As inconceivable as it may sound, your next Nvidia GPU could be manufactured by Intel. Nvidia CEO Jensen Huang held a question and answer session with the press today, and the topic quickly turned to Intel's Foundry Services (IFS) initiative that will see Intel making chips for other companies as part of its IDM 2.0 initiative. Surprisingly, Huang confirmed that his company is considering using Intel's foundry to possibly make some of its chips. Intel is now a direct competitor with Nvidia on both the CPU and GPU fronts, but Huang also explained that Intel and AMD have known Nvidia's secret roadmaps for years, so he isn't paranoid about sharing more information.
"Our strategy is to expand our supply base with diversity and redundancy at every single layer. At the chip layer, at the substrate layer, the system layer, at every single layer. We've diversified the number of nodes, we've diversified the number of foundries, and Intel is an excellent partner of ours[…]. They're interested in us using their foundries, and we're very interested in exploring it," said Huang.
Huang did hedge somewhat, explaining that operating as a foundry is very different from operating as a standard product-oriented company like Intel. "Being a foundry at the caliber of TSMC is not for the faint of heart; this is a change not just in process technology and investment of capital, but it is a change in culture, from a product-oriented company to a product, technology, and service-oriented company," Huang explained.
"A service-oriented company that dances with your operations. TSMC is a services company that dances with, what, 300 companies worldwide? Our own operations are quite an orchestra, and yet they dance with us. And then there's another orchestra they dance with. So the ability to dance with all of these operations teams and supply chains teams is not for the faint of heart, and TSMC does it just beautifully. Its management, its culture, and its core values, and you put that on top of technology and products." Huang said.
"I am encouraged by the work that is done at Intel, I think this is a direction they have to go, and we're interested in looking at their process technology. Our relationship with Intel is quite long; we work with them across a whole lot of different areas, every single PC, every single laptop, every single PC, supercomputer, we collaborate." Huang said.
Huang explained that using Intel as a foundry services partner would take an extended period of time. "Foundry discussions take a long time, and it's not just about desire. We have to align technology, the business models have to be aligned, the capacity has to be aligned, the operations process and the nature of the two companies have to be aligned. It takes a fair amount of time and a lot of deep, deep discussion – we're not buying milk here. This is really about the integration of the supply chains. Our partnerships with TSMC and Samsung in the last several years are something that took years to cultivate. So we are very open-minded to considering Intel, and I'm delighted by the efforts that they're making."
Intel now makes GPUs, and Nvidia now makes GPUs, meaning the two companies will now compete directly in several market segments. Huang also commented about potential concerns with sharing Nvidia's secrets with Intel, and his response was quite revealing:
"We have been working closely with Intel, sharing with them our roadmap long before we share it with the public, for years. Intel has known our secrets for years. AMD has known our secrets for years. We are sophisticated and mature enough to realize that we have to collaborate.[...] We share roadmaps, of course, under confidentiality and a very selective channel of communications. The industry has just learned how to work in that way."
"On one hand, we compete with many companies, but we also partner deeply with them and rely on them. As I mentioned, if not for AMD CPUs in DGX, we couldn't ship DGX. If not for Intel CPUs and all of the hyperscalers connected to our HGX, we wouldn't be able to ship HGX. If not for Intel's CPUs in our Omniverse computers that are coming up, we wouldn’t be able to do the digital twin simulations that rely so deeply on the single-threaded performance that they're really good at."
Huang's governing philosophy for how Nvidia deals with its partners is far from the "Only the paranoid survive" mantra that was popularized by Intel legend Andy Grove:
"It turns out that paranoia is just paranoia. There is nothing to be paranoid about. It turns out that people want to win, but nobody's trying to 'get ya.' So we try to take the not-paranoid approach when we work with partners. We try to rely on them, let them know we are relying on them, trust them, and let them know we trust them. So far it is working pretty well," Huang said.
Stay on the Cutting Edge
Join the experts who read Tom's Hardware for the inside track on enthusiast PC tech news — and have for over 25 years. We'll send breaking news and in-depth reviews of CPUs, GPUs, AI, maker hardware and more straight to your inbox.
Paul Alcorn is the Deputy Managing Editor for Tom's Hardware US. He writes news and reviews on CPUs, storage and enterprise hardware.
I remember long ago when I was a new engineer. Nvidia was looking for someone to make their new Geforce GPU for them. They were practically begging Intel to fab their chips for them. Intel told them to go pound sand, we're not anyone's foundry.Reply
NVIDIA see's the coming crisis in Taiwan. They are hedging bets with Samsun and IntelReply
Is this due to them annoying both TSMC and now Samsung?Reply
I feel it is all cost driven. Nvidia said that they moved to Samsung because they wanted to diversify the production of their GPUs. But if you look at the choice of node selected, Nvidia went for Samsung’s 8nm (which is a 10nm) instead of the superior 7nm. So in my opinion, the action is more to reduce cost, and diversification is just an after thought or secondary reason. In other words, Nvidia not sticking with TSMC and exploring other foundries is also going to be largely cost driven. Having said that, they will need to diversify because with what we see with Ukraine, there is always a risk of Taiwan getting invaded. So if that is to happen, then they will lose all chip making capability. I am sure the other big chip makers will also have this in mind, thus, we see them using Samsung, and for some, hear rumours of them checking out Samsung foundry, i.e. AMD.thisisaname said:Is this due to them annoying both TSMC and now Samsung?
TSMC is the definition of Geopolitical risk and for some reason that is on everyone's mind.Reply
Also pricing in a market with three players is greatly improved over a market with one.
"Get three bids" is basic advice for any purchase process.
jkflipflop98 said:I remember long ago when I was a new engineer. Nvidia was looking for someone to make their new Geforce GPU for them. They were practically begging Intel to fab their chips for them. Intel told them to go pound sand, we're not anyone's foundry.
and that's why TSMC finally able to surpass intel when it comes to fab advancement.
maybe that is one of the reason but nvidia has been looking for alternative other than TSMC for a very long time now. the talk to use samsung starts in 2012 when in the same time nvidia also openly talk about their dissatisfaction on the situation with TSMC 28nm. i think it is also around the same time when nvidia publically said that intel should open their fab for others to use.digitalgriffin said:NVIDIA see's the coming crisis in Taiwan. They are hedging bets with Samsun and Intel
thisisaname said:Is this due to them annoying both TSMC and now Samsung?
nvidia might annoy TSMC but ultimately TSMC have said that they want nvidia to fab with them more. this is after nvidia end up choosing samsung as their main partner to produce their 30 series. plus for 2021 7nm wafer allocation nvidia actually got a lot of space from TSMC. AMD being the biggest TSMC 7nm customer end up getting 27% of TSMC 7nm capacity and they make CPU, GPU and console on that node. Nvidia is TSMC second largest 7nm customer with 21% of 7nm allocation. and nvidia mostly only make A100 on TSMC 7nm. TSMC giving nvidia quite a lot of capacity just for that compared to AMD that pretty much make everything on than node.
Samsung? they still need a lot of proving to be done before they can get annoyed with nvidia. nvidia is very well known to have crazy demand. TSMC has been dealing with that side of nvidia since the 90s. right now Qualcomm are quite dissatisfied with the yield on Samsung 5nm process. Qualcomm take both samsung and TSMC 5nm and the yield on samsung 5nm is 35% while on TSMC the yield are over 70%. there are talks about how samsung are lying on yield to convince some chip maker to use their process. i heard samsung intend someone inside the company to take responsibility on this issue. if anything samsung will want nvidia to stay on their foundry and help them invest on the fab advancement the same way nvidia did with TSMC years ago.
If Intel itself outsource their GPU fab to TSMC, why would NVidia would use Intel Fab to produce their GPU ? I suspect that money the most likely answer ... :sneaky:Reply
samopa said:If Intel itself outsource their GPU fab to TSMC, why would NVidia would use Intel Fab to produce their GPU ? I suspect that money the most likely answer ... :sneaky:
intel is increasing their fab capacity going forward. they scoff at nvidia idea about opening their fab to others before forever thinking that no other fab can ever surpass them but 14nm happen and TSMC finally showed them that working with hundreds of different client with different needs will give you an experience that you will never had if you only make your own product.