Report: Intel on verge of settlement with FTC and nVidia

[Creig]

Why do you bother arguing with someone who can't even spell "Jedi" correctly?

Maybe he was referring to a Jedi Yeti?

 

[Compddd]

I don't see how this benefits Nvidia at all

 

[Genx87]

NV is only just doing well enough to keep their GPU R&D pipeline resourced enough to stay competitive...where would they even get the resources necessary to develop a competitive product to AMD's and Intel's offerings in the six years time that will pass before NV would even bring a product to the market?

Look at how much money and time Intel threw at Larrabee. Look how long Via has been developing their x86 processors and we see just how competitive their products are with AMD and Intel.

NV jumping into x86 would simply be a fool's errand for them. It would be perfect for AMD and Intel if NV starved their GPU development teams while attempting to fund an x86 development team.

That is how I feel about it as well. x86 for Nvidia's business plan really doesnt seem to fit. Mobile devices wont be using x86. Mobile devices are going to move a hell of a more product in the future than x86 devices.

 

[HurleyBird]

Now Nvidia can buy Via.

 

[Wreckage]

Now Nvidia can buy Via.

That would be huge!

 

[HurleyBird]

That would be huge!

Yes, it would. I've been thinking for awhile that it's only a matter of time before Intel and AMD squeeze Nvidia out of the market. If Nvidia bought VIA and obtained x86 rights, then we could see some very interesting things.

 

[WelshBloke]

Now Nvidia can buy Via.

Isnt Via part of some massive Taiwanese conglomerate?

 

[PandaBear]

The problem with x86 is not really licensing, but manufacturing process. You cannot, without a FAB, design an x86 that build and runs well to compete with Intel or AMD. Having an ATOM like platform will not make a lot of money and to be honest you need to sell in high, mid, and low end to make money selling CPU.

Then a few years later Intel will put the GPU on die, and render the entire market of chipset gone. AMD has ATI so they are not worrying too much, Intel may buy nvidia if they are not banned to do so. In the mean time, how much longer can nvidia sustain before Intel offer to buy it?

The problem with mobile market is that nvidia is not going to be the only one that can do it, and since everyone can put an ARM core on it, it'll be free for all and not a huge money maker like the current chipset market.

 

[Wreckage]

The problem with x86 is not really licensing, but manufacturing process. You cannot, without a FAB, design an x86 that build and runs well to compete with Intel or AMD.

Technically AMD no longer has a fab.

NVIDIA has been dabbling in x86 for awhile now even hiring x86 engineers. I don't know if they will buy VIA. I don't even think they need x86 all that much.

However if they can get VIA at a good price, they can attack Intel/AMD on the low end immediately. A few years down the road they could could even compete top to bottom.

This ruling certainly opens the door for such a scenario.

 

[HurleyBird]

The problem with x86 is not really licensing, but manufacturing process. You cannot, without a FAB, design an x86 that build and runs well to compete with Intel or AMD. Having an ATOM like platform will not make a lot of money and to be honest you need to sell in high, mid, and low end to make money selling CPU.

And without a licence you can't build any at all... so yeah, the licence is just a little bit more important.

 

[jvroig]

However if they can get VIA at a good price, they can attack Intel/AMD on the low end immediately. A few years down the road they could could even compete top to bottom.

Hehe, yeah, after all, designing and making x86 CPUs is not that hard and not at all different from designing and making GPUs.

I take it you weren't able to read Idontcare's post a few posts just above yours? Even if you weren't, it should be obvious that R&D and manufacturing costs for CPUs will get them nowhere. They'd need intellectual capital (read: the right people), they'd need expertise (which comes with experience, and that comes at the cost of years), and they'd need to at least double their R&D budget to make way for any x86 efforts. Attacking Intel/AMD in the low end "immediately" is so far out of the realm of possibility right now. If barriers to entry were that low, there would have been more than just AMD and Intel on the playing field. The fact that the rest of the CPU players don't bother competing in x86 shows that it just isn't a market that's easy to penetrate at all.

 

[Idontcare]

Isnt Via part of some massive Taiwanese conglomerate?

Formosa Plastics Group

 

[BenSkywalker]

What I think could be interesting on a potential level is nV building a VIA/nV SoC to compete against the likes of Atom. Previously such a setup would have been impossible due to legal issues, but it sounds like such a setup would be covered under 'joint ventures'.

 

[HurleyBird]

Attacking Intel/AMD in the low end "immediately" is so far out of the realm of possibility right now. If barriers to entry were that low, there would have been more than just AMD and Intel on the playing field. The fact that the rest of the CPU players don't bother competing in x86 shows that it just isn't a market that's easy to penetrate at all.

Ugh. If Nvidia bought VIA they *would* be able to immediately compete at the low end since VIA already has low end x86 CPUs.

x86 isn't hard per-say. Not even Intel or AMD use x86 anymore aside from a translation layer. Other CPU vendors don't compete because they are artificially locked out due to not having a license. Another barrier would be for specific architectures like VLIW that are by their definition are very hard to translate any foreign code to since they need the help of the compiler to extract performance.

 

[Wreckage]

Hehe, yeah, after all, designing and making x86 CPUs is not that hard and not at all different from designing and making GPUs.

I take it you weren't able to read Idontcare's post a few posts just above yours? Even if you weren't, it should be obvious that R&D and manufacturing costs for CPUs will get them nowhere. They'd need intellectual capital (read: the right people), they'd need expertise (which comes with experience, and that comes at the cost of years), and they'd need to at least double their R&D budget to make way for any x86 efforts. Attacking Intel/AMD in the low end "immediately" is so far out of the realm of possibility right now. If barriers to entry were that low, there would have been more than just AMD and Intel on the playing field. The fact that the rest of the CPU players don't bother competing in x86 shows that it just isn't a market that's easy to penetrate at all.

Look up NVIDIA and Transmeta and then get back to me. ^_^

 

[jvroig]

Look up NVIDIA and Transmeta and then get back to me.

Transmeta never made a dent in the x86 market. I don't see why you think nVidia + Transmeta would do any better.

Don't get me wrong. I don't hate Transmeta. I actually like them even if they weren't successful in x86, and this positive impression I have of them is simply because they hired Linus Torvalds way back when, and I'm a Linux guy, so in my book that automatically made Transmeta one of the good guys.

And I don't hate having another player to compete with AMD and Intel in the x86 space. In fact, I'd love to have another real competitor in there. But given barriers to entry, it just isn't feasible. Intel and AMD's tech advantage there is too much.

But after this point, I suppose we'd just be trading "what ifs" and opinions, so I guess this won't go anywhere. That's what forums are for, I suppose

 

[Wreckage]

Transmeta never made a dent in the x86 market. I don't see why you think nVidia + Transmeta would do any better.

Don't get me wrong. I don't hate Transmeta. I actually like them even if they weren't successful in x86, and this positive impression I have of them is simply because they hired Linus Torvalds way back when, and I'm a Linux guy, so in my book that automatically made Transmeta one of the good guys.

And I don't hate having another player to compete with AMD and Intel in the x86 space. In fact, I'd love to have another real competitor in there. But given barriers to entry, it just isn't feasible. Intel and AMD's tech advantage there is too much.

But after this point, I suppose we'd just be trading "what ifs" and opinions, so I guess this won't go anywhere. That's what forums are for, I suppose

My point is a lot of what was Transmeta went over to NVIDIA. So they have the expertise/right people. They also have a billion or 2 in the bank so they have plenty of R&D money. If they picked up VIA they would also have a chip to build off of.

Besides how much market share would NVIDIA need to be successful?

AMD has very little, do you consider them a failure in the x86 market?

http://www.xbitlabs.com/news/cpu/di...Share_Drops_Intel_s_Rises_in_Q1_2010_IDC.html

 

[tommo123]

question, can nvidia even do anything useful with an x86 licence?

i mean, amd64 is AMDs own right? they share that with intel in exchange for various SSE x access?

wouldn't an x86 licence mean nvidia can do something pre mmx? a chip thta wouldn't run anything really these days?

 

[jvroig]

AMD has very little, do you consider them a failure in the x86 market?

Nope, because I can go to most brick and mortar stores and buy an AMD CPU. I can also go online and order AMD CPUs. I can easily find motherboards that accept AMD CPUs, because they are supported by all the big players - ASUS, Gigabyte, MSI, etc. It is also quite hassle-free to find laptops, desktops, and servers that use AMD CPUs - ASUS, MSI, Acer, Dell, HP, Provantage, Supermicro... they'd all gladly give you all the AMD-powered devices you can order.

Naturally, the same goes for Intel.

But the same cannot be said for Transmeta.

 

[bryanW1995]

NV is only just doing well enough to keep their GPU R&D pipeline resourced enough to stay competitive...where would they even get the resources necessary to develop a competitive product to AMD's and Intel's offerings in the six years time that will pass before NV would even bring a product to the market?

Look at how much money and time Intel threw at Larrabee. Look how long Via has been developing their x86 processors and we see just how competitive their products are with AMD and Intel.

NV jumping into x86 would simply be a fool's errand for them. It would be perfect for AMD and Intel if NV starved their GPU development teams while attempting to fund an x86 development team.

well, as we have seen it doesn't take the fastest processor, or even a very fast processor at all, to max out most video games these days. Nvidia needs a ready-made mareket for their gpus, and here are their options:

1. amd - has their own gpu division
2. intel - hates them with a passion, cuts them out at every turn.

not exactly good options right now, are they? maybe they could enter into an agreement with via instead of an outright purchase. or, if nothing else, they at least have more options today than they did before.


A point that many of us are missing is that via could be worth a lot more today than it was yesterday. nvidia is certainly not the only, or even best, potential suitor for via. it is in nvidia's interest that SOMEBODY develop a cpu as a 3rd option to intel and amd, however, and this ftc settlement makes it more likely now that this can happen.

 

[Fox5]

Isnt Via part of some massive Taiwanese conglomerate?

They could buy Nvidia and then have the world's 3rd best x86 cpu company, 3rd best cpu manufacturing, and....eventually degrade nvidia into the world's 3rd best gpu company? (behind amd and powervr I'd imagine, but intel IGPs COULD get better)

 

[Hard Ball]

After reading through the settlement and trying to parse the legalese, I must say that it does not stray too far from what I expected. Intel becomes much more restricted in terms of marketing, licensing, design and manufacturing in terms of the possible directions that they can possibly pursue in the future. But the requirements on changing current practice is minimal, mainly because Intel has already back off many of the unsavory conduct from earlier this decade.

Couple of minor surprises though:

• Via got off much better than I expected. While they were always an essential component of FTC's complaint, I didn't expect Via getting several provisions specifically targeting them, including the requirement to renew their x86 license, and specifically to put the legitimacy of their x86 business beyond the shadow of doubt.
• The specificity of the language addressing possible M&A activities of the current x86 designer/manufacturers. I don't think this provision would be in there, unless the FTC has some inkling of what might be ahead for some transaction of the sort involving AMD and/or Via in the next few years. This provides legitimacy for the x86 business of any currently non-x86 producing entity, should they enter into such a deal with AMD or Via; this really got me wondering.
• nVidia didn't get much in terms of the potential harm that intel has caused them in terms of the chipset license lockout. The question here is whether they would continue to pursue some legal avenue in this area. Perhaps someone with some legal expertise in this area would be able to answer better.

 

[lifeblood]

I'm happy with the setlement as it provides more protections, especially for AMD. I am disapointed with the setlement as nothing changes (that I can see) as far as chipsets. I was really hoping that would be clarified so nVidia could start making chipsets again. Competition is good.

 

[Wreckage]

I'm happy with the setlement as it provides more protections, especially for AMD. I am disapointed with the setlement as nothing changes (that I can see) as far as chipsets. I was really hoping that would be clarified so nVidia could start making chipsets again. Competition is good.

I'm pretty sure NVIDIA has their own case in court against Intel, that is separate from this FTC agreement.

 

[Hard Ball]

I'm pretty sure NVIDIA has their own case in court against Intel, that is separate from this FTC agreement.

Well, we will see about that in the near future.

I think if I read the document right, there is a short period in which Intel would be required to make a good faith effort in negotiating and renewing licenses that would mitigate their past wrong doings. We will see if they extend something significant to NV (Via almost certainly, since that's explicitly spelled out); if the answer is negative, see if NV would again go the legal route.