have to catch up or always be behind in tech). It’s just too hard to avoid using the patents these two (amd/nv) own to do anything FUN or Professional on a PC at this point as they’ve been alone at the top pioneering the tech for a good 15yrs (longer as others all died or got absorbed).

      OR, if Jen wants to sell, they could get bought 🙂 But I doubt that, he wants to grind Intel into the ground at some point (as best he can) with their own in house ARM cpus. Revenge for the chipsets being killed by Intel. That is the point of making a cpu, getting games on arm/opengl (where x86 won’t be needed, no wintel needed at all), NVlink to avoid busses from Intel/amd etc (pcie, hypertransport, infiniband all supplanted by NVlink), and finally pro apps in a desktop box with NV cpu and DISCRETE gpus that will already be a great gaming box by that time. A total NV box with a multitude of FREE OS choices or Tri/Quadboot boxes having them all (steamos on ARM, linux, Android & or Chrome).

      • 32Devlin1

        By the time Nvidia comes out with any ARMs CPU that can do what you want it to AMD well kill them with there own ARMS CPU a company that knows how to make a real CPU or Intel will just brief all the OEMs not to use Nvidia’s ARMS CPU, remember Intel is alot bigger then Nvidia ever will be so that’s a dream that will never happen.

        • nobodyspecial

          ROFL @ the idea AMD can beat anyone in cpu at this point. Nvidia now spends more in R&D than AMD and AMD has a far larger product portfolio they spend that R&D on (console, server cpus etc). AMD just got an ARM license to make IN HOUSE cpus in May2014 (they are expensive), so they are already behind NV who’s got one coming with Denver next month (they expect in house in 2016, but we’ll see considering their R&D has been dropping for 4 years). The NV design team is no slouch either, chock full of expert cpu jockeys from all major cpu companies, not to mention NVlink to work with it (replacing Hypertransport/Infiniband/PCIE).

          http://www.forbes.com/sites/patrickmoorhead/2013/10/10/nvidias-mobile-custom-64-bit-arm-cpu-its-sooner-than-you-think/

          “Let’s look at Nvidia’s processor team. They have been in existence since
          2006 and have hardened multiple, “off-the-shelf” ARM cores. Unknown to
          most, Nvidia’s engineers have been working on their ARM 64-bit Denver
          for at least three years, since before the CES 2011 announcement.
          Hailing out of Portland, Oregon, the team consists of former CPU jockeys
          from Intel,
          AMD, HP, Sun, and Transmeta with experience in superscalar, OoO (out of
          order) execution design, micro-code, VLIW, hyper-threading, and
          multi-core. Does this experience and background guarantee success? No,
          but it provides the opportunity to succeed, and succeed big if you look
          at what others have accomplished.”

          These guys know what they’re doing and have experience with every part of the cpu/busses etc that is needed to field a great cpu. For the record AMD has NO ARM soc for mobile yet, while NV is on rev 5 right now, and rev6 is just 7-8 months away. So it is likely NV will be on v7 when AMD gets out their first mobile soc on ARM.

          There is nothing stopping NV from amping up volts on Denver and running with a heatsink & fan at massively higher wattage than works in a tablet today. You are mistaken. The only reason we don’t see it is there isn’t much point until you have a 64bit ARM OS (AndroidL) etc to take advantage of what you’d get in a FULL PC LIKE box today (4-32GB ram etc). I suspect they’ll wait for some apps that are created for 64bit before moving in this direction, but that won’t take long once AndroidL hits.

          Rimm was big once too…LOL. Intel profits have been dropping yearly for 3yrs from 12.94B to 9.6B or so now. They are losing 4B per year now on mobile if the current trend continues (1.1B loss last quarter, up from 930mil the quarter before that, so accelerating losses already to try to get people to use baytrail). Even with Intel basically paying EVERYONE currently to use their chips (baytrail or atom), Chromebooks stole a decent share of notebooks. The same will happen on desktops (and even more notebooks). Nothing stopping NV from putting their desktop gpus on an NVlink bus and dropping them in ARM64 boxes also.

          The only thing Intel can do to stop NV is buy them now that ARM/Android has become such a large market (if you can’t beat them, and Intel can’t, you join them, and only NV is buy-able really for Intel, AMD would be worthless to them and they can’t buy Qcom). You do realize NV is about 4x bigger than AMD right? You do realize NV owns 65% of the discrete market and a good 85% of the pro market right? NV makes 1/2bil per year while AMD keeps getting losses (lost 6B+ in the last 10yrs). NV has 4B+ in the bank, while AMD has 1B (and a few bil in debt too). You do realize AMD has laid off 1/3 of their workforce in the last 3-4yrs (tons of engineers) also right? This isn’t AMD hate, but rather realizing the financial, market, and employment data at AMD and recognizing the costs all that info implies.

          With current and future price cutting that will be needed to fend off maxwell attacks AMD will again probably lose money for the next 2 quarters. Console chip sales may save one of the two, due to already being in the quarter for the MS ramp in many NEW countries for a temporary pop on xbox1, but likely break even at best for the next two overall. I could go on but you should get the point. AMD can’t do squat to NV, gpu or cpu/soc wise.

          One last point, Nvidia isn’t alone gunning for Intel, rather it is a whole ARM armada coming for WINTEL in general. The first casualties will be DirectX and it’s games first, along with consoles, then apps as 64bit ramps on android. Sheer numbers of units on the ARM side will cause devs to follow in apps, just as they have in games massively. I’m not saying WINTEL will disappear. But Intel’s competition isn’t AMD any more, rather it is ARM. Wintel’s new competition isn’t just linux and some chip, now it’s the entire ARM armada (I’d include apple devices here too), linux, steamOS, android, iOS, chromeOS and the entire ecosystems surrounding them all growing yearly. As your profits go, so does your R&D (see AMD 10yr, RIMM etc). In just 3yrs Intel’s profits have slid from $13billion to 9.6Billion. If that trend continues for another 3yrs you’ll be looking at $6B. That’s what Qcom makes right now! Samsung makes $25-35B, Apple $45B, Google 13.2B not to mention TSMC etc. GF/Samsung now have a process that can be shared so if one can’t make enough chips you can immediately ramp at the other. Fabs are closing, device makers are closing, R&D from all of them combined dwarfs Intel’s (10Bil TTM, and google almost does that alone, Amazon hits ~7.3B, never mind samsung, tsmc, apple, qcom etc). All of them are in some way contributing to competition for Wintel.

          Of course things change if either of them (Intel/MS) buys NV (goog/samsung could buy them and benefit also and I wish it was google), but until then I see NV slowly gaining as cpu/modem are no longer kings in their fields. NV’s cpus will be “good enough” for a large portion of the public (amped to 4ghz+ and ~80w may actually win a lot of benchmarks too at some point) and we know their gpus will be top notch for anything graphic related and gpu is becoming king quickly in many things. But it may not even be possible for Intel to buy if Jen really hates them (he’s said CEO of Intel or bust, and it seems they won’t do that), and of course they both have hate in the engineer side for years. No doubt Jen dislikes them for killing his chipset business already and sees the long game here as revenge in the end anyway.

          But make no mistake, all major soc players have DESKTOPS/NOTEBOOKS in mind, no doubt about that. Some of the next generation of kids may never even see a windows desktop ever if their job doesn’t show it to them. I’m in IT so I’m on a Wintel treadmill for years to come probably, but my parents/sister & her son, could easily quit WINTEL any day of the week as could many in the public at large.

          • 32Devlin1

            What has Nvdia done with there ARMS chips don’t see them in any Phones O they have that hand held game pad that no one is buying O thats right they make a few chips to put in cars but from what i seen from there last call they did not make any money from that so all they have is Video cards witch they do do a good job at but as far as CPUs they will never come close to Intel or AMD even if AMD CPUs are slower then Intel.

          • nobodyspecial

            The first few revs were just to get name recognition. NVidia is playing the LONG game here. The point was to get to maxwell, where you have the first “from the ground up” built for mobile gpu while having the gpu roads on discrete and mobile meet. The race is just now starting. Look at the 70w drop maxwell causes on the same 28nm process and think 16nm Denver+Maxwell next year. The real war is just beginning, up to this point it was just showing you were in the game and developing the ecosystem to support you once you were ready for the real war. IE, Cuda took 7yrs to get to where it is today and now owns 80% of the workstation market, controls the green 500 top 10, and is taking over supercomputers with NV cards.

            Intel is losing 4Billion+ on mobile yearly now and accelerating. You should be able to see where this is going. Intel has went from 12.9B profits down about a billion each year to now 9.6B. A this rate they’ll be at 5B profits in a few more years as ARM starts to take a bite out of desktops and higher end notebooks etc. You need to look at the bigger picture. This isn’t just about a phone. 😉 Sure NV would have loved to get into more stuff with the first few revs, but it didn’t matter as Jen Hsun is playing a LONG game just like he did with Cuda, and in this game he has a large number of partners (google on OS +app stores, Valve who hates windows/directx, device makers, sheer # of units 2b+ sold on ARM etc etc, now 51.5% of devs making GAMES for ARM, only topped by PC’s at 53%, consoles FAR behind this in dev interest, fabs amping up, some working together etc).

            The writing is on the wall for anyone willing to read it. GPU is becoming king, DirectX is losing power, Portability is king (see how fast NV has converted games, 2 weeks etc for halflife 2, trine2 etc), and Intel will have to lower prices (killing profits) to compete with ARM boxes soon. MS already doing this on their phone/tablet OS. Wintel is losing it’s dominance. I predict a large portion of the next generation of kids may never know what WINDOWS looks like if it isn’t used by them on their job. Don’t forget most could leave now, as all the majority of us do is read the web, email, facebook, twitter, casually game, and consume video/audio. All of which can be done by ARM today. Only some people like me (IT guy) will be forced to stay longer (or forever…LOL) on the WINTEL treadmill. Let’s just say I’m going to start learning more linux crap just in case though 😉

          • 32Devlin1

            Yea but what is Nvidia going to do when Intel No longer supports PCI express and goes it a lone with just there build in GPU.

          • nobodyspecial

            Sell their own cpus in a full fledged arm box with their own DISCRETE gpus running NVlink bus instead of PCIE, Hypertransport, or Infiniband. It’s purpose is to replace all three in any scenario. Doing so for Intel would be suicide vs. a gaming box that would crush their built-in cpu on the ARM side that can go either way. As AMD gets weaker Intel either has to put out a gpu that can compete with Nvidia or gaming will slowly migrate for WINTEL to ARM+Nvidia with a multitude of FREE Operating systems. That is the goal of all the arm players (not just NV).

            IE: A full fledged 64bit box (just got 64bit android piece of puzzle) with 8-32GB ram (can do this now too), huge HD, apu/cpu with huge PC like heatsink/fan, 500w+ PSU’s etc etc. Exactly like A PC box but without the say $100-150 off Intel’s fee and ZERO OS fee (saving another $100-135). Yet that machine will perform the same running NV discrete cards if desired or just the amped up socs themselves (just like Intel/amd now) running 4ghz+ etc. Apps will follow after games and already show signs of this, but will really ramp up now that they have 64bit and more room to make fully fledged PC like apps. The full assault is coming soon (right after Volta hits I’d say, as they’re working out the NVlink kinks now on supercomputers).

            http://www.extremetech.com/computing/194283-ibm-and-nvidia-will-build-two-ultra-efficient-150-petaflop-supercomputers-for-the-doe

            We may see the first full box on Maxwell, but Volta gives Google another year to polish up their 64bit OS, get some apps going on it etc. Either way, ARM is headed for your PC box and it will be supported by NV discrete cards if the user desires more than what comes on the amped up SOCS. Wintel won’t die, but there sure will be a lot of users who can switch (especially developing nations who need the free OS and $100-150 off the cpu side). It is much easier to sell a $500 box vs. $750 with both performing the same. The arm side could come with a triboot of Linux, SteamOS (porting to ARM is surely on the way from NV/Valve) and andorid. ALL FREE.

            Having said all that, I have ZERO fears any hardcore gamer will give up their discrete cards any time soon for integrated crap (from AMD or Intel) and casual gamers are slowly figuring out why you WANT discrete as NV’s quarterly reports show (records). Integrated is having ZERO effect on gpu sales and will always be behind. Even if I end up wrong on that (they’ve been claiming it for ages though), NV how has the same thing on ARM side and all the crap they need to make a box the same as any PC (64bit OS, gobs of memory etc now). The question you should be asking is, once NV makes more than 500mil profits on the ARM PC box side what does Intel do when they quit allowing their discrete gpus to run on WINTEL boxes? LOL. AMD will be gone by then or so weak they don’t matter.

            NV only makes 500mil roughly now supporting Wintel. So if ARM can get them up to a billion in profits in the next few years you could argue they should cut WINTEL off at the knees and leave them with no discrete options making the ARM box the default gamer powerhouse. Shortly after that apps will be there too if not already by then (pro apps I mean, like Adobe suite etc) and content will be made there by then too. Why pay Intel $340 for their top chips and a lic fee for windows when you don’t have to?