Intel Foundry Direct Connect Keynote (Replay)
As a reminder, today's presentations include forward looking statements subject to a number of risks and uncertainties. More information can be found on this slide and in Intel's filings with the SEC. I take a breath. I'm not going to lose.
I won't look down. Got no regrets. I was made for this. I was made for this. Born in the wild. Formed in the fire.
Built for the battle. I was made for this. I was made for this. Mind over matter. Silence the doubters I have the power. I'm a soul on a wire.
That's where I feel alive. Open up the skies. I was made for this. I was made for this. Born in the wild. Formed in the fire. Built for the battle.
I was made for this. I was made for this. Please welcome the CEO of Intel, Pat Gelsinger. - Welcome to Direct Connect 2024. And I am just thrilled and excited to have all of you here with us today.
Last week marked my three year anniversary as the CEO of Intel. And when I came back to the job, I said, we have three goals. We're going to rebuild this iconic company. The company that Grove, Moore and Noyce built. Second, we're going to restore the critical role it has and plays in the technology industry writ large. And third, we're going to rebuild Western manufacturing at scale, resilient, sustainable, trusted supply chains.
And I think of today as a day, three years in the making, as we open the doors on delivering on that mission. When I came back to Intel in 2021, we presented this event just a couple of months after my return called Intel Unleashed. We shared a perspective on the tremendous opportunities ahead. The plans that we had in motion for this fundamental rebuilding of Intel, setting a course for a new era of technology and innovation. We announced plans at that point to become a world class foundry, and we said we're going to be a major provider of US and European based capacity.
And when we rolled that strategy out, some were like 'yeah'. Others were like 'yeah right'. And others were just 'no way are they going to be able to pull this off'. They had fallen behind. They stumbled. Could they possibly get back to leadership? And our employees as we spoke to them, they said 'yeah we can do this'. We're going to reestablish that leadership position.
And our pessimists were challenging. Some of them I call perma bears. They were so far in hibernation they ain't ever coming out. But here we are today, three years later, and today this vision of unleashed becomes real. Today, the best of foundry, the silicon ecosystem. 'What could be today' is 'what is'. And with that, I'm thrilled today to announce, simply put, Intel Foundry.
This is a renaming, a reorganization, a new organizational model for Intel that includes our three major elements: technology development, leading, bending physics, our global manufacturing and supply chain, and our Intel Foundry services and ecosystem operation. All three together, Intel Foundry, a rebuilding of Intel. And for that, as I said, we're not fixing one company. We're establishing two vibrant new organizations. Intel Foundry, to serve internal and external customers at scale, to manage supply chains, to assure capacity, corridors. And Intel products,
our client data center and networking products. Two distinct and separate organizations. And as I think about that, they're dependent: the wafers, the factories, that leadership technology thrust. I need the product group and the foundry group working intimately together. Who's going to drive the wafers to run those fabs? Intel products. But they're also independent because you, the foundry ecosystem, need to have an independent.
Your designs, your fab, your intellectual property, your capacity corridors. In a word independent, in a word dependent or simply put, interdependent Intel products, Intel Foundry all under the Intel banner that begins today. And when we think about this, this era of AI, making it more accessible at scale and delivering AI end-to-end.
And as we say, from the data center to the cloud to the networking, to the edge and to the client, this AI continuum. And as I came back to Intel, I was sort of thinking about computing. And I'm young and vibrant yet, but it's been over 40 years since I've been in the computing industry. And with that when I showed up, it was like, this is boring.
You know, let's add a few more cores on the chip. You know, let's make PCIe a little bit faster. You know, let's increment the DDR. Boring!
And then AI happened. End of boring. This is transforming everything about computing. The explosion and systems and thermal envelopes and interconnect capabilities. Everything that we do is coming to bear in this AI period.
And as I think about Intel, we're engaging in 100% of the AI Tam clearly through our products on the edge and the PC and clients and in the data centers. But through our foundry, I want to manufacture every AI chip in the industry, those internal that are being done by the cloud service providers, those merchant providers, the technology providers. We are engaging in 100% of the AI Tam. And with that, we need a new model of what the foundry requirements are for that industry.
And that's, simply put, what we call the systems foundry for the AI era. Where semis are essential. This lowest layer, this world class foundry. And as you hear from us today, Moore's law, alive and well. Until the periodic table is exhausted, we're not done with Moore's Law.
Come on, get excited. We bend physics, we create molecules, we do amazing things. And Moore's Law is alive and well. And we're doing it through new transistors with RibbonFET, new power delivery, with PowerVia new advanced lithography with EUV and High-NA. But it's even better than that. Because now we're doing it not just in X and Y, A monolithic system silicon.
We're doing it in Z as well, with advanced packaging X, Y, and Z and new systems of networking and glass and optics and in package capabilities and thermals. But we have to do it a different way. We have to enable globally resilient, sustainable and trusted supply chains. And at the end of our conference today, I'm joined by Sam Altman. He's made a little news of of late. You might have heard of it.
And you know, I'm a pretty aggressive guy Man, am I conservative! And he's going to challenge the capacity needs of the industry. And Keyvan is going to spend some time talking today about our capacity strategy. This next generation of world class silicon process packaging is all at your disposal.
This idea of this coming together in advanced packaging, it's not because we said, hey, let's put it all into a package. Physics is driving us that way. We're thermal envelopes. Interconnect. Speed of light is driving us to put these together. Iraq is becoming a system. A system is becoming a system on a chip or a package. And the system of chips needs new capabilities, new tests, advanced packaging capabilities. And that's what a systems foundry does.
And Stu is going to describe today how we're making all of this real and available to you. Every aspect of Intel to enable your innovation, bringing all of this together. And the people, the systems, the processes, that we're standing up with Intel Foundry - to ensure your supply chains, the confidentiality of your information, to do it with your EDA and IP partners.
And throughout the day, you're going to see this vision come to life. And a year ago, as we began this next phase of our journey, we said, we need people who are advising us on that journey. And it was a true pleasure of mine to go to some of the Intel board members, but also some of the leaders of the industry, to help us set this up. To assure, the establishment of the Intel Foundry. To assure that we've established the business processes to change the DNA of Intel to become maniacally customer focused. As we look forward to assure that we have efficient operations.
And my advisory board is here with us today, and I'd like them to stand up and acknowledge that these are the people that are helping guide us. So over here. Joe Kaeser, former CEO of Siemens. Chi-Foon Chan former Synopsis Co-CEO. My many decades long friend Lip-Bu Tan. (brother and friend) And over here. Let's get a light over here to Tsu-Jae King Liu,
one of the most honored members of the semiconductor research and academic community. So these are the people that are helping advise us and guide us on the journey to set up and become, what I've called, the world's #2 systems foundry. So thank you for joining us on this great journey. And as we describe this vision, to become the #2 foundry of the world, we realized there's only a few companies that can do this, that have the capital capacity, that have the R&D, the longevity to go do this. And we announced this vision in the middle of Covid. Boy, did we realize how much we needed resilience at that period of time. And now as we're in an economic downturn.
Ah, welcome to semiconductors. This is a tough business. And then the AI explosion and the cyclicality of the industry, and we've seen the geo instability and active wars and Israel and Ukraine and the tensions in Taiwan Straits.
This is anything but a resilient supply chain today. And even as we seek to be peacemakers, to see global prosperity, we know that the best way to do that is through resilient, sustainable supply chains. And for that, we said yes, get to be a large, meaningful - the second largest foundry. Bbecome the world's most sustainable foundry, the world's most resilient foundry. Because that's exactly what the world requires. And this exponential explosion of AI demands a new way of delivering silicon with the lowest possible environmental footprint.
And as part of Intel's climate transition action plan, we're approaching 99% global renewable energy. 100% renewable in the US for over a decade. Driving conservation in our energy, the water management processes, responsible use of chemicals, reducing, abating and replacing greenhouse gas chemicals. As I like to say, Gordon was an environmentalist before it was cool. He built it deep into the DNA of Intel.
We don't do this because it's driven by some expectation in the industry. We do it because it's the right thing for humanity and for the planet. At one time, my daughter-in-law was giving me a bad time. And hey, you're sort of proud of all the great advancements that you drove, huh? And then she said, but did you have to bankrupt my planet to do so? And I think all of us are left with that same question. Are we going to be responsibly turning over our planet to the next generation? And in the face of things like this explosion of AI capabilities, it isn't just doing good things, it's doing the right things for tomorrow as well.
So sustainable foundry, but also resilience. And we're seeing global politics has been dominated for the last 50 years by where the oil reserves are. Covid was a big wake up call. For us all you know, an auto factory building a $50,000 car, being stopped for a $1 semiconductor. How did that happen? Semis advanced computing is to the world's geopolitics, what oil has been for the last 50 years. And silicon, fortunately, isn't restricted to where it's found.
It's the second most abundant element on Earth. I call it God's gift to humanity. This 4x4 crystalline structure that we can bend and shape and shove different elements into. It's just magic that these tiny chips are enabling the modern economic cycle that we are in today.
But stunningly, in 1990, 80% of the semiconductors were built in US and Europe. Today 80% in a small concentrated area in Asia. We've seen this long, steady decline, in terms of our supply chains for the world. Nothing should be reliant on a single port to a single country, a single place in the world. We need resilient access to supply chains and capacity in the right regions at the right time.
And thus the choice, the opportunity to drive systemic change in where and how we drive the most important aspect of our future - where the technology supply chains are. And as I've liked to say, the moonshot is 80 to 20, 20 to 5050 in a decade, rebalancing the supply chains of the world. It was a proud day when I got to stand on the White House lawn when we signed into law the US CHIPS Act. A moment to rebuild the supply chains of the world. And with that, it was a thrill to do so with the President.
But the tireless championing that we saw through Secretary of Commerce Gina Raimondo. And it's my pleasure to have her join us now here at Intel Direct Connect. Secretary Raimondo. Thank you Gina. Thank you Secretary. It is such a pleasure. And you and I agree how important the Chips act has been, and I'd love for the audience to hear directly from you, your perspective and your passion for this piece of industrial policy. - Hello, everybody. Pat, I love the energy.
I love listening to your presentation and it's exciting. We've never quite done anything like this before. Or at least not for a long time. If you want to find a precedent of when the United States government has acted this strategically in terms of industrial strategy, I think you'd have to go back 60 years to the space race. I think that's the closest parallel, a time when the federal government came together with the private sector and academia at every level to spur innovation and ensure America's technological leadership in the face of fierce competition. Of course, then it was Russia.
Now it's other countries. That's what this is about. You just said a couple of minutes ago, after Covid, the vulnerabilities in our semiconductor supply chain were on full display for the world. The fact that we are so overly dependent to a couple of countries in Asia to access semiconductor chips that we need for life saving medical equipment, cars, every piece of technology showed us, we got to get to work. We need to get back to work making more chips in America. And so this is a moment, our generation, this is our moment. And it's making more chips in America, bringing back the silicon to Silicon Valley if you will. But it isn't just that.
It's then getting a massive flywheel going in the United States of America. From research and development all the way through to advanced packaging and everything in between. So and I hope in the process, we will create hundreds of thousands of high paying jobs for America. But also if we do our job right Pat, and I guess I'm talking to myself here, but I have to do it in partnership with you and the folks in your audience. Ten years from now, engineering degrees all over this country will be taught in a different way, will be tailored for the chip industry. It'll be exciting. The top graduates of our schools will say,
I want to go work in a chip factory and a chip design operation in the United States. And so I'm psyched about it. And you've been a fantastic partner. - So say a little bit more about building those. So say a little bit more about building those chips in the US. Why so important? And a little bit about the AI surge and how you see that affecting the Chips act and affecting the AI future? - I want to be clear. We can't and do not want to make everything in America.
We don't want to make every chip in America. That isn't a reasonable goal. But we do need to diversify our semiconductor supply chain, have much more manufacturing in the United States, particularly of leading edge chips, which will be essential for AI, as you well know. We need to just have more resiliency and more diversification. So the goal isn't to be self-sufficient, to produce and package everything we consume in this country. But we do need a self-propelling engine of innovation and production.
And that fell out of balance. We maintained our leadership in the design and in the software, etc. but we sacrificed our manufacturing capacity and with that, our capacity to conduct the advanced research and development that we need to maintain our global leadership. Now, people ask me all the time, how do we compete with China? We run faster in America. We out-innovate the world. We manufacture in the United States. We expand and modernize our manufacturing capacity.
And that's what this is all about. With respect to AI I think that's unbelievable. First of all, as an American and as a US secretary of Commerce, it kind of gives me the chills to realize how the United States of America leads in AI, because of our great entrepreneurs, in the whole stack - up and down the stack. That's a source of competitive advantage.
And we need to continue to invest in that. But for for you, frankly, it's a source of great upside and opportunity and customers. When I you mentioned Sam Altman, when I talked to him or other customers in the industry, the volume of chips that they project they need is mind boggling. Even if you take their projections and cut them in half, it's still mind boggling. So, you know, it's exciting.
It's exciting what AI can do to advance, technology and medical care and such. Of course, we have to keep a lid on the risks, which I'm hard at work on. But I think it's really a perfect convergence for you at the time, you have IFS that, the demand for two and three nanometer and 18A is just going to explode. - Well, thank you. And maybe say - the chips act -
Are we done? Do we have all the R&D underway? Do we need a CHIPS-2 Secretary? - I'm out of breath, running as fast as I can to implement CHIPS-1. But, and it's amazing. We were at global foundries yesterday. A billion five investment. That's the third of our investments. And I think, there'll be a steady drumbeat of of those announcements to come in the coming weeks and months.
We announced a week ago that we've launched Natcast, which is a purpose built nonprofit. It'll be a public private partnership. We announced Deirdre Hanford, who many of you know, she was at Synapsis for a long time. She's going to run this. And that's going to be our hub for research and development and workforce training. We will be putting billions of dollars into that.
All of that being said, I suspect there will have to be, whether you call it CHIPS-2 or something else, continued investment. If we want to lead the world, look, we fell pretty far. We took our eye off the ball. We used to manufacture 40% of the world's chips, in this country. We used to manufacture-- you know the numbers better than I, but I don't know, what, 12- 15% of leading edge chips. And now we're down to almost none in this country.
So if we want to really compete globally, we're going to have to continue to invest. - Yeah. Thank you. And, maybe, just as we finish up here, obviously we haven't announced our chips grant yet. Very soon. We're making that happen. But maybe just a final message for this audience, and here, the Intel Foundry Day. What's your message to them?
- Get excited. Get ready. Get ready to lead the world. America is a great country, because we are competitive. We out-innovate, we out-compete. And what CHIPS is about is much bigger vision than just, incentivizing ten new fabs all around the country.
If that's all we do, then shame on us. What this is about is getting the flywheel going, developing a deep and sustainable ecosystem again in this country. Intel is this country's championship company. It's an American champion company with a very huge role to play in this revitalization.
And I'm just, hum, I'm excited. I'm ready to go. We did a big announcement yesterday. There's more to come. And I believe we will be very, very successful. Well, thank you so much for joining us today. Happy to be here.
Thank you Gina, my friend, my partner in this great, incredible journey. Thank you again. You know, we've described this as, this critical underlayment. And I just ask you, what aspect of your life is not becoming more digital? Well, everything is your health care, your financial, your social. And with that,
we simply call it Siliconomy. Silicon and the economy becoming fused together in an inextricable way. Today, you know, 15% of every economic endeavor. It is expected to be 25%.
And that's before we account for the implications of AI driving that more rapidly. And as stewards of Moore's Law, we see this relentless pursuit of more efficient, more capable, more scalable computing. And for that, we've been on this journey. And when we announced our, you know, Intel Unleashed, we also announced that we're going to get five nodes in four years. We're going to do something unheard of in the industry to return Intel to process technology leadership.
And while we're not finished today, we see the end as soon in front of us on that journey. And Intel 7, you know, shipping and ramping in volume. Intel 4 with our Core Ultra launch, shipping and ramping in volume.
Intel 3 is production certified and will be with our server products launching in the first half of the year going into volume production. So with this, we've gone on an incredible, you know, journey, but then it continues into what we call the Angstrom era. And for this, you know, Intel 20A and Intel 18A, the adoption of RibbonFET, a new transistor structure, of PowerVia, a power delivery technology, the embrace of EUV. And for this, the first major new transistor rearchitecting since 2012. And I'll tell you, I've been studying SEM diagrams, you know, for over 40 years. This is a Mona Lisa. No, no, this is a Rembrandt.
No, no, no, I think it's a Michelangelo. Right? Sculpted in silicon. Right. For these truly are works of art. And I am thrilled for the progress our TD teams are playing to bring us back to technology leadership. And with that, the finish is 18A.
And with that we've already sent into fab our first 18A products. And this is a test chip for Clearwater Forest; Clearwater Forest is Intel 18A top die. But then we're putting it on Intel 3-based die, a construction that you saw in the cartoon there that's being broken out that takes advantage of EMIB. It also takes advantage of Foveros direct,
the first copper-on-copper without solder ball bonding. That will allow us to go to below ten micron pitches between the interface of top and bottom die. The first ever time. You know you're going to hear Choon talk today about our assembly and test technology, singulated die testing, advanced packaging capabilities, but the testing to go with it as well, you know, and how we're working with our partners and EDA ecosystem that are represented here. Synopsys, Cadence, Ansys, Siemens to enable these capabilities, delivering these advancements in technology, but also not just in our products, but for the first time ever, making them available to the entire industry. So I'm thrilled. This is what we call a family photo moment. So kids, come to Papa, here we go. Five nodes in four years.
And I do want to announce, describe, you know, and give a moment to our latest, newest 18A customer, my decades long friend Satya Nadella, speaking for Microsoft as the newest 18A customer. So let's hear from Satya now. Thank you so much, Pat. It's great to join you at your launch event. It's clear that we are in the midst of a very exciting platform shift that will fundamentally transform productivity for every individual organization and the entire industry. To achieve this vision, we will need a reliable supply of the most advanced, high performance and high quality semiconductors. And all of us at Microsoft are committed to supporting Intel's efforts to build a strong supply chain right here in the United States. That's why we are
so excited to work with Intel Foundry Services, and why we have chosen a chip design that we plan to produce on Intel's 18A process. We look forward to sharing more details in the future, and I can't wait to see all that we will be delivering together for our customers in the years ahead. Thank you so very much. Thank you Satya. And you know, now that, you know, we see the light at the end of the tunnel and finishing five nodes in four years. Are we finished, now that we've gotten back to leadership? No! having gone on this grueling trek to five nodes in four years, Moore's law, alive and well. And you're going to hear more about that today.
So can I tell you more about what comes after 18A? Yes. And today we're announcing that we're extending these nodes. We're adding major and minor nodes to it, a combination of older and leading edge nodes to ensure our customers have access to the process technology they need. Today, we are announcing Intel 14A for the first time. You can think about this like 1.4 nanometer technology, but Intel 14A, venturing deeply into the Angstrom era. 14A, first processing.
But we're also announcing that we're extending our nodes, as you see on here, adding P nodes, enhancements to those existing, you know, adding performance capabilities, adding T nodes, through-silicon-via, new feature enhancements with E nodes on the roadmap, filling out that roadmap of capability. You know, today we're announcing for the first time Intel 16-E, enhancements to our Intel 16 technology as well. So we're filling out that full set of nodes in the roadmap that we have to go beyond it. And for that we're, you know, continuing to see, a broader set of portfolio of interest from the industry.
And we're thrilled to have Jason Wang here, the UMC president, we just announced the partner. Where are you, Jason? Here. You're supposed to be right there. So right. So but Jason Wang from UMC joining us as well, a partnership to add a 12 nanometer node, you know, building out a real foundry roadmap and working with our customers to continuously evolve our offerings on Intel's leading and mature node technologies. But as we've seen as we've gone through this period of time, this AI era explosion, wafers are cool. Packaging, something that we sort of kept in the corner of Intel and private, has gotten to be really cool.
So Intel Foundry offers a broad set now of advanced assembly and test technologies. And we're under volume ramp. You know, we're seeing more and more customers taking advantage of the capacity and the technology that we offer to optimize power to innovate in X, Y, and Z as well. Being able to take these technologies of EMIB and Foveros, and Foveros Direct, bringing higher density pitch and working on next generation. Intel, 25 years ago, drove the standardization of organic packages. Now we're driving the next generation of glass based packages, and with that, the ability to directly interface with optics and waveguides directly into the package construct for the most advanced system capabilities as well.
Choon Lee, this afternoon, we'll give, you know, a full roadmap of these capabilities now available to the industry. And we've seen extraordinary interest from customers and the momentum that they have. You know, Intel Foundry has added a number of additional AI customers to our portfolio of packaging offerings as well. AI era needs advanced wafers, but it even needs more systems and packaging capabilities, an area that Intel is the clear leader in.
And this now includes, as customers, some of the largest AI leaders in the world. And I'm happy to share that we now have expected lifetime deal value of over $15 billion in our foundry customer business into the future. Including the Microsoft announcement today, the advanced packaging customers that we described on leading edge nodes as well as mature nodes.
We're continuing to see great response from our customers for Intel Foundry. You know, as I conclude my time on stage. Today is a day, three years in the making. And I couldn't be prouder of the team at Intel that has rallied behind this rebuilding of this iconic company. And you're going to hear from a number of those leaders today, bringing together the world's first system foundry capabilities for the AI era. And with that,
it's my pleasure to introduce to the stage a friend for decades now, and the zealous leader of our foundry services, none other than Stu Pann. Please welcome Stu to the stage. Good morning. Thank you all for coming this morning. What I'd like to do is break down that top sentence. "A systems foundry for the AI era." First off, I want to show you why we believe we are a foundry. We have the table stakes.
We have the ability to deliver what you need when you need it. I want to establish what is a systems foundry. A lot of questions. We've talked to the press like...
Systems of foundry. Foundry of systems. What does it mean? And then I'm going to talk about why the AI era is driving significant upside demand for all of us and is an incredible tailwind for us. But first, I want to offer up a personal perspective on Intel, the company. Um, the thing that the IBM PC and I have in common is that we both came to Intel in 1981. It's a long time, 43 years. And so I have a sense of history about this place.
And one of the things I'd like to talk about first is the people, because it's the people that make this company. Our people can sense and feel this momentum. You can hear it in their voices, you can see it in their actions.
You even get the occasional high five in the factory. They know that you don't need to be in Taiwan to build the world's most advanced semiconductors. They've realized this five nodes in four years. This audacious goal that Pat laid out. You can see the check marks across all the boxes here. So it's an honor to represent the 52,000 scientists, engineers, factory people, logistics people, imagine, all this organization.
We're ready to earn the right to be your foundry supplier. Nobody's going to give us that. We want to earn it. So that's one perspective. I want to offer up a second perspective. And this comes from Chris Miller, the author of the book Chip War. This is a New York Times best seller, Financial Times Book of the year. It took Chris ten years to write it.
And it's incredibly timely considering what Secretary Raimondo just talked about. He talks about the development of transistor on up through, like, you know, the last few years. Now, Chris, in his speech in October, said this quote to the employees, he said, quote unquote, Intel is the most important company of the last 50 years. So I called Chris and I said,
are you okay if I say this publicly? He said, yeah, okay. Go ahead. Good publicity for the book. Um, you know, I've been associated with Intel, I said, for 43 years, 37 years as an employee, six years as a chief supply chain officer of HP. I came back to Intel three months after Pat did, because I believe this quote to be true. And I believed it long before Chris said it.
And what you're going to hear in the next 40 minutes or so is all the reasons why Chip War season two is about to begin. So let's start from the beginning. Gordon's original paper was called Cramming More Components on Integrated Circuits. It was published in 1965. Every time I read it, I'm stunned by how prescient he was. But if you read all through the paper and by the way, I know Pat talks about physics and Moore's Law and periodic tables.
Read the paper. It's also about economics. And in fact, it's the foundation for Moore's Law. It's about driving strong economics. In the last section of the paper, Gordon's remarks of we've come to a day of reckoning. Quote unquote. It may prove to be more economical to build large systems out of smaller functions, which are separately packaged and interconnected. And this is why I believe we're in the area of the systems foundry. You can no longer do just monolithic devices.
You have to break it up. You have to adjust for thermal profiles. You have to address cost, you have to have flexibility. And this is all driving a new level of systems thinking into the foundry business. And at Intel, we're no strangers to systems thinking. We started in 1980 with Multibus, 1990 with PCI express.
The idea that you could put a common bus architecture inside of a PC and create whole new markets. We moved on to Centrino, where wireless created an explosion of notebook demand. And today we're at the AI Systems Foundry. This idea that you have to do open architectures, reference platforms, open standards in order to foster all the innovation you need to go meet the demands of what AI requires.
In this systems era, you not only have to have open standards between devices, you have to have standards on the device. Think about the fact, to do a training model today requires 100,000 CPUs, all running in concert, all on the same data set. The next round of training models will require a million CPUs. That's something that requires standards everywhere, from chips inside to chips outside. This is what's necessary for our customers to succeed. And this is the fundamental reason why Intel is a systems company turning into a foundry, not the other way around.
And that's what makes our entry in this market so powerful. So now that we've described, you know, system thinking, let's look at how the original foundry model was created. Now, since C. C. Wei has been kind enough to mention us in his last few earnings calls, I thought I returned the favor and talk about TSMC in my presentation. Now, this was a presentation that Maurice Chang gave at MIT last October, and in it he describes the foundry model.
He says it's research and development, it's wafer fabrication, it's advanced packaging. And he says, in the red line, that's what TSMC does. And he said in the blue line, Intel, everybody else does everything else. You might even call that an IDM 1.0 kind of manufacturer. Now, TSMC has been incredibly successful with this model.
Disciplined execution, disciplined strategy, consistent innovation. But to quote Bob Dylan, the times, they are a changing. Let me tell you why. There's an idea of systems, technology, co-optimization where you look at application workloads, software, system architecture, memory, interconnect, advanced packaging, packaging technology, core, all these different things. What happens today is people focus on their layer and maybe the layer up on top of it.
And in fact, this is really what a classic foundry does today. But we're now in the realm of the exponential. You know, Sam Altman is going to come out later on this afternoon and talk about, he doesn't have enough capacity to do what he wants to do. And as we talk to our customers around the world, we're now realizing that it's not just enough to do Moore's Law.
It's not just enough to do systems, you know, kind of implementation. You have to look at all of these combinations. You can get a couple of things right, but to do a system that coordinates the activities of solving a training model across 100,000 CPUs, requires you to get all the gear ratios right. If you're mismatched in memory. If you're mismatched in networking, you wind up throwing away valuable cycles and valuable resources. So we've got to get 100 x more out of what we're doing.
So we describe the evolution. Now let's build out the strategy and talk about what's the revolution in all of this. We'd like to think about this in three basic layers. First off, we have to be a world class foundry, right? Roawen Chen, the COO of Qualcomm, talks about the fact that silicon speaks and silicon speaks in four different ways: performance, power, area and cost. Without that, you're not in the business. And so we're absolutely focused on making sure we are world class across all four of those dimensions.
On top of that, you have to have an ecosystem. We're going to talk about the ecosystem at length throughout the day. That ecosystem is what allows you to take advantage of PPAC. The next layer of the triangle, uh, the pyramid is all about resilient, sustainable supply. And that's what Pat just talked about earlier.
And that's what Secretary Raimondo talked about. This need to have capability around the world to build this and to build it in a sustainable fashion. The top layer of the pyramid is this idea of systems of chips. We're putting system inside a chip, and we're working with our foundry partners to create systems of chips. So let's walk you through what that means.
Now, Pat obviously was very proud of his children, his grandchildren or relatives. I'm not quite sure. Um, so I won't go into a lot of detail on this, but I will point out to you and you'll hear from Ann a little bit later on this afternoon, that what we do with 18A has been incredibly well received by the Hyperscaler community. Why? Because 18A has two key attributes. One, it has, we think, the world's best transistor structure, but also it has this idea of PowerVia, this idea that you supply power to the bottom of the device. Why is that important? Because these AI devices are thousand watt devices. They take hundreds of amps of current.
To do this the right way means you have to come up with a totally different way of developing backside power. And I will talk to you how we've developed that and why we think it's so robust. Pat mentioned we're extending out our node families. The thing I want to spend a little bit of time on, since you're going to hear a lot more about the roadmap, is the bottom, where he talked about Tower, the partnership we signed with them just recently for our New Mexico factory and what we're doing with UMC. What's so important about those partnerships, those collaboration agreements, is that they give us a way to load balance our factories, to give us the best possible cost structure, take advantage of the investments we've made in tools and people and buildings. And rounds out our cost structure dramatically. And in fact, we just are now in conversations with Tower about ways we can develop the next extension with them, possibly even 40 nanometer, just started this recently. This ability to take advantage of this world class factory network and extend it in multiple ways is the thing that helps us get the C part of PPAC. The cost part of PPAC.
By building at scale, we'll absolutely be able to hit that. The next part of the ecosystem. Is around all the folks you saw walking in this hallway, the 30 plus suppliers, the EDA vendors, the IP vendors, the people bringing complex IP so that when you get our process, you have the capability to design all sorts of things. Because we have a rich assortment of IP and EDA solutions. We're not asking you to pick a certain path. We're making sure we cover as many paths as possible.
So the way you do your business now is the way that you can do it with us. We round that out with design services and cloud providers. We just announced a partnership with Faraday in order to help our customers build out complex Asics. Our cloud providers run all these EDA systems in a cost effective way. And lastly, we're creating a separate ecosystem for the military and aerospace customers. It's important that their needs are taken care of.
So folks, with Cadence, Synopsis, with Siemens, Flex Logix, Draper, Trusted, these are the folks who help us develop specific things for military applications. And it's a very powerful thing for us as a country when they do that. So this is how we make it easy for our customers. And I would encourage you as you're walking out through the showcase. And by the way, we sort of made it so you had to walk past them, duty free, same kind of thing. Um, talk to them. They have made an incredible investment with us,
and we're going to talk a lot more about this a little bit later. Now, the next layer of the pyramid. This idea of resilient, sustainable supply. You know, Pat touched on this and later on in the afternoon, Keyvan will lay out targets for what that means, because sustainable supply is not about PowerPoint slides. It's about a culture that works at this for decades, and I would strongly invite all of our competitors to match the targets that Keyvan is going to lay out. They are tough.
They are demanding. They are essential in this world where we all want sustainable, resilient supply. Now there's another box in here on security.
Why is security so important? In a speech given by Matt Kay at one of our federal conferences a few months ago. Doctor Kay talked about the need to give our troops asymmetric advantage in the battlefield. What does that mean? It means we give them today's technology today, not technology that was done ten years ago. And, you know, in our discussions with the defense community, they've always wanted this, but we haven't found a way to take the things that they regard that give us that asymmetric advantage, the things that are highly confidential.
How do you build it cost effectively? How do you get it out on time? By working closely with our DoD partners, we have solved that problem. And that's one of the reasons the US government gave us $1 billion contract a few months ago, was to go take that to the next step, to go be able to create this trusted, secure environment. So the things that our colleagues in DoD know about, the things they want to embed in our silicon, but the things they want to protect, will be done so in a secure fashion.
And it's super powerful because we all want to give our troops asymmetric advantage. Now, I talked earlier about this day of reckoning. What that means it's happening now. We can no longer do designs at a monolithic level. We are now at radical limited designs, design sizes that are 800 square millimeters of silicon.
The vast majority of customers we talked to are absolutely moving this idea of disaggregated design, because we have to move around beyond recall limits and thermal constraints. And by the way, even cost constraints, because when you're building these really big die sizes, they're really expensive. Isn't there a better way to take advantage of this? Pat talked about with Clearwater Forest to take smaller tiles on the more advanced nodes, get better yields out of them, package them together, have more flexibility. To do all that really requires, if you will, a system on a chip. So you can see in the animation here how we build it out.
And this is literally how we're going to build out in the factory. The idea of a substrate, the idea of base dies, the idea of logic tiles, the idea of IO tiles on the side. Why do this? It gives our customers the ability to optimally trade for what they need for their design.
The things that you do for a training engine will be different than what you do for an inference engine, and only by having all these levers to go pull, can you get this done. Now. We learned a lot through a device called Ponte Vecchio, or as the branding people call it, Intel Data Center GPU Max series. I call it Ponte Vecchio. It's an SoC. It's a billion transistor, 100 billion transistor SoC.
It's dozens of chiplet tiles, 47 of them. It's multiple suppliers, by the way. We coexist with TSMC in the same package. Right.
We developed testing techniques to go off and do that. We do this idea of singulated die test. What does that mean? It means every single die that goes into that package is a known good die. Why is that important? Because you want every one of them to be good.
The assembly test yield on this device is 95% +. It is the Super Bowl of integrated design. Now, what do you do with it? Well, if you're Argonne National Labs and our partners at HPE, you build a really big supercomputer and they build a computer that was 66,000 Ponte Vecchios, 20,000 Sapphire Rapids. And it looks like this. Okay. What do you do with something like that? Well, you solve some really hard science problems.
If you want to model the airflow across the wing, you can do that on a workstation. If you want to model the airflow across a plane, you do it on this. If you want to model fusion reactions, which are pretty tricky things to model, you do it on a device like this. If you want to model, if you want to model cancer curing drugs at the molecular level, you do it like this. It's 600 tons of compute.
It's four tennis courts. It is the weight of an Airbus. It has 300 miles of optical cable. It takes 34,000 gallons a minute to cool.
By the way, your faucet at home, it's a gallon a minute. 34,000 gallons a minute. So when we talk about how to design this stuff, we have to find ways to make this more power efficient, to make this more cost effective.
This today is the second fastest supercomputer in the world, by the way, at 100,000 CPUs, roughly. You know, to handle the demands of AI, we're going to go far beyond that. And that's why this idea of scalability is so important. Now let's go through the top of the pyramid. This idea of systems of chip. You know, systems a chip requires. Pat mentioned great packaging technology, and we're making that packaging technology available to all of our customers. It requires this idea of standards.
We've created a standard called UCIe, which allows chip to chip connectivity. Think of it as the PCI express of what we did back in the mid 90s. This idea that you can add, mix and match, by the way, you can mix and match different foundry suppliers. We like that because we're sort of an underdog in all this. Some of our competitors, not so much, but. Our customers want this kind of flexibility.
And we're going to talk extensively about this later on in the afternoon, Choon is going to talk about optical interconnects as an example. But there's one thing I really want to point out at the bottom. And that's that little IEEE symbol for Ethernet. So as you can see from the rendering, what we're doing is taking, systems on a chip.
And with new Ethernet standards, we are working closely with a number of partners. You are going to be building systems of chips. Why do you need systems of chips? Because it's the same problem I mentioned earlier. You need to train models with 100,000 CPUs now and a million potentially down the road, and then maybe up to 10 million. When you're moving and coordinating data across all of these devices, you need to have standards and connectivity.
And that's what we provide. And really, when you think about it. What is? What's it going to take to bring AI everywhere? How do you make it cost effective? How do you make it capital efficient? You know? Yes. You know, I know Sam's asking for trillions, but we want to make sure he's spending all that money in the most cost effective manner possible. So let's break that out.
You have data center chips doubling year on year. But the efficiency needs of the things that are really eye catching, you know, New York Times ran an article that AI could soon need as much electricity as an entire country. So I'm sure you're like, which countries? Sweden, the Netherlands, Argentina. If you were to run all of the AI servers that market estimates have on DGX-2, DGX H100, those kind of devices, you would take 85 to 134 terawatt hours, terawatt hours. By the way, the great State of California,
its entire power generation capability today is 30 terawatt hours, for the entire state. So bringing AI everywhere is going to require us and our foundry partners to figure out how to do this cost effectively. And this is why the essential elements of an AI foundry are these things. Start out with us being the stewards of Moore's Law, right? The idea that we're going to double the amount of transistors every couple of years and make them power efficient. Add that with continuing systems innovation.
You know one size won't fit all. And in this AI era, foundries are going to have to do a lot more. So I'd like to do is sort of build that out for you for the next five years, because this is what this roadmap is going to take.
And this, I think, is what makes us different, than other foundry approaches. You have to start off across, first off, with table stakes. As I mentioned earlier, you have to start off with a great process and with what you'll hear from Ann, and Choon, Keyvan, we have a great process delivered at scale. Add to that packaging, which is, we believe, a unique differentiator for us, why? We build a lot of server parts, and we're taking everything that we learn from our server business and offering up to our foundry partners. These are table stakes, and this is what it takes to be a foundry player.
And we'll innovate for the next five years. And I think this is what one of the conversations we have with customers that's so intriguing is, we're not just one and done. We have this planned out literally across all these dimensions for the next five years. So on substrates, Intel is the world's largest consumer of substrates. So as such, we play a pretty influential role in how substrates get driven. Later on today, Choon is going to talk about this idea of glass substrates.
Why do you care about glass substrates? You can put a lot of parts in a package on a glass substrate. It doesn't bend. And we're going to work closely with our partners to give the world a new way of looking at substrates through glass.
Let's talk about cooling. I mentioned that Argonne needs 34,000 gallons a minute to cool it off. The next wave of devices are going to have to be immersion cooled.
And today, Intel's Xeon product line is the only product that offers an immersion cooling warranty. Immersion cooling allows us to deliver power much more effectively in a data center. We're going to take what we've learned there, and we're going to offer that up to our foundry customers so that when we start looking at 2000 watt devices five years from now, we're going to have a way to cool those.
Memory. Pat mentioned this idea of a base die technology. We're working with all the major memory manufacturers right now. How do we optimize that interface for HBM 3E, for HBM 4, how to perhaps, do we put memory on the device itself? So it's more computationally cost effective. And over the next five years, you're going to hear us talk a lot about new technologies and new ways to increase memory bandwidth while decreasing the need for energy consumption. A five year kind of look. Interconnects.
Interconnects between chips. The idea of having high speed SerDes, high speed interconnect. All the things that you expect from a foundry provider is what we're going to give you, and we're going to not plan out for just what's out there today. Keep in mind, as a standards company, we do this across all different standards, and it's our job to make those standards available to all of you. And last, you know, put on that, networking. Think about a NIC card that's capable of handling demands of AI modeling.
What you do with Ethernet today isn't good enough. So working closely with a number of Ethernet standards partners to figure out ways, how do we make Ethernet more capable to develop the idea of systems of chips and Ethernet provides that low cost, high bandwidth potential to go off and do that. Add to that photonics, add to that technologies beyond that. Here again, it's another five year roadmap. And then lastly. We have 18,000 software engineers at Intel.
Why don't we harness those software engineers and help people figure out how to make these systems boot up, how to develop the firmware and the software necessary to optimize and get those gear ratios absolutely correct. So for us, this is what a true systems foundry means to us. And it's why we believe we have a differentiated approach in the marketplace. The only way we're going to get to this exponential
scale is by taking all these elements and multiplying them together. And that's why we think we have a new way of looking at the foundry business for all of you. To do that. We can't do it alone. We have to have partners. We're doing everything we can to give our customers choice, to have a full stack of availability from a number of EDA partners, to do systems technology co-optimization, to have open standards and reference designs available to speed the innovation that these fundamental technologies provide. And we're not going to just do it with the folks of today.
We're also going after partners to fuel the next wave of innovations. Intel Capital makes equity investments in startups. They make strategic investments, strategic investments in scale partnerships. They make ecosystem investments for foundry at scale. And there are a number of examples that Intel Capital has successfully invested in that people are bringing unique technology, Ayar Labs, a photonics startup doing some incredible work in photonics. That's how we're going to capture customers large and small.
But we have to start back even further than that. We have to go to the universities and talk to them about what do the students need to learn. As Gina Raimondo said, how do we create an engineering population capable of taking what we're doing in these factories and making it available broad base? So I'm proud to announce today that we have our first 18A partners, the University of Michigan, my home school, and Berkeley, Tsu-Jae's home school, no connection whatsoever.
Um, 500 academics, 60 research group introduced test chips. What we're doing with U of M and with Berkeley is we're finding a way to bring 18A and make it available to students. Today's technology, not yesterday's or three years ago, but today's technology. And we're developing innovative ways to provide little tiny shuttle seats that students can run their test chips through our factories. And it's going to be fascinating to watch this develop as we go out and embrace the university community with this new technology. So.
We have one more thing to talk about. Oops. I'm sorry. Go back. One slide. We're not only. Doing this with partners, and universities, with Intel Capital. This morning we're announcing a new partnership with ARM, Emerging Business Initiative.
How do we take advantage of all the programs that ARM has to offer to bring design capability, design education out to all of their customers? We're doing this with ARM. We will make Co-investments, we'll do joint programs, we'll provide shuttles at scale, ARM will provide IP at scale. And this is how we're going to fuel this next wave of innovation. And it's truly exciting.
Now I mentioned we have our 30 ecosystem partners out there in the hallway, but there are five of them coming on stage today that I'd like to talk to you about. ARM. There we go. ARM. Ansys. Cadence. Siemens. Synopsis.
The big five. The folks that you turn to when you're doing your designs. So after the break, you'll hear from the EDA CEOs and a research report that I read yesterday, one of the commentators remarked.
Is the ecosystem really going to show up? They're really going to be here. What are they going to say? What you're going to hear from these four companies is hard data, not marketing slides. Although there are a couple of slides on AI. You'd expect that. Hard engineering data. What does 18A performance look like? What does it take to design with it? What are they seeing from us? Why do they like what they're seeing from us, and why do they view us as a great partner in this journey? But for the next few moments, I want to talk about our most important partner.
And that partner is. Believe it or not. Wait for it. Rene Haas from ARM. So what I'd like to do now is invite Rene to come up on stage with me and talk to you about what we're doing with them. Hello, sir.
So good to have you here. So, Rene, you know, first off. You know, it's really unusual for an Intel business unit leader to make this statement, ARM is my most important business partner. And what kind of universe would you have ever thought that you'd see ARM and Intel standing together? But you know, this is a brave new world for us.
And when we started talking with Rene and his team about a partnership. You know, we rapidly came to the realization that 80% of the wafers TSMC runs, has an ARM device in them. There is no way you can be in the foundry business without a partnership with ARM. And so we kicked out discussions months ago. They have been absolutely fascinating . And I think you'd like to probably tell us about a few of them, starting with perhaps the announcements you made this morning. Well, thank you for having me.
Um, as you said, this is a bit of strange bedfellows. I was trying to think of a parallel that I might give relative to this story. And the only thing I could kind of think of is for those who can hearken back, is when Walt Mossberg asked Steve Jobs what it was like to see iTunes run on windows, and I think he said it was like an ice water in hell, but I won't go that far. I won't go that far. No, it has been, it's been fantastic. We started these conversations, you know, not long after I took over as CEO at ARM, because we felt, you know, fundamentally that the technology that Intel was bringing to bear, as you've just shown earlier, between you and Pat, is industry leading, industry changing.
And we need to be a part of it. So thank you so much for having us. Great to have you here. Yeah. So you know, the announcements that we made earlier today was around our Neoverse product line, which is the product line that we use for the data center.
Which has just been exploding. It was in a very, very high growth trajectory prior to the AI wave, and now it's become even stronger. The Neoverse V3 that we announced today, which is 50% faster than the Neoverse N2, and then our N3 Neoverse, which is 20% faster, but also much more efficient.
And when you think about these AI data centers, which are pulling hundreds of megawatts
2024-02-26 13:17
