Speaking at a press conference earlier this evening, Nvidia CEO, Jensen Huang, and Intel chief Lip-Bu Tan took questions from the media about the newly-announced collaboration between the two.
While both CEOs refused to be drawn too far in on questions regarding which process the chips created would be built on (or whether Nvidia would commit to making use of Intel's foundry in the near future), the two did have a lot to say about the project, and the scope of the market it's attempting to capture.
"With respect to the size of the market," said Huang, "the data center market and the PC market are both large, and we're going to build revolutionary products, first of its kind products. Nothing of its kind has ever been built before for the x86 market."
The products in question are said to be both Nvidia-custom x86 CPUs integrated into AI infrastructure platforms for data center usage, and system-on-chips (SOCs) with integrated Nvidia RTX GPU chiplets for the consumer market.
"I think these are going to be revolutionary products," Huang continued. "I know that all of us working on it are super excited about it, the architects working on it are super excited about it. And so we're looking forward to telling you more about it over time."
Despite the reluctance to give exact details of the project, Huang did give some hints as to how the two architectures might work together for the SOC-based consumer chips:
"Intel has the Foveros multi technology packaging capability, and it's really enabling here. And the reason for that is because, as we all know, Nvidia's GPU technology is based on TSMC foundry. And this is one of the extraordinary things that you can do, connecting Nvidia's chiplets with Intel CPUs in a multi-technology packaging capability and multi-process packaging technology."
"And so it's really a fabulous way of mixing and matching technology, and that's one of the reasons why we're going to be able to innovate so quickly and build these incredibly complex systems and deliver it as a multi-chiplet systems package."
The new chips will likely be a straight-up fusion between Nvidia's TSMC-based GPU tech and Intel's mostly-TSMC-based tech for now.
So, it seems as though the new chips will likely be a straight-up fusion between Nvidia's TSMC-based GPU tech and Intel's mostly-TSMC-based tech for now. When asked for a second time about the prospect of Nvidia someday using Intel as a foundry partner for its advanced AI chips, Intel CEO Lip-Bu Tan said:
"As we mentioned earlier, this is more the product collaboration announcement. We both [have] a lot of respect for TSMC, C.C Wei, Morris Chang, and we continue to work with them.
"In terms of the Intel foundry, we continue to make progress. And then in terms of the yield performance, 18A, 14A, clearly we want to qualify. Jensen and I will review that, but overall we're going to continue driving our success on the process side, and then win the customer's confidence and trust, and then one step at a time."
Exciting stuff nonetheless, isn't it? I don't think any of us would have had an Nvidia/Intel collab on our books this morning, but it seems the two have been working together for longer than anyone expected.
"The two technology teams have been discussing and architecting solutions now for probably coming up to a year," said Jensen. "This partnership, we kept it really quiet."
No kidding!
1. Best CPU: AMD Ryzen 7 9800X3D
2. Best motherboard: MSI MAG X870 Tomahawk WiFi
3. Best RAM: G.Skill Trident Z5 RGB 32 GB DDR5-7200
4. Best SSD: WD_Black SN7100
5. Best graphics card: AMD Radeon RX 9070
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
