Intel and Nvidia team up on Intel x86 RTX SoCs for future gaming PCs

[Mr Majestyk]

Give it a decade, nVidia will own intel.

Highly doubtful any regulatory authority would ever allow that. Nvidia couldn't even buy ARM and with good reason.

That aside, this is huge news and should scare the crap out of AMD. Competition is great and I saw AMD slowly becoming like Intel of old, so hopefully this is a big wake up call in consumer area at least, but also they will now face some real competition in data centre too.

 

[human7]

This makes me a bit worried for Intel's Arc Graphics Cards. They're the only ones with a decent affordable card on the market right now, and (tinfoil hat time) I'd hate for this to be nvidia's attempt at putting those down before they start eating up market share.

-Nvidia doesn't give a flying fart about arc. It's so far off their radar they might as well not even know it exists.

At the Time of writing: Intel stock up 34%, Nvidia up 4%, AMD down 9%

Intel and Nvidia already dominate the GPU market, which could spell trouble for AMD's APU and Radeon products if Nvidia starts encroaching further on Intel. It might also lead to higher prices for consumers, especially if mid-range GPUs like the RTX 5060/70 end up integrated with Intel CPUs in the future.

It’s hard to imagine Nvidia letting Arc survive much longer as well.

Yeah, this is squarely aimed at the laptop market. If Nvidia is preemptively looking for a cushion in a post-AI bubble burst world, this looks like it could be a big part of it. Per The Register:
"In addition to a roughly 4 percent share of Chipzilla, Nvidia's partnership with Intel gives it tighter cooperation on co-designing hardware, and greater access to Intel's customer base.

"The return on that investment is going to be fantastic, both, of course, in our own business, but also in our equity share of Intel," Huang said, estimating the addressable market for datacenter CPUs at about $25 billion, and the laptop market at 150 million units a year.

The latter appears to be what Nvidia is really paying for. By integrating its GPUs directly into Intel's processors at the chiplet level, Nvidia gains access to a market that was previously inaccessible to it. Nvidia is certainly no stranger to the PC arena, but as Huang pointed out, its offerings have largely been constrained to high-end gaming systems.

"There's an entire segment of the market where the CPU and the GPU are integrated. It's integrated for form factor reasons, maybe it's for cost reasons, maybe it's for battery life reasons, all kinds of different reasons. And that segment has been largely unaddressed by Nvidia today," Huang said. "That segment of the market is really quite rich, and it's really quite large, and it's underserved today."

Nvidia has been moving in this direction for a little while now. Some versions of the tiny GB10 Superchip built in collaboration with MediaTek are also expected to see wider availability at some point. So, it's not just Intel CPUs getting the Nvidia graphics treatment.

But the deal could mean bad news for Intel's in-house graphics division. Intel has been trying to expand its influence over the PC graphics market with its Arc GPUs for several years now with limited success."

 

[Melkor Unlimited]

If Nvidia were to acquire Intel outright, Intel would lose its rights to use AMD’s x86-64 extension, and AMD would lose rights to Intel’s x86 base. That would effectively cripple both companies’ ability to produce modern CPUs.

How long takes patent to expire? 20 years?

AMD64 was released September 2003. So it is expired now.
As is SSE3 (2004)
There is comming expiration of some more extensions: SSSE3 (2006), SSE4.1 (2007), SSE4.2 (2008).

What counts and will be counted for some more years are: AVX, AES, FMA4 (2011), AVX2, FMA3 (2013), RdRand (2014), AVX-512 (2015) ... and newer extensions.

TL;DR - after terminating this cross-llincensing
- AMD could make x86-64 CPUs with SIMD up to SSE3 and some extensions which brought into pile
- Intel could make x86-64 CPUs with most of extensions except those made by AMD and not expired yet
It would hurt AMD more.

 

[SRB]

If Nvidia were to acquire Intel outright, Intel would lose its rights to use AMD’s x86-64 extension, and AMD would lose rights to Intel’s x86 base. That would effectively cripple both companies’ ability to produce modern CPUs.

- Intel owns the original x86 instruction set, while AMD owns the x86-64 extension (aka AMD64).
- They cross-license these technologies to each other, allowing both to produce modern x86-64 CPUs.
- The license includes a “poison pill” clause: if either company is acquired by a third party, the license automatically terminates.

Instead of a full acquisition (or why Nvidia will never buy Intel), Nvidia is currently pursuing deep collaboration with Intel, the $5 billion investment (is just part of that strategy) for joint development of custom x86 SoCs that integrate Nvidia GPU chiplets. This lets Nvidia tap into Intel’s x86 ecosystem without triggering the license termination and further push into gaining market share.

The licenses were not set up that way. The both have ongoing cross license that give the rights to both to use each others extensions, etc. The problem is that one side cannot sell their license without the other sides permission. If they do, the license agreement voids the ability of a new company to make x86 if sold without permission.

This was the poison pill section of the agreements that Intel put in because AMD was on the rocks when they started this, and Intel thought they would never have to worry about it.

 

[SRB]

NVLink could be beneficial for GPUs that have low VRAM (like 6GB or 8GB), because it has been proven that these low-VRAM cards benefit most from higher PCIe bandwidth, and NVLink offers higher bandwidth and lower latency than PCIe

I seriously doubt Nvida/Intel would spend the money to engineer NVlink into 8gb entry level cards. If that were the case, they could improve performance by using 256 bit busses and 16x pcie lanes. While high end AI cards are designed with NVlink for rack to rack and GPU to GPU communication with HBM, That set up would cost multiples more that just putting 16 or 24gb memory on in the first place

 

[SRB]

Yeah, this is squarely aimed at the laptop market. If Nvidia is preemptively looking for a cushion in a post-AI bubble burst world, this looks like it could be a big part of it. Per The Register:
"In addition to a roughly 4 percent share of Chipzilla, Nvidia's partnership with Intel gives it tighter cooperation on co-designing hardware, and greater access to Intel's customer base.

"The return on that investment is going to be fantastic, both, of course, in our own business, but also in our equity share of Intel," Huang said, estimating the addressable market for datacenter CPUs at about $25 billion, and the laptop market at 150 million units a year.

The latter appears to be what Nvidia is really paying for. By integrating its GPUs directly into Intel's processors at the chiplet level, Nvidia gains access to a market that was previously inaccessible to it. Nvidia is certainly no stranger to the PC arena, but as Huang pointed out, its offerings have largely been constrained to high-end gaming systems.

"There's an entire segment of the market where the CPU and the GPU are integrated. It's integrated for form factor reasons, maybe it's for cost reasons, maybe it's for battery life reasons, all kinds of different reasons. And that segment has been largely unaddressed by Nvidia today," Huang said. "That segment of the market is really quite rich, and it's really quite large, and it's underserved today."

Nvidia has been moving in this direction for a little while now. Some versions of the tiny GB10 Superchip built in collaboration with MediaTek are also expected to see wider availability at some point. So, it's not just Intel CPUs getting the Nvidia graphics treatment.

But the deal could mean bad news for Intel's in-house graphics division. Intel has been trying to expand its influence over the PC graphics market with its Arc GPUs for several years now with limited success."

There are a few practical problems with this. First, Nvidia's ARM laptop chip is way late to market, reportedly because of NVlink issues. Second, NVlink is used for GPU to GPU and rack to rack communications. It may not scale down for chiplet communication very well. Intel still sells the lion's share of laptops WITHOUT discrete graphics. If anything, chiplet wise, the only thing they'd compete with is the Ryzen AI Max chips, which probably have great margins. Price wise, they're super expensive and not a huge market share.

While this could be a turning point for PC's and especially laptops, I'm just not seeing the Intel/Nvidia superchip selling in big enough numbers to be a sea change.

 

[airbornex]

The US government, SoftBank, and now Nvidia all owning chunks of Intel feels like the tech version of a fantasy football league. Wonder who gets bragging rights if these chips actually take off.

 

[Watzupken]

Give it a decade, nVidia will own intel.

Having burned their fingers trying to acquire ARM, it seems like they won’t be trying to acquire. And I feel Nvidia don’t have to either because this partnership is mostly beneficial for Nvidia, less for Intel. Sure Intel chips may sell well, but how much margin are they making given that Nvidia is the dominant partner here? And it also makes Intel dependent on Nvidia, not the other way round.

 

[Vulcanproject]

I think you misunderstand the console market. While consoles may someday go chiplet, so far the have been monolithic designs, same with the handheld market. While NV link may be fast, I seriously doubt it is power efficient. Consoles play in a unique space just like the handhelds do. For that matter, Nvidia hasn't had a lot ot luck getting the bugs worked out with their ARM pc chip with NVlink.

You overstate the issues for a home console application. If they can be used in notebooks they can be comfortably viable for a games console. A single custom design with fixed hardware in the tens of millions of units is the perfect application. More so than variable notebook configurations that prevented Kaby Lake-G succeeding.

Whether or not they want to offer some sort of alternative to AMD console silicon is the real question. There was plenty of talk that Sony shopped around at Intel for PS5 but saw there was no viable alternative at the time. Going forward there is another door now open, albeit slightly.

 

[human7]

There are a few practical problems with this. First, Nvidia's ARM laptop chip is way late to market, reportedly because of NVlink issues. Second, NVlink is used for GPU to GPU and rack to rack communications. It may not scale down for chiplet communication very well. Intel still sells the lion's share of laptops WITHOUT discrete graphics. If anything, chiplet wise, the only thing they'd compete with is the Ryzen AI Max chips, which probably have great margins. Price wise, they're super expensive and not a huge market share.

While this could be a turning point for PC's and especially laptops, I'm just not seeing the Intel/Nvidia superchip selling in big enough numbers to be a sea change.

If they can get Microsoft to approve an Nvidia iGPU for their AI PC labeling, they might be able to corner the business market. Conditioning on the killer app in that category, of course. Big if.

 

[pokan]

They missed the perfect opportunity to name them RTx86 SOC's.

 

[Puiu]

Until I see proper products from this, I'll stll consider the future of Intel to be bleak. They can get the best GPUs from Nvidia to put inside as chiplets, if the CPUs are far behind AMD... it will only see very limited adoption by the market.

 

[gamerk2]

If Nvidia were to acquire Intel outright, Intel would lose its rights to use AMD’s x86-64 extension, and AMD would lose rights to Intel’s x86 base. That would effectively cripple both companies’ ability to produce modern CPUs.

- Intel owns the original x86 instruction set, while AMD owns the x86-64 extension (aka AMD64).
- They cross-license these technologies to each other, allowing both to produce modern x86-64 CPUs.
- The license includes a “poison pill” clause: if either company is acquired by a third party, the license automatically terminates.

Instead of a full acquisition (or why Nvidia will never buy Intel), Nvidia is currently pursuing deep collaboration with Intel, the $5 billion investment (is just part of that strategy) for joint development of custom x86 SoCs that integrate Nvidia GPU chiplets. This lets Nvidia tap into Intel’s x86 ecosystem without triggering the license termination and further push into gaining market share.

Tbf, there's nothing that would preclude NVIDIA from purchasing Intel (voiding the agreement) then signing a new one with AMD after the fact.

 

[gamerk2]

Until I see proper products from this, I'll stll consider the future of Intel to be bleak. They can get the best GPUs from Nvidia to put inside as chiplets, if the CPUs are far behind AMD... it will only see very limited adoption by the market.

Intel made a massive mistake skipping over EUV @10nm, in order the reduce costs in the short term. They're still playing catchup, and since they've lost share since Ryzen they don't really have the time or funds to really fix their problems.

 

[Puiu]

Tbf, there's nothing that would preclude NVIDIA from purchasing Intel (voiding the agreement) then signing a new one with AMD after the fact.

I think anti-competitive laws should prevent Nvidia from merging with Intel (like with the Nvidia and ARM deal that was prevented by the EU). Unless Trump says "it's ok" and he orders the supreme court to vote in favor since these are US companies.

 

[gamerk2]

I think anti-competitive laws should prevent Nvidia from merging with Intel (like with the Nvidia and ARM deal that was prevented by the EU). Unless Trump says "it's ok" and he orders the supreme court to vote in favor since these are US companies.

If I were in charge, I'd make it so companies can be broken up for no other reason then being "too large". To me, none of these companies should have been allowed to exist to start with.

 

[Loadedaxe]

And as of now, AMD stock is down 1.9%, Intel is up 22% and Nvidia is up 3.6%.

People read through this announcement and have seen that Nvidia is finally acknowledging they are worried about AMD, and perceived the company as a threat.

This is an admission that Nvidia ARM venture's was a bust.
This is also an admission that Intel GPU venture's was a bust.

Now I wonder if Grace, Nvidia ARM for desktop, ARC and Gaudi are going to be scrapped.

Boy, the conspiracy theories ...

That’s not what’s happening at all. Nvidia picking up Intel shares doesn’t suddenly mean “panic over AMD” or that their ARM venture or GPU strategy is a bust. Nvidia has more cash on hand than they know what to do with, and buying into Intel is just a strategic investment in a company they already rely on for fabs, packaging, and server synergy.

If Nvidia was worried about AMD, they wouldn’t be growing revenue at a rate that dwarfs both AMD and Intel combined. Grace is ramping, Hopper is selling as fast as they can make it, and Blackwell is already locking up hyperscale contracts. Intel’s ARC isn’t scrapped, Gaudi is gaining traction in AI, and ARM isn’t going anywhere...especially since Nvidia is using it successfully in Grace/Hopper superchips.

This isn’t an admission of failure, it’s just smart business. Nvidia benefits if Intel stabilizes because it ensures more capacity outside of TSMC and strengthens the ecosystem Nvidia already dominates.

Contrary to what you all believe, Intel's fabs are a necessity for competition to keep TSMC in check.

 

[Loadedaxe]

If I were in charge, I'd make it so companies can be broken up for no other reason then being "too large". To me, none of these companies should have been allowed to exist to start with.

That sounds nice in theory, but too large isn’t a legal or economic standard, it’s just subjective.

Breaking up companies purely because they’re big ignores why they got big in the first place. Innovation, efficiency, and market demand. If you punish success by dismantling it without evidence of monopoly abuse, you destroy the incentive for companies to grow, compete, and invest.

Antitrust law is supposed to protect consumers, not handicap companies simply for being successful. If a company dominates because it’s delivering the best products and customers choose them freely, that’s not a problem....it’s literally how markets are supposed to work.

 

[Loadedaxe]

The licenses were not set up that way. The both have ongoing cross license that give the rights to both to use each others extensions, etc. The problem is that one side cannot sell their license without the other sides permission. If they do, the license agreement voids the ability of a new company to make x86 if sold without permission.

This was the poison pill section of the agreements that Intel put in because AMD was on the rocks when they started this, and Intel thought they would never have to worry about it.

That’s a common misconception. The cross licensing agreement between Intel and AMD isn’t some trap that prevents AMD from surviving or selling parts of its business, it’s a standard safeguard that keeps the x86 IP from being freely transferred to third parties. Both sides agreed to it, and it’s been renewed multiple times without issue because both benefit.

If AMD were ever sold, it doesn’t automatically mean no more x86. It would just require renegotiation with Intel...something that would absolutely happen, because Intel gains nothing by nuking its only real x86 competitor especially with regulators watching.

The poison pill narrative makes it sound like Intel designed some secret kill switch, when in reality it’s just basic IP protection and contract structure. AMD isn’t trapped, it’s still here decades later, competing stronger than ever.

 

[Puiu]

If I were in charge, I'd make it so companies can be broken up for no other reason then being "too large". To me, none of these companies should have been allowed to exist to start with.

Such laws already exist, but when you have legal "bribes" large corporations think they can do whatever they want without any consequences (and history agrees with this statement).

 

[Hecate91]

Nvidia buying Intel shares isn't about Nvidia bring scared of AMD, it's about the leather jacket man wanting to corner the market on laptops and APU's which AMD has been quite successful.
And Nvidia buying shares is at an easy time for them when antitrust regulators will look the other way while two monopolistic companies are allowed to partner up. There is a difference in a company being successful and the consumer losing out because a company has nearly majority control over the market. In the long term I would expect Nvidia to buy Intel's fabs unless tons of cash flow from from the leather jacket man and the US govt can get Intel to improve yields.

 

[Loadedaxe]

Nvidia buying Intel shares isn't about Nvidia bring scared of AMD, it's about the leather jacket man wanting to corner the market on laptops and APU's which AMD has been quite successful.
And Nvidia buying shares is at an easy time for them when antitrust regulators will look the other way while two monopolistic companies are allowed to partner up. There is a difference in a company being successful and the consumer losing out because a company has nearly majority control over the market. In the long term I would expect Nvidia to buy Intel's fabs unless tons of cash flow from from the leather jacket man and the US govt can get Intel to improve yields.

I don’t think it’s as simple as cornering the laptop/APU market. Nvidia doesn’t even have x86 rights, and Intel’s cross license agreements with AMD are structured so neither can just sell off that license to a third party without it being voided. That was Intel’s idea to prevent exactly this scenario.

Buying a minority stake doesn’t give Nvidia control of Intel, its fabs, or its x86 IP. What it does give them is influence, closer cooperation on foundry capacity, and a hedge in case their reliance on TSMC becomes a liability.

Regulators would step in long before Nvidia could actually swallow Intel whole, because letting one company hold GPUs, AI accelerators, fabs, and x86 CPUs under the same roof would be the textbook definition of antitrust violation.

And honestly, this isn’t even about gaming. Can gaming benefit? Sure. But Nvidia’s primary focus now is AI, datacenters, and securing reliable manufacturing pipelines. Gaming is just the side benefit, not the driver of these moves.

TechSpot should honestly be ashamed for even framing it the way they did...it’s clickbait speculation at best.

 

[Loadedaxe]

Such laws already exist, but when you have legal "bribes" large corporations think they can do whatever they want without any consequences (and history agrees with this statement).

Laws do exist, yes, but the key difference is enforcement. Corporations always test boundaries...it’s what they do, but that doesn’t mean they’re free to do whatever they want.

Antitrust regulators, especially in the U.S. and EU, have been far more aggressive recently with blocking mergers, imposing fines, and dragging companies through years of litigation.

Nvidia trying to outright buy Intel, or even just its fabs, wouldn’t skate by under the radar no matter how much legal lobbying money is spent. History also shows plenty of examples where regulators stepped in and stopped things dead in their tracks.

 

[gamerk2]

Breaking up companies purely because they’re big ignores why they got big in the first place. Innovation, efficiency, and market demand. If you punish success by dismantling it without evidence of monopoly abuse, you destroy the incentive for companies to grow, compete, and invest.

Except we see the opposite: Big companies that dominate their market are inefficient and lead to stagnation.

Such laws already exist, but when you have legal "bribes" large corporations think they can do whatever they want without any consequences (and history agrees with this statement).

The only anti-trust laws that exist are for anti-competitive practices; it's impossible to break up a company that's already formed unless you can *prove* they are engaging in anti-competitive conduct. That's why you see ever major US industry dominated by just two or three large players, with everyone else squeezed out.

 

[Loadedaxe]

Except we see the opposite: Big companies that dominate their market are inefficient and lead to stagnation.

That’s only half of the picture. Big companies can get bloated and complacent, sure...but they also often drive the kind of massive R&D investment smaller players simply can’t afford.

Intel, Nvidia, AMD, Apple, none of them are “stagnant,” they’re all throwing billions at AI, chip design, and advanced nodes because competition requires them to. What actually leads to stagnation isn’t just size, it’s lack of competition or lack of pressure.

As long as rivals exist and governments enforce antitrust when needed, even the giants have to keep moving forward.