NVIDIA Files Complaints Against Samsung and Qualcomm for Infringing Its GPU Patents

From a purely strategic PoV, what does Nvidia lose by pursuing this?
Worst case scenario: No case to answer, they're out legal fees, and the patents aren't used for further litigation.
I can think of a few worse cases. NV win and upset Arm who now lose a bunch of income from there GPU License who no one wants to use any more due to paying extra for it. Arm cancel the CPU license NV have which kills Tegra. Unlikely but possible. Even if NV lose what if they upset ARM enough to cancel the license? Or if not cancel it not license any new CPU tech with NV.

Another option the mobile worlds stick a finger up at NV and moves towards ray tracing GPU’s faster then expected to avoid paying extra to NV. Ray tracing chips I believe wouldn't be covered by the NV patents in the case. Then the mobile world all stop doing business with NV. My seem far fetched but if the case drags out a few years wouldn't ray tracing mobile GPUs be near or here already?

Or if I really dream the courts decide in fact the PowerVR’s patents hold and predates NV. So NV own money and back payments for the past decade worth of desktop GPUs. Unlikely but who knows what will happen.
 
That is silly. There is no reason for NVIDIA to go directly after ARM and ImgTech, because these companies do not directly bring their tech to market. NVIDIA will look for graphics patent licensing or cross-licensing agreements with the major companies that actually bring graphics technology to market.

As for the notion that the mobile world will "stop doing business with NV", simply because they are trying to protect their intellectual property, again that is silly. NVIDIA's ultra mobile SoC's are good enough nowadays to win some major design wins in the consumer, automotive, and embedded space, and NVIDIA's graphics and visual computing IP (including fundamental graphics patents) is arguably second to none in the industry. In fact, both AMD and Intel already have graphics patent cross-licensing agreements with NVIDIA.

Essentially what Samsung is doing is failing to take any responsibility for the third party IP that they use in their products while placing the blame on their suppliers, while Qualcomm is simply ignoring the issue of graphics IP licensing altogether, which is why the patent litigation was initiated in the first place.
 
Last edited by a moderator:
That is silly. There is no reason for NVIDIA to go directly after ARM and ImgTech, because these companies do not directly bring their tech to market. NVIDIA will look for graphics patent licensing or cross-licensing agreements with the major companies that actually bring graphics technology to market.

As for the notion that the mobile world will "stop doing business with NV", simply because they are trying to protect their intellectual property, again that is silly. NVIDIA's ultra mobile SoC's are good enough nowadays to win some major design wins in the consumer, automotive, and embedded space, and NVIDIA's graphics and visual computing IP (including fundamental graphics patents) is arguably second to none in the industry. In fact, both AMD and Intel already have graphics patent cross-licensing agreements with NVIDIA.

Essentially what Samsung is doing is failing to take any responsibility for the third party IP that they use in their products while placing the blame on their suppliers, while Qualcomm is simply ignoring the issue of graphics IP licensing altogether, which is why the patent litigation was initiated in the first place.
I don’t see it as silly but I can agree on very unlikely if not extremely unlikely. If we are talking worse case things can spiral out of control, companies can hold grudges or refuse to do business with others.

If the courts make the royalties to high it might not be cost effective for companies to use ARM’s GPU. The way I see it is if NV win then either ARM have to lower the price for there GPU so less profits or if the price stays the same the other companies have to pay more to use ARMS’s GPU due to the extra cross licensing fees to NV. Paying extra could mean the companies now decided to go for Tegra or a different none ARM GPU to be more cost effective. Either way it could hurt ARM. Worse case is ARM response is no new CPU license to NV so Tegra dies and ARM gets a chance to re-enter the market. It might be extremely unlikely but surly it’s possible as worse case?

As for stop using NV it’s even more extremely unlikely then above but again its possible worse case that if court case spirals out of control. Last time I looked NV was under 1% market share of the mobile world. I believe they have gone over 1% now? It might have increased but if they start suing all the big players they could end up in a situation where no one wants to deal with them. NV are not an important part of the mobile world with such a tiny market share. Finding alternative to cover the 1.5% ish NV do is a possibility .

Anyway I wasn’t trying to suggest these things will happen. I was trying to point out this isn’t entirely risk free. Worse case this could back fire on NV.
 
Another option the mobile worlds stick a finger up at NV and moves towards ray tracing GPU’s faster then expected to avoid paying extra to NV. Ray tracing chips I believe wouldn't be covered by the NV patents in the case. Then the mobile world all stop doing business with NV. My seem far fetched but if the case drags out a few years wouldn't ray tracing mobile GPUs be near or here already?



Well even if we assume that the "mobile world" was some sort of hive collective and was also run by children that still wouldn't happen. The APIs and technology don't yet exist to make ray tracing a viable alternative on the desktop and far less for mobile.

Licensing IP is standard industry practice. It's not a reason to throw a hissy fit.
 
NVIDIA Files Complaints Against Samsung and Qualcomm for Infringing Its GPU P...

Anyway I wasn’t trying to suggest these things will happen. I was trying to point out this isn’t entirely risk free. Worse case this could back fire on NV.


Yes, it could backfire but the consequences are very limited. nVidia has basically no exposure to Samsung and Qualcomm. ARM probably doesn't give a hoot - their bread and butter CPU IP is unaffected and nVidia is investing heavily in bringing ARM to servers.

In the end mobile manufacturers will use nVidia's tech if it's good because consumers will demand it. nVidia have nothing to lose since like you pointed out they have no mobile market share. Besides, there's always China...
 
Ok.

Either way, if Samsung is anything like the big conglomerate I've worked for (they are), the Samsung Laptop division isn't going to care one bit about what's going on at the Samsung Mobile group. (In fact, if they are like the one I've worked for, the other divisions might be cheering right now. :D )

Samsung Laptops is probably a different legal entity, it wants to make products that people want to buy. Minor stuff like this in a remote other part of the company is no big deal. (Yes: minor stuff. This is going to be about a few $100M maximum per year. Peanuts for a division that makes tens of billions in profits per year.)

You may be right, as I've heard that Samsung is quite compartmentalized.

Still, you'd think at some point, someone higher up in the Samsung Electronics food chain might say: "OK, you sue one of our divisions, we'll mobilize every other one to fight back". I mean, what's the point of being a ridiculously huge conglomerate if the separate entities don't collaborate at all? But since Samsung doesn't use NVIDIA GPUs—why?—I guess the point is moot anyway.

I do wonder what other OEMs might think about this. Many of them are in the same situation as Samsung.
 
Well even if we assume that the "mobile world" was some sort of hive collective and was also run by children that still wouldn't happen. The APIs and technology don't yet exist to make ray tracing a viable alternative on the desktop and far less for mobile.

Licensing IP is standard industry practice. It's not a reason to throw a hissy fit.
I thought both IMG and the other ray tracing companies had the technology up for license now along with API’s. First mobile products are due this year from the other companies if no delays. EDIT: Clearly not ready for widespread use now but in a few years perhaps?
 
Last edited by a moderator:
Alexko,
Samsung is a very complicated case, consisting of over 85 companies. When Apple first sued them many made false assumptions that Samsung would dump their manufacturing or any other childish revenge against Apple. In the end did any of those theories turn out to be true?

I kept insisting back then that it's not in both Samsung's and Apple's best interest for the latter to continue manufacturing at Samsung. Funny how some sites that claimed that Samsung will crush Apple, now happily report about the 14nm deal.

To bounce back on topic: legal matters are one story and each companies interests another. No Samsung will ever hurt its interests by embargoing NV no matter where because of some legal patent dispute which seems pretty common practice in the industry.
 
Alexko,
Samsung is a very complicated case, consisting of over 85 companies. When Apple first sued them many made false assumptions that Samsung would dump their manufacturing or any other childish revenge against Apple. In the end did any of those theories turn out to be true?

No, however, Samsung did increase the cost of manufacturing to Apple. Apple then attempted to move more manufacturing away from Samsung due to the increased costs. However, they couldn't replace all the capacity at other companies and thus have had to remain with Samsung...despite the higher cost of doing business with them after the lawsuits.

So, Apple won. But Samsung got some revenge as well. And while it may have allowed Apple to win some small bit of marketshare from all of it, I'm not sure the increased cost to doing business was worth it.

Regards,
SB
 
Putting aside SoC manufacturing, Apple products use a lot less Samsung parts than they used to. Compare new and old BOMs for things like DRAMs, NANDs, and displays. But I have no idea who punishing whom in this, if either side really is.
 
Well Apple scored some major victories. Especially with the President of the USA personally siding with them, first to uphold the injunction on Samsung products and then to throw out the injunction against Apple products via presidential mandate.

Those definitely impacted Samsungs ability to compete in the short term. But in the long term they are likely to recoup some/most of those financial losses through increased prices on components.

And the short term impact to Samsung didn't put a significant dent in them taking marketshare (not sales) away from Apple in the USA. And it taught them a valuable lesson. That Obama will more than happily throw out US law in order to prop up US corporations that contributed to his campaign. Which put a stop, at least for now, in Samsung going after Apple in the courts.

Regards,
SB
 
Isn't this type of patent case practically impossible to argue in court ?

How does one confirm that a particular chip does a particular small part of graphics processing in the same way as is covered in a patent ?

There is no way it can be reversed engineered by a 3rd party. So is it you ask the defendant how they do something, and ask a 3rd party "expert" or range of experts to determine if this violates the terms of the patent.

That requires

a) The defendant telling you the truth, with very little way of proving that it is the truth.

b) Several (at least one from each side) experts in the fine details of GPU IP design and implementation. One assumes anyone with that level of expertise is already employed in the industry, so how unbiased can they be, after all there can only be 1/2 a dozen of so GPU IP design houses, and most of them will have some interest in the outcome.

Anyone not employed in the industry at a very advanced level of involvement couldn't possibly make a reasonable determination, IMO.
 
No, however, Samsung did increase the cost of manufacturing to Apple. Apple then attempted to move more manufacturing away from Samsung due to the increased costs. However, they couldn't replace all the capacity at other companies and thus have had to remain with Samsung...despite the higher cost of doing business with them after the lawsuits.

So, Apple won. But Samsung got some revenge as well. And while it may have allowed Apple to win some small bit of marketshare from all of it, I'm not sure the increased cost to doing business was worth it.

Regards,
SB

Yes but that kind of (let me call it) "love affair" had absolutely NOTHING to do with their legal disputes over rectangle bullshit patents and not. That's the real point here.
 
Isn't this type of patent case practically impossible to argue in court ?

How does one confirm that a particular chip does a particular small part of graphics processing in the same way as is covered in a patent ?

There is no way it can be reversed engineered by a 3rd party. So is it you ask the defendant how they do something, and ask a 3rd party "expert" or range of experts to determine if this violates the terms of the patent.

That requires

a) The defendant telling you the truth, with very little way of proving that it is the truth.

b) Several (at least one from each side) experts in the fine details of GPU IP design and implementation. One assumes anyone with that level of expertise is already employed in the industry, so how unbiased can they be, after all there can only be 1/2 a dozen of so GPU IP design houses, and most of them will have some interest in the outcome.

Anyone not employed in the industry at a very advanced level of involvement couldn't possibly make a reasonable determination, IMO.

I'd say that's one one of the reasons why NVIDIA actually targetted Samsung f.e. in the first place.
 
Isn't this type of patent case practically impossible to argue in court ? How does one confirm that a particular chip does a particular small part of graphics processing in the same way as is covered in a patent ?
It's actually quite simple: during the discovery phase of a patent case, an independent third party (basically, your patent lawyers) will get access to the emails/design documents that are related to the technology in question. They are under the obligation to not disclose anything to their client.

The defending party will fight in court to limit the amount of data they need to dig up. That's one of the many legal side battles.

After that battle, the defending party will have to submit that data. The plaintiff will then often have a second legal battle complaining to the judge that the defendant didn't provide everything and so forth.

A common tactic is for a defendant to provide the plaintiff lawyers with a haystack. That is: overwhelm then with data whether it's related to the case or not.

The deliverables will usually include the actual source code, humongous amounts of emails etc.

It's very rare that plaintiff's lawyers will abuse this privilege and pass on this highly confidential information to the plaintiff, but it does happen: a very good recent example is one of that company that's ethics deprived to the core, Samsung, where it used discovery documents of an Apple-Nokia agreement during negotiations with Nokia. http://www.fosspatents.com/2013/10/sanctions-loom-large-samsung-execs-were.html?m=1
 
Assuming the details of the IP ever become relevant, the response to nVidia's accusations would likely start by trying to show how the ideas as described in the patents shouldn't have been patentable in the first place, then show examples of prior art that also fit nVidia's description in the patents, and, if necessary, finally show how the allegedly infringing graphics IP is substantially different from the description in nVidia's patents (the last one can be difficult due to the vagueness and intentional ambiguity of patent wording).

If it gets that far, the jury will have to hear far more about graphics architectural details than they might've ever cared to know.
 
Common sense would dictate that most patents are issued and approved because they are in some way, shape, or form an improvement over prior art. Anyway, when it comes to graphics IP and graphics R&D over the last 15 years, Samsung and Qualcomm are not considered to be innovators in the field, while NVIDIA is considered to be one of the most innovative companies in the field, with more fundamental graphics patents issued and approved than any other company in the world (and by a large margin too). So I seriously doubt that they would have any legs to stand on. Clearly NVIDIA chose these seven specific graphics patents for a reason, and I'm sure they feel extremely confident in being able to both defend their patents and in being able to prove that Samsung/Qualcomm GPU's infringe on their patents.
 
Common sense would dictate that most patents are issued and approved because they are in some way, shape, or form an improvement over prior art. Anyway, when it comes to graphics IP and graphics R&D over the last 15 years, Samsung and Qualcomm are not considered to be innovators in the field, while NVIDIA is considered to be one of the most innovative companies in the field, with more fundamental graphics patents issued and approved than any other company in the world (and by a large margin too). So I seriously doubt that they would have any legs to stand on. Clearly NVIDIA chose these seven specific graphics patents for a reason, and I'm sure they feel extremely confident in being able to both defend their patents and in being able to prove that Samsung/Qualcomm GPU's infringe on their patents.

Samsung's graphics credentials have zero to do with it, they aren't the GPU designers in any of the chips in question.

You might as well say that nvidia has very little cpu credentials, that doesn't have any bearing on the legality of the their CPU IP in tegra.
 
Samsung's graphics credentials have zero to do with it

In the context of this discussion, sure it does. In my opinion, it would be pretty illogical to suggest that a company like Samsung or Qualcomm--who are not considered to be innovators with respect to graphics IP over the last 15 years--would be magically able to invalidate seven very fundamental graphics patents from a graphics IP innovator such as NVIDIA.

If you read the actual complaint, you will see that NVIDIA is very specific about each individual line item within each patent that they believe is being infringed upon. They are in a much stronger position here than Samsung or Qualcomm because they have a variety of issued and approved graphics patents and they appear to have evidence that demonstrates infringement of these patents on a line by line basis.
 
Last edited by a moderator:
In the context of this discussion, sure it does. In my opinion, it would be pretty illogical to suggest that a company like Samsung or Qualcomm--who are not considered to be innovators with respect to graphics IP over the last 15 years--would be magically able to invalidate seven very fundamental graphics patents from a graphics IP innovator such as NVIDIA.

If you read the actual complaint, you will see that NVIDIA is very specific about each individual line item within each patent that they believe is being infringed upon. They are in a much stronger position here than Samsung or Qualcomm because they have a variety of issued and approved graphics patents and they appear to have evidence that demonstrates infringement of these patents on a line by line basis.

Considered by whom, exactly? Apart from you, that is.
 
Back
Top