Good morning, everyone. I'm Eileen Mangalindan, Vice President of Corporate Communications at Intel. Thank you for joining us to discuss the press release we issued today regarding a collaboration between Intel and NVIDIA to jointly develop AI infrastructure and personal computing products. With me on the call today are Pat Gelsinger, Chief Executive Officer of Intel Corporation, and Jensen Huang, Founder and Chief Executive Officer of NVIDIA. At this time, all participants are in listen-only mode. After prepared remarks, we will conduct a Q&A session. As a reminder, this call is being recorded. The content of today's call is Intel's property. It can't be reproduced or transcribed without prior written consent. During this call, Intel and NVIDIA may make forward-looking statements based on current expectations. These are subject to a number of significant risks and uncertainties.
For a discussion of factors that could affect their businesses, please refer to the disclosure in Intel's and NVIDIA's most recent Forms 10-K and 10-Q, and the reports that they may file on Form 8-K with the Securities and Exchange Commission. With that, let me turn the call over to Pat.
Good morning, everyone. AI is driving a reinvention of every layer of the computing stack. Sixty years ago, IBM introduced the System/360, the first general-purpose computer. It launched the modern computing era, powered by Moore's Law and CPUs, programmed line by line by human hands. General-purpose computing has reached its limits. To keep advancing, we invented a new way forward. NVIDIA pioneered GPU-accelerated computing, increasing performance by orders of magnitude, tens, hundreds, thousands of times faster, while dramatically improving energy and cost efficiency. That innovation opened new frontiers in science and industry, and it sparked the Big Bang of artificial intelligence. Today, we're taking the next great step. Just moments ago, NVIDIA and Intel announced a historic partnership to jointly develop multiple generations of x86 CPUs for data centers and PC products. This collaboration will tightly couple and optimize Intel's x86 CPUs for NVIDIA's AI and accelerated computing architecture.
Together, our companies will build custom Intel x86 CPUs for NVIDIA's AI infrastructure platforms, our data center platforms, bringing x86 into NVIDIA's NVLink ecosystem. For personal computing, we're going to create new Intel x86 SoCs that integrate NVIDIA GPU chiplets, fusing the world's best CPU and GPU to redefine the PC experience. This partnership is a recognition that computing has fundamentally changed. The era of accelerated and AI computing has arrived. Today is a very exciting day and a very big day. Intel and NVIDIA are partnering to drive it forward. I'm delighted to partner with Lip-Bu Tan, long-time friend, and many of my colleagues at Intel in this great partnership, this historic partnership. Now, let me turn it over to Lip-Bu to tell you about the exciting partnership we're entering into.
Thank you, Jensen, and thank you all of you for joining. Jensen and I know each other for more than 30 years, and I still remember Jensen had the vision of building the system, platform, software, and with the CUDA, and the long-term vision that he had for the company. I have to salute to him. He's done a fabulous job building that AI platform, driving the whole new market opportunity. I'm so excited to be able to work together with Jensen to build a new era. This is a historical collaboration between the two companies, and I think this is a very big, important milestone. I call it a game-changing opportunity that we can work together. I would say that we are proud that NVIDIA is an investor in Intel, and thank you for supporting and confident in us and trusting us.
I think this milestone is the critical role that Intel can play in the ecosystem, and I'm grateful for the confidence that you place with us. In terms of product perspective, I just make three points just to add on to building on the Jensen comments. Number one, this collaboration is built on the core strengths of both companies. NVIDIA is a clear leader in AI-accelerated computing. Intel is a leader in the data center and client PC CPU. This collaboration brings all together for the best for the industry going forward. Number two, this collaboration unleashes the new era of x86 innovation, and x86 is a foundational role to play in the next era of computing. I'm excited about what can be created together in the NVLink stack scale solution that's based on x86.
x86 client SoC with NVLink interconnect and united memory, and also the advanced packaging that can bring it all together and then to make it together for making the company and for the industry. Number three, all of this is great for our customers. Our motto is very simple: make great products and delight the customer. That is what this collaboration is all about. Our teams are ready and excited to work together, both teams to make it a success together. Now I open up for some questions.
This concludes our prepared remarks. Operator, please open the call for questions. Thank you.
Thank you. If you would like to ask a question, please press *1 on your telephone keypad. If you would like to withdraw your question, simply press *1 again. Please ensure that your phone is not on mute when called upon. Thank you. Your first question comes from Jim Cramer of CNBC. Your line is open.
Yes, thank you. First, congratulations, gentlemen. Can you help us understand what the market landscape looks like with this product partnership, and how does it expand the opportunity and growth for both companies?
Yeah, thanks a lot, Jim. Lip-Bu, I'll take the first shot at it, and you could help me out. If you look at the AI computing world, let me take it in two segments, the AI computing world first. These are supercomputers. As you know, Jim, we recently introduced the scale-up NVLink 72 rack-scale computers. That involves designing a custom CPU we call Vera. It's tightly integrated with the GPUs, Blackwell GPUs, so that we can disaggregate NVLink switches, scale it up into a rack-scale system, and essentially have an entire rack behaving as if it's one giant computer, one giant GPU. In order to do that, we have to really customize the CPU to do that. This architecture, the NVLink 72 rack-scale architecture, is only available for the Vera CPU that we build, the ARM CPU that we build.
For the x86 ecosystem, it's really unavailable except with server CPUs over PCI Express. That has limitations in how far you could scale these scale-up systems. The first opportunity is that we can now, with Intel x86 CPU integrated directly into NVLink ecosystem, create these rack-scale AI supercomputers. The second thing is there's 150 million laptops sold per year, and NVIDIA's market is largely targeted squarely at gaming and workstation markets where discrete GPUs are used. We're very successful there, and we continue to grow there, and we're going to continue to grow there. There's an entire segment of the market where the CPU and the GPU are integrated, and 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. That segment has been largely unaddressed by NVIDIA today.
What the Intel team and I are doing, NVIDIA is doing, is that we're creating an SoC that fuses two processors. It fuses the CPU and NVIDIA's GPU, RTX GPU, using NVLink. It fuses these two dies into one essentially virtual giant SoC, and that would become essentially a new class of integrated graphics laptops that the world's never seen before. That segment of the market is really quite rich, and it's really quite large, and it's underserved today with state-of-the-art, world-class GPUs like NVIDIA is able to build.
I think, Jensen, if I can just add on to it, it's all about the scale, in terms of the best of GPU accelerator and then the best of x86, with the NVLink linked together and able to scale. Some of the markets that we both are doing well, we can expand even more in terms of some of the application solutions, the vertical market we can go after.
If I may just follow up for a second, Jensen, will you be using Intel's foundries to make high-end chips like a Grace Blackwell or, more importantly, a Vera Rubin, the kind of best-in-class semiconductors that right now you have made with TSMC?
We have always evaluated Intel's foundry technology, and we're going to continue to do that. Today, this announcement is squarely focused on these custom CPUs. With this partnership, with this agreement, we're essentially going to be a major customer of Intel server CPUs. This is the very first time. At the moment, we buy CPUs, the ARM-based CPUs from TSMC. The x86-based PCI Express CPUs are sold openly in the marketplace. In the future, we will buy x86 CPUs from Intel, and we would fuse it with NVLink into our rack-scale system. We are going to become a very large customer of Intel CPUs. The second thing is that we're going to be quite a large supplier of GPU chiplets into x86, Intel x86 CPU SoCs. In that particular case, we're going to be a supplier into a market segment we've never addressed.
In the server, we're going to be a customer, a major customer of Intel's, and we'll integrate it into NVLink 72 and resell CPUs, essentially. This is going to be a great growth business opportunity for both of us. Thank you.
Your next question comes from Ian King of Bloomberg. Your line is open.
Jensen, thank you very much for doing this. This is an interesting time we're living in. I wondered if you could talk a little bit about how long you've been working on this agreement. Obviously, you talked a lot about custom solutions and custom parts. Those don't happen overnight. When might we get to the point where we're seeing devices in the end market for sale, please?
The two technology teams have been discussing and architecting solutions now for probably coming up to a year. The two architecture teams, well, it's three architecture teams, are working across, of course, the CPU architecture as well as product lines for server and PCs. The architecture work is fairly extensive, and the teams are really excited about the new architecture. The teams have been working a while, and we're excited about the announcement today.
Your next question comes from the line of Michael Afton with Financial Times. Your line is open.
Hi, both. Jensen, why can't you commit to Intel as a foundry for your most advanced AI chips at this point? Does this pave the way for sort of deeper manufacturing collaboration or not? Are you confident Intel is going to get there? Secondly, I'd be interested, what sort of involvement did the Trump administration have in this agreement, if any?
The Trump administration had no involvement in this partnership at all, and they would have been very supportive, of course. Today, I had the opportunity to tell Secretary Litnick, and he was very excited and very supportive of seeing American technology companies working together. I think Lip-Bu and I would both say that TSMC is a world-class foundry. In fact, we're both very successful customers of TSMC. The capabilities of TSMC from process technology, their rhythm of execution, the scale of their capacity and infrastructure, the agility of their business operations, just all of the magic that comes together for being a world-class foundry, supporting customers of such diverse needs is really quite extraordinary. You just can't overstate the magic that is TSMC. Today, our conversation, our partnership is completely focused.
It is 100% focused on the custom CPUs that we're building for the data center that now has NVLink capabilities and can connect to the NVLink and the NVIDIA AI supercomputing ecosystem. It's completely focused on mobile SoCs or consumer PCs that now fuse the world's best CPU and the world's best GPU for consumer products. That segment, the first segment, the data center CPU segment, is probably something along the lines of $30 billion a year or so. In the case of, and this combination of Intel and our technology is going to address a fairly significant swath of that because it is the fastest growing segment. We all can agree that the future of computing is going to be about AI through and through. This is an exciting partnership for the data center market.
For the consumer market, it's 150 million laptops sold each year, and we're now going to combine the best CPU and the best GPU. It's really, really exciting. Lip-Bu?
Yeah, I think clearly this is historical. This is also my sixth month as Intel CEO. From day one, Jensen and I, we work on it, and then we accelerate that process. The two teams work together to this game-changing opportunity. It's a deep partnership, and we're looking forward to multiple ways we work together.
Next question comes from Laura Bratton with Yahoo Finance. Your line is open.
Hi, thanks so much for taking my question. I'm really curious about which manufacturing process these new CPUs will be made on. I know that Lip-Bu, you said that Intel 14A is only going to go forward if it has meaningful volume or customer commitment. I just wonder, can you comment about which manufacturing process the CPUs will be made on? I know you said this is strictly a product announcement, but can you all comment at all on whether this might pave the way for NVIDIA to collaborate with Intel Foundry Services in the future? That's all.
Yeah, I think this announcement is more on the product collaborations. Clearly, like Jensen mentioned, the TSMC has been a great partner, long-time partner for NVIDIA and also for Intel. We're going to continue doing that. I think this kind of, in terms of process, later on we can describe more, but right now we are focused on collaborations. At a certain date, we can have more announcements down the road when the product is ready.
Yeah, I think it's safe to say that the partnership that we're entering into is going to address some $25 billion, $50 billion of annual opportunity. This is a very significant partnership, and we're completely focused on that. One of the things that I will say is that our ARM roadmap is going to continue, and we're fully committed to the ARM roadmap. We have lots and lots of customers for ARM. We're building the next generation of Grace, the next generation Grace called Vera, and we have a next generation after that. We have exciting CPUs that we're building based on ARM. We're building ARM, of course, robotics processors. Our latest one is called Thor. It's used for robots and, of course, for autonomous driving. We also have a new ARM product that's called N1.
That product, that processor is going to go into the DGX Spark and many other versions of products like that. We're super excited about the ARM roadmap, and this doesn't affect any of that. NVIDIA's architecture, accelerated computing, covers just about every CPU architecture. Our most important interest is for whatever general-purpose computing platform that has market reach, we would like to be able to accelerate it to its fullest capability. Today, we have the benefit of partnering with Intel on a CPU platform that unquestionably has the largest enterprise industrial space, cloud, consumer footprint of any CPU in the world. This is a really exciting partnership, and we're going to revolutionize this general-purpose computing platform by adding and fusing the NVIDIA accelerated computing and AI computing architectures.
The next question comes from Stephen Nellis with Reuters News. Your line is open.
Thank you, gentlemen, for taking my question. I have a question for Jensen and Lip-Bu. Jensen, why did you feel it was appropriate or necessary to also make an equity investment in Intel along with this product collaboration? Lip-Bu did note sort of a string of equity investments we've seen from folks. Can we expect other ones from potential partners or foundry customers in the future?
I appreciate that question because we thought it was going to be such an incredible investment. This is a big partnership, and we think it's going to be fantastic for Intel. It's going to be fantastic for us. We're building revolutionary products that are going to address some $50 billion annual market. How could we, on the one hand, be excited about the products and how revolutionary they are, on the other hand, not be excited about the opportunities ahead? We're delighted to be a shareholder. We're delighted to have invested in Intel. 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. I think it's going to be fantastic for Intel. It just reflects how excited we are about this partnership.
Thank you, Jensen. To answer your question, clearly, as I mentioned, my top priority, top 10 priority, one of them is to strengthen my balance sheets, and that's I'm focused on that. Secondly, in this particular situation, first of all, I'd like to thank Jensen for the confidence in me. Our team and Intel will work really hard to make sure it's a good return for you. More important, it's a strategic partnership to drive the products and go to market together to win. That, I think, is very meaningful for us. Thank you.
The next question comes from Robbie Whelan with WSJ. Your line is open.
Hi, thanks to both of you for the announcement and congratulations. There have been a lot of questions about whether or not NVIDIA will someday use Intel as its foundry partner for its most advanced AI chips. With these CPUs that we're talking about under this partnership, will TSMC be fabricating most of those CPUs for the foreseeable future? Also, just really quickly, what's the bigger addressable market under this partnership, data centers versus PCs and edge computing? In other words, do you expect to be making more CPUs for PCs under this new arrangement or more CPUs for data centers? Lip-Bu, do you want to take the first part?
Sure. I think in terms of the, 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. In terms of the yield performance, Intel 18A, Intel 14A, clearly we want to qualify, and then we're going to decide whether this is the right one for doing our foundry. I think we continue to improve at the right time. Jensen and I will review that. Overall, I think we're going to continue driving our success on the process side and then win the customer's confidence and trust, one step at a time.
Yeah, I think one of the things, Lip-Bu, that I would also add is that Intel has the Foveros multi-technology packaging capability. It's really enabling here. The reason for that is because, as we all know, NVIDIA's GPU technology is based on TSMC's foundry. This is one of the extraordinary things that you can do, connecting NVIDIA's GPU chiplet with Intel's CPUs in a multi-technology packaging capability and multi-process packaging technology. It's really a fabulous way of mixing and matching technology. 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 multi-chiplet systems packages. I'm really excited about that. Yeah, in terms of the size and, oh, go ahead.
Yeah, thank you. I think, Jensen, you brought up a good point about our advanced packaging. Foveros and also the EMIB is a really good technology. We will definitely continue to refine it and make sure that's reliable and the yield improvement. I think that part definitely will explore the collaboration opportunity.
Yeah, with respect to the size of the market, the data center market and the PC market are both large. 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. I think, if my recollection is correct, the data center CPU market is about $25 billion or so annually. Just the notebook market is 150 million notebooks sold each year. That kind of gives you a sense of the scale of the work that we're going to do here. We're going to address, in terms of the consumer market, a vast majority of that consumer PC market, consumer PC notebook market. With respect to the data center, we're going to bring NVLink, which is the scale-up interconnect, the fabric of NVIDIA, the computing fabric of our company. We're going to bring that capability to Intel.
I think these are going to be revolutionary products. We're looking, I know that all of us working on it are super excited about it. The architects working on it are super excited about it. We're looking forward to telling you more about it over time.
The next question comes from Edward Dulac with Bloomberg. Your line is open.
Jensen, Lip-Bu, thank you very much for your time and willingness to answer questions. Jensen, I appreciate you talked a lot about the addressable market. Could you explain how NVIDIA participates in the economics of x86? Because you make money on ARM-based CPU, right? If you could explain how it will work for NVIDIA on the top and bottom line, I'd be grateful. Lip-Bu, congratulations. You've been in Silicon Valley, so to speak, for a long time, right? If you stand on the Intel roof, you can look across the freeway at NVIDIA. Would you just explain the culture and sentiment inside Intel today in reaction to this new partnership and what it means for the trajectory of you and your staff? Thanks.
In the case of the data center, the server CPU, it's like us buying Grace CPU from TSMC, integrating it into our rack-scale systems, and selling that. It's basically the same idea. We're now, instead of for x86, we don't buy any CPUs. We let the market sort it out. The CPUs are really sold as discrete servers, separate servers that are then connected with our GPUs in a data center. That architecture, basically using PCI Express retimers and things like that, basically PCI Express retimers and repeaters, essentially. Instead of building servers like that that really don't have the ability to scale up to NVLink 72 large fabric systems, we now are able to do that with Intel x86 CPUs. We'll buy those CPUs from Intel, and then we'll connect it into superchips that then becomes our compute nodes that then gets integrated into a rack-scale AI supercomputer.
In the case of our consumer PC, the current idea is to sell NVIDIA's GPU chiplet either in a pass-through way with Intel or sold to Intel. That is then packaged into an SoC. We buy a server die, server chip on the one hand, I guess server chip on the one hand. We sell a GPU chiplet on the other hand. In both cases, it expands the market for Intel very significantly, and it expands the market for NVIDIA as well.
In terms of your question about the culture changes at Intel, first of all, this is a new Intel culture I'm trying to build, and it's going to be engineering-focused. In extending my relationship with Jensen from Cadence NVIDIA partnership in terms of driving innovation, we are so excited to have this partnership to collaborate on the engineering on x86 and also the GPU accelerator and on the AI side and with NVLink. I think there's a lot of engineering collaboration together. The culture I want to have is really lean, fast-moving, so that we can match up with the Jensen fast-moving culture. I think that's something that I'm looking forward to, to build that culture that can match each other to drive the best solution for the market.
Once again, if you have a question, it is *1 on your telephone keypad. Your next question comes from Christina Partsinevelos with CNBC. Your line is open.
Thank you both for setting this up and taking my question. Jensen, you spoke about the ARM relationship, which is great. It's continuing, and maybe the reaction today was a little overdone. Given SoftBank's position across ARM, Intel, and now your partnership with Intel, is there just like some type of broader coordination that I'm missing here? I just have a follow on China.
Yeah, this today should have no impact on ARM. With respect to the second question, not that I'm aware of. There were no communications with anybody else except for between Lip-Bu, myself, and the technical teams that were working on this partnership. We kept it really quiet. Obviously, it's a very substantial partnership. This is going to expand the market opportunity for Intel in AI infrastructure that is largely unexposed to them today. It's going to expose Intel in the consumer notebook market where really exquisite GPUs are necessary. These two markets are unexposed to Intel today, and it's going to be brand new growth markets for Intel. I think, in all the due diligence that we've, between our two teams and all the work that we did, gave us a lot of confidence about the future of Intel.
We're really betting on, I've become quite a significant shareholder because we believe in this, and we have confidence in them to partner with us to create these amazing products. All of these discussions have no relationship to any of the things that you were talking about.
Just a quick follow-up. Intel definitely, we know, faces different regulatory constraints than you do. All the stories that are coming out of China are just constantly on CNBC. We're talking about it all the time. Is that part of the calculus here that Intel's difference in regulatory constraints would help you in the medium to long term?
I don't think there's any relationship there either. I don't think there'll be any impact either way. Thank you.
The next question comes from Anissa Garzi with The Information. Your line is open.
Thank you two so much for the time and to ask the question. My question is for Jensen. I was curious what types of NVIDIA customers are interested in the x86 architecture for the CPU, and do you expect any customers that currently use the ARM CPU to switch to x86 in the future?
ARM in the world's CSPs is growing, but the vast majority of the world's CSPs are still x86. The vast majority of cloud instances for enterprise users are still x86. I think the x86 footprint is still quite large. NVIDIA addresses it in one of two ways. In the case of ARM, we could scale up to a full rack, rack-scale NVLink system. In the case of x86, we address it through external PCI Express retimers, and we scale up to NVLink 8. In the case of x86, we scale up to NVLink 8. In the case of ARM, we scale up to NVLink 72. Now we could, with x86, scale up also to NVLink 72. I think this is a really great growth opportunity for both of us.
It also creates a product that for many customers who are still x86-based, and basically the vast majority of the world's enterprise is still x86-based, they now have state-of-the-art AI infrastructure.
The next question comes from Eva Du with The Washington Post. Your line is open.
Thank you for holding this. President Trump, manufacturing in the U.S. is such a big push for him, and both your companies have committed to multi-billion dollar investments in this area. Could you talk a bit about what are realistically the prospects of manufacturing your chips in America? Like what proportion of your chips do you expect to be made in the U.S. in the near future, and what are the challenges to doing more of the production here? Thank you.
Lip-Bu, would you like to go first?
Sure. Yeah, clearly, you know, we like President Trump's focus on manufacturing in the U.S. I think it's important to address that and then the opportunity we have in front of us. Meanwhile, we also have the footprint from Intel globally, so in a way, we just meet customer requirements, including NVIDIA, so that they have the flexibility which is best suitable for them. Meanwhile, we continue to improve our yield of performance. Also, the other part is the advanced packaging we just talked about. I think it's a great opportunity for both of us.
This concludes the question and answer session. I'll turn the call to Mylene for closing remarks.
I want to thank all of you for joining us today on this historic day, in a historic partnership. I want to thank Lip-Bu for his leadership and the management team of Intel that we've had the great privilege of working with, architecting two really exciting product lines and architecting this partnership. We are going to go and address a new computing era where accelerated computing and AI are essential to every aspect of computing, whether it's in the data center, in the cloud, or in mobile devices and personal computers. I'm super excited to start the projects and our partnership. I want to thank Lip-Bu again and the Intel team for this exciting announcement and this exciting new partnership. Lip-Bu?
Yep, thank you. First of all, I want to thank Jensen and NVIDIA for the trust and support of Intel. We will work hard to make sure that this will be a great success for both of us. I think more exciting for me is the collaboration, the best of the acceleration AI and also the x86, then using the NVLink to scale. I think this is a new compute platform that we are moving forward. I'm super excited about the opportunity in front of us and a lot of execution. We're going to be doing that. Stay tuned. We're going to update you when the time comes. I just want to thank all of you for attending this announcement together.
Thank you, Lip-Bu. Thanks, everybody.
Thank you, everyone, for your participation. That ends our call today. Thank you.