NVIDIA Grace CPU Specs Remind Us Why Intel Never Shared x86 with the Green Team

[Count von Schwalbe]

For Arm, the profit margins simply aren't there unless you're putting your CPUs in everything, or building highly specialised CPUs for highly specialised niches. The latter is where NVIDIA has, rightly, focused their attention by augmenting their primary area of expertise (graphics) with CPUs that can help feed those graphics. They don't want to be a CPU company because CPUs are ancillary to their core focus.

Which once again brings us back to their attempted acquisition of Arm; I still struggle to see the reasoning behind it. The argument that it was to integrate NVIDIA graphics into Arm CPUs doesn't wash because NVIDIA's focus has always been high-performance high-power graphics, not low-end low-power ones as found in typical Arm applications, so they would essentially have to build an entirely new product. The thing is though, that doesn't require them to buy Arm; if NVIDIA already has a low-power GPU capable of competing with what's typically found in smartphones, there's absolutely nothing stopping them from just licensing or selling it as a standalone product.

The cynical take is that it's simply so NVIDIA could increase Arm licensing fees and reap the profits, but I really don't see that panning out well for them; it would almost certainly have pushed a lot of Arm licensees towards the royalty-free RISC-V, which makes it a self-defeating proposition.

And Intel is a CPU company. That did not stop them from trying to build a graphics card division. Twice.

 

[bug]

I now it's not their business model, but it's stupid IMO and more fool them.

Their profits and market cap tend to disagree with you

 

[Assimilator]

And Intel is a CPU company. That did not stop them from trying to build a graphics card division. Twice.

The fact that they've failed twice, going on three times, would suggest they should either not have wasted their time and money, or tried harder.

There is nothing wrong with trying to diversify and capture a slice of other markets, but you have to make sure you set yourself up to succeed. Intel has consistently failed to appreciate the magnitude of the task involved in building a competitive GPU division, consequently has consistently failed to do that setup, and thus has consistently failed to produce a competitive product.

 

[TechLurker]

However, you forgot the best part of it: Intel and its advanced foundry services, now available to third party customers might as well be fabbing both aforementioned companies' products themselves!

A Ryzen on Intel 20A? I want to see it, to be honest with you.

Probably unlikely unless TSMC fell to a sino invasion. That or they shift to Samsung, with GloFo for backup. If they did though, there'd be a lot of extra legalese to prevent Intel from stealing IP to funnel back into their own designs, which could potentially lock Intel into legal and production hell if their own chips happen to be remotely similar to some IP from AMD or NVIDIA.

 

[Dr. Dro]

Probably unlikely unless TSMC fell to a sino invasion. That or they shift to Samsung, with GloFo for backup. If they did though, there'd be a lot of extra legalese to prevent Intel from stealing IP to funnel back into their own designs, which could potentially lock Intel into legal and production hell if their own chips happen to be remotely similar to some IP from AMD or NVIDIA.

Gelsinger is not a fool, they would probably completely isolate Intel's foundry services division from their other design teams.