The Six Five Pod | EP 255: From Intel to Innovation: Pat Gelsinger’s New Ventures

From OpenAI’s $300 Billion Valuation to shake-ups at Intel and more, hosts Patrick Moorhead and Daniel Newman dive into the week’s biggest tech stories with a very special guest: Pat Gelsinger. Get ready for insights and unfiltered perspectives you won’t hear anywhere else. The handpicked topics for this week are:

1. AI Model Mania: From the latest from OpenAI, Microsoft, and Google, we break down the key AI announcements and what they mean for the future of the industry.
2. The OpenAI Valuation Debate: Is OpenAI’s $300 billion valuation justified? We go head-to-head on this hot topic, exploring the factors driving investor enthusiasm and potential risks.
3. Pat Gelsinger’s Next Chapter: We sit down with Pat Gelsinger to discuss his new roles as Executive Chairman of Gloo and General Partner at Playground Global, delving into his passion for deep tech innovation and the future of faith-based technology.
4. Semiconductor Sovereignty: Pat Gelsinger shares his views on the critical importance of US semiconductor manufacturing and the ongoing efforts to strengthen domestic supply chains.
5. Powering the AI Revolution: The increasing power demands of AI, and the crucial need for more efficient computing architectures to sustain future growth.

For a deeper dive into each topic, please click on the links above. Be sure to subscribe to Six Five Media so you never miss an episode of The Six Five Pod.

Daniel Newman: For everyone out there, though, by the way, you heard the Big Pat, little Pat thing. There is something to that. It’s that Pat Gelsinger is just a heck of a lot stronger than Pat Moorhead. We’re going to talk about this, by the way, the Pat segment Pat. Because you and I need to debate this because I think you know I’m kidding.

Patrick Moorhead: I mean, that will be next week’s flip. But just listen, for the record, I did see Big Pat in the gym.

Patrick Moorhead: Daniel, my bestie. It’s great to see you. Welcome to The Six Five Pod. It is episode 255 and we have not been cancelled yet. Don’t intend to, but just glad to be on air.Daniel it is a rainy day. You know, I was in Seattle for three days and it’s like I just brought that weather with me. It’s raining outside. It’s foggy. What’s going on here?

Daniel Newman: I knew it was you. It was beautiful until you got back. No, actually yesterday the weather was pretty crappy here as well. Are we really starting off talking about the weather?

Patrick Moorhead: Yeah, I know, I know. What a snare.

Daniel Newman: I know you’re old and grumpy and now you have to do the weather thing, but in all seriousness. No, it’s great to be back. It’s been a busy week. I think you and I are suffering. Have you complained lately on Twitter about how many weeks on the road you’re about to, you know, have to participate in? I think it’s like 12, 13.

Patrick Moorhead: I got that whining out about two weeks ago where I basically laid out my four weeks of straight, straight travel. Oh, woe is me, right?

Daniel Newman: Well, I’m still tan. I mean, you know, I went on vacation. I mean, you went to GTC and some people thought that would be a good way to spend time. I thought it was better to spend time in Turks and pretend to be at GTC, which I think I did pretty effectively.

Patrick Moorhead: No, I like that. I’m going to consider that for the next GTC next year. But listen, we’ve got a great show. This is the new format of The Six Five Pod. You know, we’re just getting all the jets going here. We have The Decode, right where we decode. We basically talk about some of the big news announcements that came out, what was important, trying to sort through the signal and the noise. And we have The Flip where we flip a coin and we debate. Kind of like the old CNN Crossfire on, you know, we just pick a side. Not necessarily a side that we believe in, but it’s more a thought exercise of, playing both sides of the coin. And then at the end, we’re kind of flexible on the segments. Right? We have basically, you know, we did a market segment last week, but this week we are doing what’s called Off the Record with a special guest. I’m not gonna leak this out just yet, but somebody, I’m pretty sure that all of you know who follow the pod here. So with that, Daniel, let’s jump in to decode. Dan, let’s jump into the decode. We’ve got some great topics here. I mean, we’ve got AI models and feature wars. I mean, I’m pretty sure everybody made an announcement regarding LLM services, agents. There’s a rumor Nvidia’s acquiring somebody that is supposedly a GPU Neo cloud. And then we might talk about, you know, what the three Intel board members stepping out might mean or not mean. And if we have time, we have a couple other topics. So, Dan, let’s jump in here. I mean, OpenAI, Microsoft, Google with Gemini, Perplexity. It’s just a cornucopia of AI’s.

Daniel Newman: There was only really one this week. It was the Ghibli. I mean, everything happened, but nothing happened. I mean, the whole world turned to Ghibli. It was kind of funny. There was a bit of a back and forth even. I think Sam Altman was sort of cooking on one of the Google engineers about, you know, that basically their product didn’t get any attention today because OpenAI just ran them over. But having said that, Google got some really great reviews in terms of its new model and its actual performance. But do we give a crap about performance anymore? It feels to me, Pat, like we just want to see all of our memes in anime. And I don’t know, I was talking on Twitter. I was just tweeting back and forth with Kelly Evans. You’ve been on her show. I’ve been on her show. We’re both kind of like, I don’t get it. Did I need to watch more anime in order to give a darn about this? Because, like, yeah, it’s pretty cool. I mean, I’m sure, Pat, that our producer should probably drop right now an anime of you and I doing something that would be really fun because our audience would appreciate that. But mostly it’s like the memes. You know, you look it up. You remember that meme of like the guy walking down the street with the girl but he’s looking over his shoulder back. All these memes that have been so popular now, have been animated. So Pat, you know, we know this LLM space. It’s a leapfrog thing. It’s jump over, jump over, jump over, jump over. The 4, 5, 6 players are all going to keep doing this. It is news. But really in the end it’s who’s paying for what and who’s going to continue to pay for what and is this market there? And you know, maybe even in The Flip we’ll talk a little bit more about how absurd and obscene some of these valuations and things are becoming. If really it’s nothing more than anime.

Patrick Moorhead: Yeah, it’s interesting. Two things that stood out for me first and foremost was the Gemini went from almost worst to first with 2.5 pro. You know, I’ve not been thrilled with Gemini since inception. I keep going back and kicking the tires but you know, pretty much Perplexity, Groq, and OpenAI but also based on what I do, which is a lot of research. In fact OpenAI Deep Research is absolutely phenomenal for stuff like healthcare. It’s crazy. I almost self diagnosed my dad and some of the issues that he’s been having and one announcement I did want to call out is that Microsoft is leveraging OpenAI’s Deep Research and O3 mini reasoning model for two offerings. And Dan, I don’t know if this is going to put you know, everybody in the research community out of business. It’s not. But there’s a researcher that’s leveraging that deep research capability that I talked about but doing it for businesses. Right. And then with different APIs for data you can actually pull in additional data that you’re looking for and these are called connectors. And then there’s an analyst, right. It’s billed as letting you think like a skilled data science going from raw data insights in minutes. That’s the architecture, that’s the marketing. But it is great to see Microsoft like it has leveraging OpenAI and what I don’t yet consider enterprise grade and Microsoft applying its enterprise grade construct and ecosystem to bring this to big businesses. So those are really the two standouts for me. It is interesting. You know, Perplexity is going literally straight after Google. Okay. With, you know, you can go shopping, you can do travel, you can do all the things here with a higher level of specificity. And while you have to click and self select on what you want, Perplexity says it’s going to get better just based on sitting at the search bar, on knowing with which data path and model path take you down there. Perplexity already has that for world knowledge, right? You pick auto and it, it has a decent idea and then it kicks you off, you know, into a reasoning model or it’s pro search or just a basic search that you’re looking for. So interesting stuff there.

Daniel Newman: Yeah, it’s funny, Pat, I know we got to keep moving on, but Apple should just buy Perplexity and make it their search engine. I’m just saying, like this is the most obvious.

Patrick Moorhead: I know

Daniel Newman: Thing ever, because Perplexity will not, I mean, look, I’m calling it out now. It will not be a big success on its own, but it could really bring Apple to where it needs to be. And if Apple backs it, it will succeed. And I mean the return on investment there could be massive. It seems so obvious to me, but I’m imagining Tim Cook’s just going to put like three more people in charge of Siri so it can continue to suck for eternity.

Patrick Moorhead: Ouch. Ouch. Yeah, we’re gonna, we’re gonna leave this topic at that. Yeah, I made the Perplexity call a while back, so I’m in complete plea agreement. But hey, let’s go into this next topic. There was a TechCrunch article that I think was driven by an information article that Nvidia is going to be buying in terms of what it characterized as a GPU provider. But first of all, let me clarify and do a fact check. Lepton doesn’t, Lepton AI doesn’t actually have GPUs. They actually are an intermediary between companies that have GPUs. Okay. And what they are is a bunch of smart people who created things like Pytorch, Onyx, Cafe, okay. Literally on their websites is from the creators of Pytorch, right? So a bunch of Meta folks. And what they do is there is a tool layer to not only give you better efficiency of your GPU spend, but they also have these monitors to make sure that you know the health of your GPU. And as we know we’ve discussed it on the pod before. The biggest challenge with training runs and even inference runs are GPUs that quite frankly are either burning up they’re sitting __.

Daniel Newman: What did Sam say? My GPUs are melting because of the Ghibli anime?

Patrick Moorhead: No, no, no.

Daniel Newman: He tweeted that 3 times.

Patrick Moorhead: That’s exactly right. So what you have is. It’s really a tool set. So you know, it’s kind of funny. Everybody said this was an acqui-hire. Oh boogeyman. Gosh. They’re going to be competing with AWS. Listen, their investment just in CoreWeave is the competition with their customers. Lepton AI is not a vertical integration play. It is not necessarily an acqui-hire. It’s the tools part. I could see them integrating Lepton AI dashboard and their tools right into the enterprise AI software stack or integrating this as a bolt on onto Dynamo.

Daniel Newman: I know this is somewhat, this isn’t a flip, but I would actually say it is somewhat vertically integrated, but it’s not quite what people are making it out to be. I think this would be more useful to, to basically diffuse this to its reseller community and let the resellers have a service to make money because they keep taking all the opportunity out. Like what is the, the whole CoreWeave IPO getting crushed thing is: it’s a company that’s 100% dependent on Nvidia that’s now got really no product service differentiation except the fact that they have less constraints of getting more GPUs. So these are things that these firms and nebuses and the CoreWeave’s and these other firms could put into play to add more value to their customer. If Nvidia takes it out, it’s just another layer where they can’t make money. Having said that, just don’t think Nvidia trusts its ecosystem right now to get it done. I mean I think that’s a lot of what we’re seeing right now is they’re watching everything and they’re saying if we let everyone else control this, we will get caught. So we’re just going to control it all and we’ll take our chances on pissing some people off. I don’t think Nvidia will get hammered the way Z systems hammered AMD in terms of how it hurts, even though it was really supposed to help. But having said that, historically speaking buying your customers has never been a great strategy for companies in this space, I don’t love it. Like I said, I can’t pretend to fully understand how he plans to utilize it. Like you said, good idea with Dynamo, I’m just, I’m skeptical. And by the way, maybe Jensen’s strategy right now is while the stock’s sucking and the whole market’s sucking, do everything to piss people off all at one time. Piss off your biggest customers. Buy your customers. You know, do all this stuff, get it out of the way and grow. And you know what? Sometimes you have to be bold. So I’ll give them some, some chop props. Maybe that’s what’s happening here.

Patrick Moorhead: Yeah, I know it is war. And I think that would be a great conversation for potentially the next.

Daniel Newman: Jensen should come on our show. I think Jensen should come on the show.

Patrick Moorhead: I think he should come on and address it in Off The Record.

Daniel Newman: I’ll wear a black leather coat.

Patrick Moorhead: No, I love it.

Daniel Newman: In honor of him. Won’t look as good, but I’ll do it.

Patrick Moorhead: Exactly. So hey, let’s dive into the next topic. Three Intel board members are going to step down. Dan, like does this, what do you think? Does this, is this a meltdown? Is this the first of many changes we’re going to see at Intel? What does it represent and what doesn’t it represent?

Daniel Newman: Look, this is always bound to happen. You brought in a new CEO in Lip-Bu. We know he probably came with a series of conditions. He had been on the board, so he was deeply familiar with the construction of this board. We know that a lot of the reason the company suffered to get through some of its transition was it had a lot of academics and philosophers and non operators on the board. And I think Lip-Bu is just steadfast. He does not want to surround himself with people kind of living in academia. You know, there might be a role for that, maybe one seat on that board, maybe even up to two. But like right now, this board and some of the people that added, you know, Steve Sanghi, it added the former CEO of ASML, if I remember. This is the kind of people they need to be guiding them into the future right now. There may be a time at some point where they’ve recaptured the market and they have a dominant share and they want to get back to R&D philosophy, academia. But right now it’s about bringing back practical advice. If you’re Lip-Bu I don’t want anyone on the board that hasn’t done it. I want board members that have actually built companies, been through transformations, hopefully even been through something like this Pat. I mean look, you and I, you know, we’re building companies. Who do you take your advice from? I don’t go back to my college textbooks and look for how to do the next thing. I mean, if I’m actually going to stop and get advice, I want people to have done something that looks like the problem I’m trying to solve so they’re doing the right thing. He’s making the changes. And by the way, it’s good to see him making hard changes up and down the org. Because if you’re going to ask the people below you to operate differently, you better ask the people above you to be prepared for what that might mean.

Patrick Moorhead: Yeah, I think that’s a good take. So Intel has 14 board of directors and in fact so many it’s two pages, you have to actually click the pages. The typical tech company has seven to 12 directors. Most public companies have nine to 10. So first of all they had a lot of directors. Now if I look at the three board members who are stepping down, you had one Omar Ishrak from Medical Design Co. And there was a time, if you remember during Brian Kurzanich’s reign, that the future was all going to be about IoT. And about small chips. This made sense. And then you have Tsu-Jae King Liu, I hope I said this correctly. She was dean of the College of Engineering at Berkeley. So you have an academic, a smart academic there. And then you had Risa Lavizzo-Mourey, professor specializing in population health and health equity at University of Pennsylvania. All three of these were brought in under the Brian Kurzanich BK regime. And yeah, first of all getting people with more actual experience. Eric Maurice and Steve Sanghi first were former president, CEO and Chairman of ASML. Sounds like a really important thing to have. And then Steve Sanghi, this interim CEO at Microchip. So this makes a lot of sense to me. So focusing on what their core competencies are getting away, less academia and more into I would call it real world delivery just like you said. Okay, let’s jump into The Flip now. OpenAI valuation is the, the topic is it worth $300 billion which is a derived value based on a softbank investment that it made. We’ve debated this Daniel. So let’s jump in. Is OpenAI worth $300 billion all right, Dan, you’ve got to defend that $300 billion valuation.

Daniel Newman: Anyone that’s seen my Twitter, my X, would know where I stand on this. This is absolutely the pinnacle of AI and LLMs. This company is so far ahead, has so much more technology, has so much more market share. It has redefined the business. And every company in the world is trying to keep up with them, whether it’s Google, whether it’s Microsoft. Microsoft is so scared, in fact, of what OpenAI is doing that they’re running away, trying to build their own competitive products because they just can’t keep up. Sam Altman is an ultimate leader. He’s a machine. He’s driven the academics away from his business because he knows how to turn this thing into a company that’s worth a fortune. That Koenigsegg that he drives down the street, that’s not the only one he owns. He owns many because this guy knows what he is doing. You look at the revenue, you’re talking $12 billion in revenue over the next couple of years. You’re going to see this company turn for profitability. Of course they’re losing money. You can’t make money this early on. When you’re making this much investment, you need this many GPUs because you’re doing stuff that nobody else has done. The future requires risk. And if you’re an investor, you want to be on board with the most valuable companies in the world. And look, every time this company goes to raise money, there are investors. And these are not fools, these are not clowns. These are smart people. These are PhDs. These are the most successful investors in the world. They’re putting their money up and they’re back in the company. When it was $10 billion, Satya put up his money. The last round was 160 billion. And there were people piling in to invest in this thing. You could not get after this thing. I don’t care if their GPUs are melting. I don’t care if they’re nowhere near AGI. It’s worth it because it’s what people are willing to pay. And now you hear SoftBank has got $40 billion pulled aside with more people ready to fund this thing so they can buy more GPUs. They could build $7 trillion worth of custom chip industry development. They’ll do what they want because OpenAI is the one. They are the only. Sorry, everybody else, you’re just a follower at best.

Patrick Moorhead: Yeah. So I think the $300 billion valuation for OpenAI is laughable. Okay, so first off, like you know when a ship’s going down, right? The rats jump off and then the fleas jump off the rats. Microsoft basically is dumping OpenAI, pulling back on all their capacity. They totally do not believe that A, OpenAI will have a payoff or B, that their technology isn’t impenetrable. Okay. You know, OpenAI used to, it used to look like, oh, they had a year advantage and now it’s like a couple months. You know, I prefer, you know, deep research. And it’s the best, but quite, you know, it’s 200 bucks a month and it didn’t get lapped. But DeepSeek gave it an absolute run for its money at reportedly a lower test time, inference cost. The company’s not supposed to be profitable till 2029. Okay? Like literally cash flow. No, cash flow positive. Okay, so four more years of dumping cash into this thing and, and whose forecasts are ever on? Maybe they won’t be cash flow positive till 2031, maybe 2032. You know, who knows? Already talked about the competition, but you’ve got. Models are pretty much a commodity at this point. We saw it with DeepSeek, right? We’re seeing it with Quinn, we’re seeing with Llama. Okay. I just think that it’s kind of laughable. And even, by the way, just even looking at the multiples, right, typical SaaS, multiple versus revenue is 8x and an AI, typical AI startup is 15 to 20x. And what you’re looking at here is OpenAI looking at a 24x multiple. So it’s, it’s absolutely at the high end. And if you’re at the high end, you’ve got to have a moat. And I don’t think that OpenAI has a moat. So all this valuation, right, depends on them generating a hundred billion dollars of revenue by 2030. 70% gross margins. By the way, how do you, how on earth are you going to hit 70% gross margins? Because at some, like CapEx doesn’t just sit there, right? It becomes a cost that goes into your PNL. And okay, we can argue that’s, that’s a, that’s depreciation, right? It sits down there, but quite frankly, none of these numbers add up for me Daniel. Any counter on that?

Daniel Newman: Well you give me the chance to counter. We ever done that before?

Patrick Moorhead: We haven’t.

Daniel Newman: All right, well, the quick counter is basically this. It really doesn’t matter whether you think it is or it isn’t a bubble, as long as people with money continue to pile in as long as the chairs keep settling and that the music keeps playing, the value keeps going up. I can make the same arguments you made for every single thing for Nvidia, for Microsoft, for Amazon, that all their valuations are astronomically too high. But having said that, in the end, Pat, it will be what it will be. And in the end, I will struggle forever to try to support OpenAI and ever make this argument again.

Patrick Moorhead: Yeah, closing statement. I mean, you talked about shares, I mean, shares in the Titanic, so, you know, let’s go.

Daniel Newman: Absolutely. This is the fun part of The Flip, is trying to somehow defend something that you find indefensible.

Patrick Moorhead: So I know, I think you and I both absolutely believe that this valuation is irrational. I looked at Amazon’s valuation, it was irrational too. I looked at Tesla’s valuation previously, and it was completely irrational before it really started generating meaningful revenue. But that was a good topic, Daniel. So, hey, let’s go into our final topic here, and that is the Off the Record. That was a great Flip section, Daniel. I’m pretty sure that, as usual, I won that debate, but we will let the viewers.

Daniel Newman: Have you won any of the debates? I mean, is this just revisionist history here?

Patrick Moorhead: I’m literally two for two. Okay.

Daniel Newman: If you say that, enough people might believe you.

Patrick Moorhead: Okay. All right.

Patrick Moorhead: I mean that’s pretty much true with everything. But hey, let’s get on to the next segment. This is Off the Record, right? We’re bringing in technology’s top voices. We’re going to get real. It’s unfiltered, unscripted, and we’re pretty sure you won’t hear this anywhere else. And it is my privilege to introduce Pat Gelsinger. Big Pat. Welcome to the show.

Pat Gelsinger: Hey, little Pat. And Dan. It’s always a pleasure to be chatting with you guys. So much fun, right? And boy, what, what a, what a history we’ve had together. So just, just a pleasure to have this opportunity to be on your show.

Patrick Moorhead: Yeah, you know, some guests, it’s like, you know, first time on the Six Five, but we’ve had the privilege to have you on the show a bunch of times, so I appreciate that.

Daniel Newman: Yeah, Pat’s done a lot of different things with us. I mean, we’ve sat down with him at his events. He headlined one of our Six Five Summits. I feel like I should drop a little promo here, but I won’t. I’ll let the team handle that later. And Pat, you know we’ve been outspokenly, in many ways, supporters of the work you did at Intel and of course the work you are going to do in the future. For everyone out there, though, by the way, you heard the Big Pat, little Pat thing. There is something to that. It’s that Pat Gelsinger is just a heck of a lot stronger than Pat Moorhead. We’re going to talk about this, by the way, the Pat segment package. You and I need to debate this because I think I’m kidding everybody.

Patrick Moorhead: I mean, that will be next week’s flip. But just listen, for the record, I did see Big Pat in the gym. I forget exactly which event that was and I’m really glad he didn’t. Pat, you didn’t ask me to do the push up challenge, although my shoulder was doing okay. I may have given you a run for your money, but. But not anymore.

Pat Gelsinger: Another day we’ll do that.

Patrick Moorhead: Yeah, let’s dive in, Dan.

Daniel Newman: We’ll let him get after it. So, Pat, you had a big week. I saw a few of your LinkedIn posts. I saw some press releases hit. Give us the just kind of what is Pat Gelsinger up to right now?

Pat Gelsinger: Yeah. And maybe to just, you know, back up a little bit. Obviously, you know, I’ve lived my career, but I’ll say on two pillars, you know, our faith, philanthropy and family and deep tech. Right. And what we can do with technology. So, you know, everything that I’ve done throughout my life is sort of centered on one of those two. So this week’s announcements were with Gloo. Right. A technology platform for the faith ecosystem. Really sort of combining those two together. And I became executive chairman of Gloo. I’ve been on the board there for a number of years, but now stepping in and we really see that the faith and church ecosystem, very fragmented, very large. 350,000 churches in America alone are just stunning. $1.2 trillion of economy, you know, flowing through that ecosystem, but it doesn’t have good technology. And what Gloo is doing is bringing the best technology. We now have over 100,000 churches on the platform and are now bringing AI services into it. So really exciting to bring those two together, you know, so that was the first announcement, but the second announcement was the deepest of deep tech. And I’m joining Playground Global as a general partner. And if you know anything about Playground. Right. You know, a lot of venture capital firms, you go into the front door, you see the tombstones, right. Of all their IPOs and what you’re telling founders is hey, we’re going to make you billionaires like them. When you walk in the lobby of Playground where you’re greeted by a robot, you see the world’s first hydrogen airplane, you see the largest 3D printer, the world’s most powerful free electron laser. This is an engineer’s playground, you know, and hey, I’m an engineer, right. And I love engineering and the idea that we can help these, you know, ill formed early startup, you know, young CEOs crazy ideas, you know, and form them into proper companies, you know, that truly are game changing and quantum computing and next generation, like, you know, next generation bioengineering. That’s pretty exciting for me. So bringing together those two, it’s been a big week for me and you know, that’s where I’ll be spending my time on those two big efforts.

Patrick Moorhead: Yeah. So Pat, I appreciated the pre briefing on Gloo. You know, I think we discussed, I think you know, the first app that I open when I get up in the morning is the Bible app. You know, I need that daily, daily devotion and reset and I’m just imagining what bringing AI LLMs into this across multiple faiths. Right?It’s not limited to Christianity but no, it’s, it’s really interesting to see and I’m excited to see where you take this.

Pat Gelsinger: Yeah. And we’re, it’s, you know, truly a thrilling time in that sense. And you know, the Bible app, you know they are going to hit a billion devices with that app this year. One of the most deployed applications in the App Store for Google and for Apple. I mean it’s really quite stunning the scale and obviously they’re a partner of Gloo. How do we help them? When you think about things like Bible translation, you know, we’ve had people, you know, pining away at word for word, you know, translation. Well, hey, AI, what a powerful tool for, you know, translation, for crossing language barriers, you know, for being able to take language and put it into spoken word. Many of these languages, you know, have high illiteracy rates yet, you know, so it’s really powerful. You know, imagine the, you know, the benefits in health and recovery services, etc. So all of those things, that’s the market for Gloo and we’re quite excited.

Patrick Moorhead: Yeah, And on the playground stuff, it’s so funny, you know, for about 10 years. Only a few VCs would even get close to, you know, some people call it deep tech, other people call it hard tech because the time to money was so long and the risk was so high. And there’s big capital investments and outlays that go in there. Are you focusing on playground Global, an IP play or like or actually, you know, we’re going to see real pieces of silicon funded. You talked Quantum. Is that part of the portfolio?

Pat Gelsinger: Yeah, clearly there’s, you know, a range of things in the portfolio, but you know, we’re the deep tech, long tech, hard tech, you know, center, you know, that’s who we are. And you’re solving these really hard problems. And so a lot of silicon work with, you know, foundries, but we don’t just work with them as, hey, let’s, let’s take your standard recipe. We’re building new recipes, new material structures. You know, the side quantum team, hey, you know, they’re building an entire new machine to do a different type of material epitaxy. Epitaxial growth. Right. To deliver next generation optic because they’re doing single photon quantum devices at scan. You know, they’re building an entirely new cryogenic refrigerator. Right. So instead of that, you know, a chandelier that gives you a little area of cryogenic temperature performance. Yeah. We’re building an entire vault. Right. That we can take cryogenic temperatures because, you know, building a couple of cubits. Okay, boring, right. You know, how do you build a million cubits? Okay, that’s a different engineering problem, you know, or the X light guys, you know, building a free electron laser that’s 4x the delivered power at the target wafer and does so with one tenth the power. Right. Of today’s LPP. You know, these are breakthrough technologies and if we’re able to bring those to light, they will change entire industries. You know, not just produce good financial results, which they will, but it’s really about changing the industry for the future.

Daniel Newman: Yes, it’s really interesting. What about a million qubits inside of a box that doesn’t require a refrigerator? Now there, there, we’d be having some fun. It really does sound like a playground for nerds, for geeks.

Pat Gelsinger: Yeah, it is. And I’m a nerd. Nerd. So right. If I was starting grad school again, you know what I would want my dorm room to be is a cot in a maker lamp. Right. Where you know, you’re just the printers and you know, all other kinds of stuff. Well, to some degree, that’s what Playground is. You know, the world’s largest 3D printer right there, right in one of our layouts. And, you know, being able to, you know, have these brilliant scientists, right? You know, and helping them to create mature, proper companies around these technologies. That’s what we do at Playground.

Patrick Moorhead: So, hey, in the spirit of, of getting real unfiltered, unscripted, I mean, we’d be delinquent if we didn’t ask you about Intel. Okay. I mean, I competed with you when I was at AMD and you were at Intel. And then, you know, you went to VMware, went back to Intel and made a bunch of changes. Just tell us about, I mean, any. What can you say about Intel and the future of Intel?

Pat Gelsinger: Well, you know, two things that I was working super hard on when I was CEO there. One was, hey, getting the chips out, getting industrial policy to rebuild this industry backed on American shores and creating resilient, sustainable supply chains. And the other was bringing Intel back to be that national champion because, you know, we need semiconductors built in the US. So manufacturing, but not just manufacturing, must also be the technology development, the R &D, the intellectual property of the transistor, the materials, etc. And those two priorities remain priorities today. Right. You know, our supply chains must be more resilient. We need to have this technology back in the U.S. we, right. You know, we have to finish what we started for it. So, you know, as we look at Intel today, you know, still working through it. 18A best process technology in the industry has to come to scale. 14A hey, you know, it is a killer process. Technology has to come to scale and we need to build the foundry, the manufacturing. And because of those priorities, hey, my absolute best wishes for the intel team, you know, for Lip-Bu, the new CEO there to finish what we started. But I’d also say things like XLITE and some of the other projects we have at Playground, hey, I want to be helping to fulfill that vision as well. And I may not be at the reins of Intel, but hey, I am still deeply committed to the success of Intel, but more importantly, the semiconductor industry. Because this is the future. Every piece of our life runs in semiconductors. And we have to have resilient, sustainable supply chains today, tomorrow, and well into the future. And now I’ll be working on some of those things a little bit further out in the future.

Daniel Newman: It’s got to be a little bit like, you know, being a grandparent, watching your grandchildren now. As opposed to being the parent, you know, in some ways because like a lot of the things you’re seeing come to fruition or it’s not like all the things that were being done during your tenure just stopped. A lot of these things, like 18A, you mentioned that. I mean really, you’re seeing it manifest, you’re seeing it proliferate and it’s looking really encouraging. So you’ve got to be really proud of that and that work. Obviously a lot to be done. Yeah.

Pat Gelsinger: And you know, we came in and I think the industry recognized that, you know, Intel had fallen years behind. Just getting back to being competitive leadership technology. And it cleared. Wow, that’s a lot. You know, getting EUV deployed on that scale. Wow, that was important. And getting the development engine running, there again super important advanced packaging and what we saw, you know, with the advanced packaging centers in the US, the establishment of secure enclaves, you know, all those things, there is no disagreement on the priority that those need to get finished. Like you say, I’m proud and really encouraged by what we were able to get done, you know, in my time there. And the absolute best wishes for those to continue.

Daniel Newman: Yeah, so just a quick double click. You kind of started with Pat’s question talking about, you know, semiconductor, US exceptionalism, the need to bring back. You’ve made some comments publicly. I’ve heard, read and seen. This is a real problem right now. You know, we kind of hear there’s this tariff war, there’s this inflation war, there’s this cost optimization thing going on. But like look across all industries, it’s probably not a more important one than chip making in terms of the long term health and wellness of our economy and of the US’s exceptionalism. Talk a little bit, if you don’t mind sharing why you’ve continued really on that, on that talk track even, you know, in your current capacity.

Pat Gelsinger: And you know, the thesis that was sitting behind the CHIPS act, which began, you know, in Trump one. Right. It was passed under Biden and you’ll see the priority not continue into Trump too. And clearly if you look at Lutnick’s confirmation hearing, you know, the priority of Congress to finish what we started. So, you know, those things haven’t changed. Every aspect of human existence today is becoming more dependent on digital and everything. Digital runs on semiconductors. Right. You know, we were acutely made aware of that through Covid and that hasn’t been fixed yet. We’re now starting to build the factories. We’re now starting to see, know, those things swing in our direction. But the chips act was, you know, essentially getting us to neutral. Right. And now. Right. We have to bring bigger economic things to play because supply chains don’t move at equal. Right. You have to have dislocation, you have to have reason to adjust it, you know, and you know, Wall street, short termism, hey, we can’t let that pervade, you know, we must build, right. These supply chains and build that resilience, you know, and for that, some of these other levers that you’ve heard the administration talk about in terms of your tariffs and how you create the economic incentives and you know, the potential for the sovereign wealth fund, I think all of these are extraordinarily important complements to what the chip stack got underway. And the other piece of the chip act wasn’t just manufacturing incentives, it was long term R and D, you know, NATcast, NAP, MP. Those must be executed. We need, you know, the only way to, you know, stay ahead is to stay ahead. Right. So we have to get back in the game and then be that innovation engine for tomorrow.

Patrick Moorhead: Hey Pat, final question here, you know, s in the past month, I mean, you see a lot of conversations out there about, you know, what needs to be done with AI semiconductors. I think we’re looking at 600 watt racks and I think the next generation could be a gigawatt.

Par Gelsinger: 600 killawatt.

Patrick Moorhead: Yeah, sorry.

Pat Gelsinger: And wow, 600 watt rack.

Patrick Moorhead: No, I know that would, that would be amazing. I think I have that on my desktop. Right, next.

Daniel Newman: Very efficient, Pat. Very.

Patrick Moorhead: Exactly. But there’s a lot of conversation, I’m curious, what, what are some of the overriding conversations? You know, you’re sitting there and you’re reading something and you’re like, ah, that’s not actually how I am seeing the world. Or I think it plays out.

Pat Gelsinger: Yeah, well. And you know, when we think about what we’ve seen with, you know, AI training and environments and people are, you know, I mean, we’ve sort of gone, you know, let’s, let’s say at the data center scale, right. You know, where we sort of quickly said, hey, we’re going to be, you know, a few megawatts per at a center to 100 megawatts per data center and a half a, you know, gigawatt per data, you know. Right. You know, okay, we can’t keep going on that path. You know, projections would say, you know, that 20% of US power is going to be consumed in AI data centers and growing, you know, at high single digit per annum growth rates. Okay, you can’t do that, right? You know, that doesn’t work. Right. You know, to continue to pursue. Right. AGI, you know, with such power consumption, you know, when I said at GTC that hey, we have to make inferencing 10,000 times more efficient, you know, not 10, not 100. 10,000 times more efficient if we’re truly going to deploy it like the Internet where everything, you know, takes advantage of AI technology and obviously the compute architectures to accomplish that. And one of the seed companies I invested in Fractile, one of the companies in the Playground portfolio, D Matrix. How do you bring compute and memory, because these are large memory workloads more efficiently together in the future. And I think steps like that architecturally are going to be super important because, you know, we just need more effective ways to execute the AI workloads of tomorrow. And as you know, some have suggested that reasoning, you know, will be a hundred times more computationally, you know, expensive than the LLMs of today.

So, you know, not only are we way out of bounds with how much power we’re consuming today, you know, but if you add 100x to that flow for additional compute requirements for reasoning, hey, this doesn’t work into the future. So we have to have more efficient architectural models and clearly, hey, Playground, we’re going to be all over that topic and they were already working on it and with me joining, okay, we’re going to get more focus on this. We need more efficient power delivery. We need next generation memory technologies. We’re going to have to move from copper to optical. You’re going at scale. You know, NVL72 is an engineering marvel and a manufacturing nightmare. Right. You know, we have to be able to create more scalable systems, you know, that are deployable, more power efficient, you know, take more efficient use of memory. So a lot of great engineering work. This is a delightful time to be an engineer and a technologist in the industry.

Daniel Newman: Well, Pat, I want to say thank you so much for spending some time with us. Exciting. I have the feeling that while Gloo and Playground will certainly keep you pretty busy, these will probably not be the only things that you’ll decide to take on your plate. I know that, Pat. Little Pat, you’re really big on sleep. Last question before I let you go, Pat. How many hours of sleep a night do you go for?

Pat Gelsinger: Well, you know, since I retired. Right. You know, my wife has definitely encouraged me to do that so you know, I’m now close to six hours a night, so, so I’m sleeping a bit more, feeling good and healthy, coming to the other side of this transition. But hey, I’m excited for the way that we’re going to be able to lean into some of these next generation technologies and technologists with playground and truly bringing my two great passions together, faith and technology, through the Gloo platform and really bringing about, I think the largest untapped segment of the economy today right for technology is the faith ecosystem and we’re aimed right at, right at making that happen.

Daniel Newman: With that six hours of sleep, you have 18 to work. You can do two hours a day for nine different roles. Pat. And you can keep each one of them as productive, as an average eight hour a day worker.

Pat Gelsinger: I am not going to share this with my wife that you have a new plan for me Dan.

Daniel Newman: I have a new plan for you. It’s going public, published online, because this is off the record. Pat, thanks so much for joining us.

Patrick Moorhead: Daniel, we made it. I can’t believe it. We actually did this in an hour. The number of topics, having Pat Gelsinger on the show was pretty awesome.

Daniel Newman: That was the highlight. Trying to do that flip was hard. I’ve tweeted, I’ve been really public, that I absolutely think OpenAI is a dumpster fire. Um, but part of the fun of that whole section is actually having to defend something that, you know, it’s like debate class, like on what side are you on? So anyways, I can almost debate anything but that one was really hard for me. But this is a lot of fun, past fun to do our first Off The Record, it was a lot of fun as always to go through all these topics that decode. Always good stuff.

Patrick Moorhead: Yeah.

Daniel Newman: But everyone out there we love all you, we hope you’re really enjoying the new format. Going to keep getting better and better and better. So hit that subscribe button if you haven’t already. Send this around to your friends. Pat’s the best analyst in the world and I’m just glad to be by his side. But for this show, for this episode, I got to say goodbye to everybody. We’re going to see you.

‍