 Hello, everyone, welcome to episode 40 of theCUBE Pod. I'm John Furrier with Dave Vellante, extracting the signal from the noise here on our podcast, episode 40. Dave, great to see you. Hey, John, I just saw you yesterday. Thanks for the ride to the airport. You were in California for the week. We had our big cyber resilience, cyber recovery event sponsored by Dell Technologies. Thanks to those guys, but we had a great event, basically a lot of scalable content when I was great, great event, great turnout, great traffic, overall great response from the community. That event was phenomenal. Great to see the team and your LinkedIn posts with the team dinner, everyone's kind of waving. Not a stage photo, got a lot of likes, Dave. A lot of people in the community love that content. So great podcast. We had fun in Mountain View. Michael's restaurant was amazing. Great, great event. I wish I was there, I missed it. Anyway, so here's what we're going to talk about in the pod today. Obviously, we had our big event last week. We'll go through that real quick. Intel had a big event in New York City. They simulcast that around the world. Intel AI everywhere, not so much, but I call it the not so big splash in the pool. Pat Gelsinger really made a big deal out of this. And we have analysis on that. I got some inside information from folks in Silicon Valley on that too. So yeah, it's interesting what you see in the press and what the analysts are saying and what really is happening when you hear what's going on in the back channel. We'll share some of that. Big funding news out there. Mistral AI got a ton of funding. French company, interesting there in Europe because there's a huge AI regulations in the EU and more guardrails are coming out. So seeing AI news, AI hype and substance coming out of AI, but the whole regulatory conversation is in top gear, Dave. It's really kind of a rant, but also a reality right now we're seeing in the marketplace. Hey, OpenAI put out a research paper. We're gonna talk about that. Some good commentary there, Dave. You made some great observations. Oracle had earnings and there's more layoffs in media in other areas, Dave. So I mean, you're seeing kind of where the results are coming in, companies are failing. A lot of areas where AI is gonna have impact like media. I mean, that's so wired laid off some people, Fox and a bunch of other companies. And then we're gonna talk about LinkedIn. I have a comment on LinkedIn as they owned by Microsoft, as everyone knows, are not moving to Azure. And that came out on CNBC. Jordan Novet wrote an article on that scooping the fact that LinkedIn's own Microsoft's own LinkedIn is not moving to their own cloud. Of course, Amazonians are like pimping it up like crazy. Oh, look at that, see, they're not even moved their own people over there. And then finally, I saw Michael Dell on Twitter today talking about. By the way, by the way, sorry to interrupt. I think Amazon still runs Oracle. It's funny. I'm pretty sure they don't talk about it anymore. We'll talk about that when we get to that section. And then finally, Dell PowerEdge. I've put it in Michael Dell on Twitter this morning. Dell PowerEdge cuts a big deal with Core Weaver and Nvidia. And why that's important is because we've been saying on the queue, and I put a video out there from our AI that actually proves that we were talking about this and actually happened that the opportunity for Dell and people with them in hardware is huge with AI. And I think the Dell PowerEdge server deal with Nvidia and Core Weaver point to the second super cloud model merging. So, little nuance point, but it's an indicator. You start to see the markers, Dave, of what we've been saying for the past year and going back six years around data and cloud and super cloud all lining up. The markers are there. Intel's AI announcement, the substance wasn't as strong but the direction was awesome, right? So, you're seeing everything going on. LinkedIn, not going to Azure. Is that because Azure is terrible? Is the fact that they try to serve all their compute for open AI? We don't know, we'll get to the bottom of it, but let's kick it off. Yeah, so our Super Studio event, seventh Super Studio event out of Palo Alto. I loved the topic. It was all about cyber resiliency. It wasn't a big product, you know, Pimpage. Actually, there was very little, there really wasn't any product talk. It was really all thought leadership. And we had the community in, obviously Dell, because they were the sponsor, it was very active, but we had CrowdStrike, we had Palo Alto Networks, we had Zias Caravalla, we had Christoph Bertrand from ESG, shared some data. We had partners like Continuity from Gil Hecht, came in from Israel, we had CISOs. It was, you know, it's a really important topic, John. You know, one of my big takeaways is people used to think DR was how they created business resiliency and they were realizing now that, wow, with cyber attacks and ransomware on the rise, the way they are, you know, DR strategies don't cut it. You have to actually build in a, you know, a combination of stop the breach and be able to recover from the breach. And it's a whole new line of thinking that is really being put forth by the community. I thought it was a really good, more than a half day session. Well, I think one of the things that you interviewed, John Skimone, Skimone, I think that's how you pronounce John Skimone. Shimone, I had it wrong too, Shimone. Shimone, I liked his interview because he talked about the Sony hack, which goes back days, years. Now, if you look at that time in history, and again, we commented on the cubes around 13 years as everyone knows, we talked about that, but that's, if you look at, go back to that Sony hack to now, it's unbelievable. It's still happening. The whole ransomware market, and so the whole cyber, that was 2014. Oh my God. This is like, go back to Stuxnet even before that, right? It's unbelievable the threats are coming. So I really think this conversation of data protection recovery is not so much about data protection but the recovery piece of it because you assume you're gonna be hit. The recovery becomes critical, not just so much protecting the data. And then the threat management is a whole other area. So I think what's going on right now in the security market that's exciting is, or actually, I shouldn't say exciting, but like important, relevant in the sense of it's exciting, it's top of mind is the threats coming in are a real issue, right? So I think that we're gonna see an explosion of threat countermeasures and emphasis on technology to manage threats, Dave. Well, threat management is just an underserved market and it's growing like a weed with the money on the table, trillions of dollars of cybersecurity fraud and threats. And threats. So Wendy Whitmore came on, she runs Palo Alto's unit 42, which is the threat intelligent unit. Delagent unit. So she has like, they're kind of like, you know, Kevin Mandiant, they have line of sight on all the bad actors, the nation states, the organized crime. And she said spear fishing is not really the most, you know, the biggest concern right now. It's this mass vulnerabilities. And so they're really changing the attack strategies. So that was kind of really eye-opening. And then I thought, you know, I screwed up John Shimoni's name, it's Shimoni. Anyway, so when you go to sort of the anatomy of these attacks, you go back to Stuxnet, it was like they got through air gaps to get to the Netans uranium enrichment facility. Amazing, you know, what they did there with programmable logic controller Siemens PLCs. And then the Sony hack that you just talked about. And that's just the advances that are coming. And then George Kurtz talks about just the time it takes to exfiltrate data is like down to, it's like 72 hours an hour, sometimes as low as seven hours. So speed becomes really, really important. It's just a really fascinating topic. And it is kind of exciting. It gets your heart pumping, to be honest with me. Well, I mean, there's so much at stake. I mean, the money on the table, everyone should go to the cube.net and check out the navigating the road to cyber resiliency, episode three. It's part of our annual program we've been doing. So that was our showcase capstone summit for the year. And it had the entire community in there, like you said, a guild from continuity was there, ZK researchers, ZS was there. Of course, a lot of Dell content, Palo Alto networks. I thought Rob Stretcher and ZS had a good point, fight fire with fire. And I think that's a threat we're coming with Forrester was on there. Mark Sorenson, an author, Kindrell Atos, Davis Trump from SiliconANGLE, our security head writer. And then of course, the town of Gilbert, Dr. Tony Bryson, cube alumni comes back again. He laid down a masterclass Dave. So, you know, this was a really good event. You're gonna see a lot more of this from the cube. Obviously, we got our super studio, we call it. A lot more scale coming out of the cube and you'll see a lot more. And again, the CEOs in our audience love it. The developers love it and the industry participants. And of course, the practitioners love it, Dave. One more stat I wanna share that came out of that, that came from Cyber Crime Magazine. I like to give those guys credit because when I don't give them credit, I forget the guy's name, he always emails us and says, hey, you gotta give us credit, that's our data. And they actually do some good data. So the cybersecurity, the spending on cyber software and services, it's like, I don't know, let's call it a hundred billion. But Cyber Crime Magazine estimates that the economic impact just last year of cyber threats, cyber hacks were three trillion. So we spend a hundred billion, but the economic impact is three trillion. I mean, that just shows you how big the gap is between what we're spending and the efforts that we're putting in and the ripple effects of the economic impact. It's not only, when you look at loss productivity, damage reputation, hits to market cap, et cetera, et cetera, add that up, it's just trillions of dollars, which is nearly the size, it's like three quarters of the size of the IT market, huge. That's a lot of stuff going on. Check that out on the cube.net. Next up, Dave, Intel. Intel announced their ultra AI chip, their core ultra, I call it the AI chip in New York City. I call it the not so splash in the pool because it seemed like a catch up or an announcement. On the heels of what AMD just did, I thought it fell flatter than I would expect Intel, but here's the news. They unveiled the core ultra for notebook PCs and the fifth generation Xeon CPU servers. AI everywhere was the event. It's basically an AI chip, but the highlights are that they introduced the idea of a neural processing unit, which is not new in the system on chip design, making it a significant piece of Intel's architecture. AMD got that first, the TPU with Google. The idea. Arm with the NPU, Apple started shipping NPUs, what, I don't know, four or five years ago. Integrating the NPU is critical. It gives them, and we talked to us on our last pod, actually, when we unpacked AMD's announcement. It's the combination of CPU, GPU, and NPU is really the whole league trinity of the three-way benefits, the flywheel for the new model of AI. And the internet working, which you've brought up many times, internet working, it's what's around the chips, it's gonna create a lot of value. So this is an architectural leap. And so what I like about the Intel announcement, it's directionally correct, but it felt a little bit me too, even though they threw a lot of fanfare around it. And again, fifth generation Xeon, that's legit. But again, the market's competitive, the AMD and Qualcomm are out there, they even introduced NPUs with Ryzen. Qualcomm expected Snapdragon X series with superior AI compute capabilities. We'll see what that comes in. But the industry implications, Dave, is that this now puts, in my opinion, the AI at the center of the conversation around chips. Okay, we've been saying it for a while, but the new step function in architecture is here. Real-world applications and NPUs are gonna be a significant contributor. Okay, as more AI workforce become cloud-based, right? This is gonna be the issue, right? Intel and AMD are focusing on this as is AWS. And then the question is with Intel, and this is what I don't know because they're kind of keeping it close to the vest, can, do they have software developer support? Right, we said this with AMD. You know, if you look at NVIDIA's CUDA, the software abstractions of the holy grail in these new environments, does Intel have the chops for the software developers? And can they bring that AI accelerated application market to the table? And that's gonna be a big issue. They've obviously got a track record in software and ecosystem. It's just that the big difference this time is they're way behind the leader, which is NVIDIA. I thought, you know, look, the announcement, you know, they talked about Gaudi and it's the Habana acquisition. So we're finally seeing some results there. I don't think, I mean, let's see. So a couple of comments. One is when AMD made its announcement, you know, that it was gonna be shipping its GPUs, the stock popped the next day, went up like 15%. Intel was only up like a point yesterday, a little over a point, but it was up. I don't think they didn't set like high expectations for this announcement. So it was kind of a meh announcement, but it's a fairness to Intel. It's not like they were saying it was gonna be more than that. It was really, you know, kind of PC AI. So that's cool. I just see, honestly, I see Intel is kind of number three in this race. I think it's, you know, NVIDIA's gonna have the big market share. I think AMD's got, you know, AMD's market cap is now well over 200 billion, maybe 222 and Intel's under 200 billion. So AMD's bigger company market cap wise than Intel. And so I think Intel's gonna be number three in this race. Dave, who would have thought, could go back a decade. Oh my God. AMD would be bigger than Intel. Who would have taken that bet? A lot of people in Silicon Valley, you know, talking to me about this and they're concerned. They're worried about Intel, but here's how I squint through the noise and all the analysts waving their hands. This is the best thing I've ever seen Intel do. It's not. Really? Is that what the analysts were saying? Cause they're all- There's a lot of analysts out there saying that, but here's how I see the strategic impact. Number one, the AI PC is a new category, right? If you look at, it's not just about the new chip. I think if you look at it and say, hey, if AI is gonna be on the notebooks, think about the gamer market. Think about all the applications it's gonna have at the edge. We talk about compute and inference at the edge. This could be a wide open new category for Intel. And AMD and Qualcomm, by the way, they're not gonna look as well as Apple. Think of Nvidia, my kids who are gamers. They all have Nvidia. In fact, they go into my PC and steal my Nvidia card and swap out the cheaper one cause I'm running basically Windows, right? So they steal my GPU while I'm not looking. They want the best gaming card. That's gonna come to everything. So that's what I like about the announcement. Number two, the landscape and computing, if you look at some of the large language models that we've been covering on open source, they're running them on PCs. And we said this with Michael Dell. And we met with the executives at Dell during their private analyst briefing. We said to them that work bench, that local host developer market is gonna be in the PC, not a server. Well, there's AI everywhere, right? I mean, but okay, so here's my question. And then third implication is, so PC has become the new AI thing, like Nvidia with gaming, to the future landscape of AI chips. And then third, the GPU market is gonna be highly contested, okay? So Nvidia Intel's gotta get in that game and that acceleration, the AI acceleration market is gonna be hot. So three areas that Intel has to win with this. We'll be watching all three. If they fail on any other of these three, I don't see them moving the needle. But I mean, so here's my question is, how new is this? I mean, how much of this is, you know, from chat GPT versus, I mean, three years ago my kid was like, I said, what do you want me for Christmas? So I want a, you know, I want a PC with an Nvidia card. They've been getting this for a while. And so in Apple has had, you know, the M chip for years, right? And so there's intelligence there. And we've had AI on our iPhone doing facial recognition for a long time. So how much of this is, like, how much of this is hype is, I guess, my question versus, and I agree with you, it's gonna be AI everywhere. And the AI is just increasingly more capable for sure. Now that there's so much investment going in, but how much of this is really, you know, new versus normal progression? That's what I'm trying to figure out. Well, your point about your kids in the gaming, I think that's the opportunity Intel has to go after. So they don't want it to be Nvidia. So remember back in the old days when Intel was the dominant one, you wanted the PC with a Pentium and the multiple cores. You know, you actually sought that out. That's the way Nvidia won that game. So I think Intel has to win on faster, smaller, cheaper, and performant. And the kids and people are gonna start tracking this. If it doesn't work, and again, unless we pointed out at re-invent and this whole past few years on theCUBE, price performance hardware matters now. People are gonna be tracking the speeds and feeds. So the AI leadership on the notebook has to be that. And it's akin to Nvidia and gaming as an analogy. For all purpose AI stuff will be, if AI is in every app, you're gonna have notebooks that perform and phones. I tell you, John, I mean, you and I know Pat, we interviewed him 30 times, have great respect for him. If he pulls this off, and I say that, because obviously there's a team effort, but he is the one to let it. He's trying to pull off a Satya Nadella-like leadership transformation. I mean, if he can pull this off, not only is it like CEO of the century, but it's also great for the country. I just, as you know, I'm really skeptical that they're gonna be able to do this at an economic price. I got Ben, how do you say Ben's last name? Baharin? Bajaran Baharin? Anyway, you know how I'm talking about the guy who runs Creative Strategies. I was talking to him at ReInvent, and he's gonna come on a breaking analysis. He was supposed to come on today, but I had my friend's service. But he's really positive on foundry, Intel foundry. Now, if he's right, and he's probably, he knows much more about this than I do, and I'm gonna have, that's why I wanna have him on, to pick his brain about this, cause he's got models and he's into the supply chain. If he's right, that is a huge win for the country. I just, I remain a skeptic until I can really see that kind of progress, but boy, it seems like Pat's doing all the right things. I just, it was very hard to catch up in semis. My first interview with Pat Gelsinger was when I started PodTech, my podcasting company 20 years ago. And he was in charge of the enterprise group. And when Otolini became the CEO, Paul Otolini, that was a moment where I think Pat should have been CEO. I think- No doubt. Had Pat Gelsinger become CEO in 2005, I think it was, with Otolini who got it. It might have been a different company. And I think he's changed, and it's changed a lot since then, because remember, he went to Dell, remember he left Intel to go to work for EMC. Yeah, Tucci hired him to replace Donatelli. So, I mean- What a power move that is. Donatelli was like a legend, certainly on the East Coast and inside of EMC. He leaves to go to HP, and then Tucci hires Gelsinger to replace him. That was like major recruiting coup by Tucci. So Craig Barrett was like one of the key, Gordon Moore, Andy Grove kind of heir apparent. So Otolini was moved up the ranks there as well. But the point is, is that you mentioned about, if he pulls this off, the question on Gelsinger, if he got there sooner, when Otolini was CEO, he might not have gone to EMC. He might not gone to VMware. We might not know what could have been. And I think that's the key. Now the question I've always said is, is it too late? A lot of people at Intel in Palo Alto, where I live and around Silicon Valley, have come to me and said, they're all cheering for Pat because he brings back that old Andy Grove mindset, Gordon Moore, the legacy, Robert Noyce. However, it might be too late. They're all, the whisper is, it's too late. Now that's, and that's what to your point, it's not too late. He made some good news with the government, get the whole geopolitical thing going on. He could pull this off. Okay, the question is, that's gonna be the open question. Can he get in the AI game? Can they get back into the cloud game and not lose that business to custom Silicon? Can they leapfrog arm and compete with AMD and Qualcomm? And there's some serious competition, Dave, okay? And that's gonna be interesting to watch. I'm very influenced by the work that I did with David Floyer on this, who's, Floyer goes back to the IBM days when IBM was the leader. IBM was the number one leader in the world of semiconductors. And that's where Floyer cut his teeth. And he always impressed upon me, Dave, leadership in semiconductors comes from volume. If you don't have volume, you can't get yields to where you want them and you can't get your costs down to be as competitive as possible. And the reason that's so important is because of things like Moore's law and whatever you wanna call this new combinatorial law of GPU, CPU, NPU, et cetera, accelerators, because when you switch to new technology, the costs go up. So you have to have what writes law in your favor, which basically says your costs drop by a constant every time you double your cumulative volume. And so in semiconductors, that's probably, let's call it 15% when you combine it with Moore's law or combinatorial law, it gets down to maybe 30 plus percent, your cost. And so you need that cost advantage. So here's my point. Arm wafer volumes are much, much higher than x86 wafer volumes. NVIDIA volumes in GPU are much higher than anybody else's. So they have the cost advantage. And Intel, when PCs peaked in 2011, they lost their cost advantage and they lost their volume advantage. And to your point about the Silicon Valley Scuttlebut, it's very difficult to regain that, especially when you're fighting a multi-front war in foundry, against AMD, against ARM, against NVIDIA, against China. It's like, let's say if he pulls this off, it's like CEO of the millennium. So we'll see, I mean, Intel has been acting weird at a corporate level. Let's see, it seems like they're hiding something. I'm not sure what the rear of the meat and the bone is. We'll have to dig into it. Anyway, let's move on to AI because that's the topic of why Intel even announced that direction and how they're gonna be able to competitive. We'll track that for sure. Our research team with the CUBE Research will. And of course, I'm writing a research note right now on this topic, I'll publish that shortly. AI. It's a date. Okay. Not a rant section, but quasi-ran to the EU is regulating AI and all by European friends like, oh my God, can they screw up enough? Oh, thank God. The EU is gonna save us from AI. Okay. Out of the EU is a French startup called Mestral, which is mega-funding. Okay, another big funding round, ironically, out of the EU. And then more guardrails coming out. You're seeing more people talk about guardrails at these conferences. And then obviously the open AI research paper came out. So let's talk about AI in the industry, a lot going on there. Let's start with the regulation and the funding from Mestral, 450 million in funding. And NVIDIA, by the way, backed. NVIDIA AMD backed $56 million round for essential AI labs led by transformer architecture co-inventors. So again, more funding, more regulations. Is it a collision course, Dave? You mean between the regulators and the technology industry? Yeah, it is. Innovation needs to run. Why are they gonna stunt at birth, the growth at birth? I mean, you've made this point since we started this Q-Pod is that this is a bad thing. If we have government trying to regulate when they can't keep up, the industry's moving so fast. You know, now, the big question is okay, well, then what's the answer? Can the industry self-regulate? And I don't really know the answer to that. I do know the governments are not gonna be in a good position to regulate. They're gonna have so many contradictions in their laws and in their policies. I just don't see how they can keep up. I think it's gonna stifle innovation if these get threatened. And by the way, it probably will fail anyway because there's so much stuff going on that they won't even be able to catch that I just, I don't see it. It's gonna kill entrepreneurship. It's gonna kill innovation, in my opinion. Here's why. Innovation in AI is like nothing we've seen before in historic shifts. We are at a time now that only happens once a generation, okay? The last time was the web before that PC. You can argue mobile, but I think that's web plus. Still a big inflection point, not generational. I'd say generational, but you will get mobile in there. In these early markets, people have to feel and see the use cases and the user experience change. And if you regulate stuff too early with AI, the companies and the entrepreneurs won't be able to showcase their stuff so people can see and see the value, then understand it better, okay? It's like shutting down the web. Oh my God, people gonna surf the web and view porn or self-serve content without some middleman. That's bad, you know? It's like, no, that's good. That's a feature, not a bug. So in like AI, you want to let it run. Like we said, to quote Andy Grove from Intel, let chaos reign and then reign in the chaos. I think that it'll hurt adoption in areas where there's innovation opportunities and cause it'll not to take a step forward. I think you can take a step forward without getting into trouble with AI and then monitor it. And then the more people step into it, the more eyes are on it, the more transparency there is, Dave. The more openness there is. And I think open source and choice are the key to success, just like the open web was. And I think open AI, open AI being the concept of open, not open AI, the company, because I think they're closed personally, but openness and choice are key. That's why regulation fails. Yeah, so I agree. I don't think, I think the government has done, generally governments that do a poor job on regulating innovation. I think they've done a very poor job on regulating tech in it with the antitrust, the history of antitrust as we've talked about and I've written about pretty extensively, is I think pretty abysmal when you look at, I mean, the example I always use is Bell Labs is now owned by Nokia, what more do you need? Have you read Chip War? Have you read that yet? Yeah. I mean, Bell Labs was this key part of that whole, that fairchild semiconductor. And it's not even the US company anymore. Anyway, the history of regulation and antitrust has not been great. Now, this open AI paper is kind of interesting. It came out yesterday. It's called weak to strong generalization, strong capabilities with weak supervision. And the idea here is normally the way you're training ML models is with human, it's called RLHF, Reinforcement Learning from Human Feedback, right? That's what you do. You get a human level of intelligence to train a less capable than human model and you try to close the gap to human, right? The idea is, here's the basic premise of this open AI white paper and they do tons of research. It's a very thoughtful paper. But the idea is when you have superhuman intelligence that's above the human, can you supervise it with AI, with less intelligent AI? The reason being superhuman intelligence is gonna be writing code, for example, millions of lines of code that humans won't be able to understand. At least certainly not in quasi real time. And so you need AI to basically supervise it. So they tested how much they could close the gap. Now, the problem is there's no such thing as superhuman intelligence yet. So what they had to do was say, can we train or supervise a GPT-4 with a GPT-2 and get it up to GPT 3.5 levels? Why is that important? Well, because what they're testing is, or will, like John, you know how when you and I golf together, I drag you down to my level, because I'm so bad? Right there, so does the more intelligent model learn from the dumber model or can the less intelligent model with multi-shot prompting and other techniques raise up that the total outcome? And so it was a really fascinating paper, but there's a lot of hype around AI, as you know, I'm waiting for the magic eight ball to be declared AI. But so one wonders if, this is like, you remember Goodwill Hunting where the student was like, this shit smells fuckin' easy for me. You know, he throws it in his face, right? You know what I'm talking about? Like, the student is like so much more intelligent than the professor, is it like that? And then the other thing, I wonder, I was talking to a friend of mine who's deep into AI and the government, and he's like, I think we're reaching a Zeno's paradox. Do you know what that is? No. It's remember in high school, your math teacher blew your mind with this when he or she said, okay, the wall is 20 feet away, if you go halfway each time and cut in half the distance, you'll never reach the wall. And you're like, what do you mean? But same point is you gotta, in order to get from point A to point B, you gotta get to the halfway point. And there's so many half infinite halfway points that you never actually get there. And I wonder if we're, you know, at that phase of artificial general intelligence. But it was a really thoughtful paper. I may tackle it in a breaking analysis It's 50 pages. I read most of it in the plane yesterday. And I may tackle it this weekend for Monday's breaking analysis. I think it's worth, I think it's worth doing. And just on a side note on that, Microsoft's multi-billion dollar alliance with OpenAI was reported in the Financial Times today that OpenAI quietly clarifies that the tech giant, Microsoft, has no equity in the company, despite the 13 billion dollars that they have. Despite the 13 billion dollar investment, but it's in line to make big profits. So, I thought they owned 49% of the company. No, that's what they're saying here. And according to, it's not. Now, this comes back down to, not to change the subject, but back to AI. This thing comes back down to OpenAI's governance. Remember, you pointed out on your breaking analysis and we ran it on the podcast, they have that convoluted corporate structure, okay? So, I'll send you the link. And even on their website, they show 49% ownership for the C Corp. This is an article that came out today, okay, three hours ago, Tim Bradshaw from, and multiple people in London and San Francisco. What publication? The Financial Times, yeah, search how Microsoft multi-billion dollar alliance with OpenAI really works. Goes into great detail. After the UK officials said they were preparing to investigate OpenAI and Microsoft's relationship with them. Remember, we've reported that last podcast, I don't know if you remember, but we kind of mentioned that, that the UK competition group, remember they went after Microsoft as well? And they had the whole team's thing. So, Microsoft's under siege from the UK. Again, my friends in the UK are like, they can't get enough of this nonsense. So like, they're like, the government's crazy. People are pissed off in the UK and in the EU, by the way. The younger generation are like, why are you screwing up our future? So that they're just killing innovation. But this article comes in from that inquiry, the government inquiry. So. Yeah, but this is consistent. This is consistent with what I wrote. Microsoft has, Microsoft's only a minority economic interest is no equity. Hold on, let's go with this. But this is word smithing because you got, here's what you have. You see that picture in this article. You got the board of directors, which is a nonprofit company. Board of directors all nonprofit, which is absurd, but let's forget that for a second. The board of directors controls the 501C3, which is a public charity. It's an OpenAI nonprofit. They wholly own the OpenAI LLC, which is the operating company, which controls the holding company for OpenAI nonprofit, plus the employees plus investors. So the investors like Vinod Kosla and the employees own that holding company. And then that holding company is a 51% owner of the OpenAI Global LLC, which is a capped for-profit company of which Microsoft is a 49% owner. So it's true that they don't have any, they have a minority ownership in that LLC, which is the capped profit company. So it's consistent with them. And this is, that's not anything new. It's just they're playing games with words. This is direct quote. However, neither Microsoft nor OpenAI's other backers, which include Thrive Capital and Sequoia Capital own any conventional equity shareholding in the company. Yeah, that's true. It's a very non-conventional. Let me finish. Instead, they're entitled to receive a share of profits from a specific subsidiary of the OpenAI up to a certain limit. So what they're doing, again, back to your point. And, and, and, hold on, hold on, and, and excluding, hold on, and excluding the AGI IP. That's why they structured it this way. So they capped the profits, they gave them basically shadow equity, I would call it. And they don't get any of the AGI IP. So yeah, it's non-conventional. Let me translate for the audience out there. That's this capital structure and Byzantine. They basically have a separate company set up for the profits. So they technically don't own any equity, but they're going to get dividends when they do the step-ups and how they sell shares. Again, this back down to the, why I think this is a set up for potential failure day because now you have a corporate governance problem, even though that Microsoft's got a board of observer seat, the intricate details of the pie, the chart, it's like a flow chart. It's like, it's like writing codes. If then this goes over here, open AI, GP, LLC, wholly owns and controls the nonprofit, but where does the money flow? We put the money in. So the open AI global is an LLC, limited liability corporation. So again, that's usually a flow through entity used to capture profits, not for shareholders. Again, different. Yeah, exactly. So I just put my post into the chat with you and Brendan. And I added some, I annotated that chart that the FT had and with some speculation on what the open AI, GP does, which is some kind of operational role. They probably have some R and D stuff. So, take a look at that, I don't know if Brendan- Hey, look at this, first of all, I wasn't really trying to litigate your post because your post was great. My point is it's a nuance to your post. No, it's consistent. It's consistent with my post. Yeah, you had it right. But the point is, is that the regulators are circling, that's why I brought it up. The regulators are circling and then they got to disclose. So more distractions for Microsoft, more distractions for open AI. They should be spending all their time blowing up this capital structure, blowing up this corporate structure and this Byzantine lines and just reset the whole thing because there's so much money to be made. The interest is gonna end up fighting over who's got what, carcass. This is what happens when a nonprofit controls a for-profit. It's like Harvard University, right? Let's face it, that's a hedge fund. Yeah, if the rest of the people die in Israel then give up their hedge fund status. Right, I mean, so it's a nonprofit. It's like the certain religious organizations, churches, right? They're really for-profit companies hiding in a nonprofit. By the way, speaking of the Harvard, our rant last week, my rant, I think our rant, I know I ranted on, I think you did too. The Harvard, the whole testimony in Congress. Boy, did that go supernova this week, Dave. The CEO of Wharton's gone. They actually voted to keep the Harvard CEO in there because he would not relent. And again, remember the Mike quote? This is the woke culture kind of backlash. Everyone's saying the same thing. Could someone please, like, can we put this chapter to bed? Like, it's terrible. They're just a hypocrisy of justifying the death of Israelis. It's just incredible. Well, the president of U-Pen, president of U-Pen, and by the way, it's like, my heart goes out to all innocent victims here. Israelis, Palestinians, I mean, the whole war just sucks. But so, but the president of U-Pen basically had, it was given every opportunity to say- As was the Harvard. I didn't see the Harvard. It was exactly the same interrogation. It was the same thing. And then the board- It was embarrassing. It was really embarrassing. And then the president of U-Pen walked it back the next day. And she still got fired. Yeah, she should have been fired. That was just poor judgment. But now you think about the open AI board, you know? I mean, what were they thinking? And then, you know, Ilya got caught up in that. And by the way, in that paper, it's funny. The Twitter comments say, where's Ilya on that paper? But he's on the paper. So I wonder if the first version of paper didn't have him on there because of the sensitivity, all the rumors that were floating around about they made some breakthrough on AGI. I don't think this is a big AGI breakthrough, but they're one step closer, I guess. And we still might be, you know, a hundred years away for all we know, or we could be 10 years away or five years or two years. It was very hard to predict. I mean, look, when in 2018, when sort of this whole thing really started to take off from an investment standpoint, I gotta believe investors like Vinod Kozla had no clue when this thing was gonna have its iPhone moment. I mean, I don't think there's anyone who could have predicted it was gonna be November 2022. And then, boom! It's like, you know, five years later, it's like, it's so much hype now around AI. And I think it's important, John, to point out to people that there is a lot of hype. And while a lot of the hype is justified, you know, a lot of it is bullshit. So you just gotta be really careful. I think, again, SiliconANGLE's got a good story on this. Rob has an article on there. And so just check out SiliconANGLE.com, we've got some great articles on this. Open AI, and all the news is happening. And again, now, getting back to kind of open AI, speaking of who's winning and losing, Oracle apparently had earnings and their GPU supply got impacted. Dave, what's your analysis? Yeah, I know, I don't know it was so much the GPU supply, but that could have been part of it. It came up in the earnings call. I know, yes, but there was also, I think, just data center supply, their ability to build data centers. So Oracle, it was kind of a mixed quarter. They paid $28 billion for Cerner, and we knew Cerner was gonna be soft this quarter, and it was, but OCI came in lower than expected. I think 25% growth versus 27% growth. And Oracle said it was because of supply issues. Part of it certainly could have been GPU, but it was also data centers. They can't build data centers fast enough. And so you could see their CAPX was lower than people had expected, but they say they have really good demand. So who knows, it was a buying opportunity for Oracle. We're not, by the way, we're not huge. Yeah, stock was down like 6, 7%, but we're not financial advisors. So you gotta do your own homework. But demand I think is pretty healthy. Their revenue growth was off, was like 5.5, 5.4 versus 6.3% expectation. But they beat operating margin, and so they were able to control their costs, and they did a big, big huge share buyback, like almost a half a billion dollars in share buybacks, which of course will prop that thing up. But I actually, so I've been peaking at the ETR data, the January data. It's really interesting, John. Microsoft, which is always amazingly super strong. I'm not saying Microsoft is gonna feel the effects here, but it's a little bit softer. And I think that's some of that is the, people shutting off co-pilots and the whole open AI meltdown. Amazon is definitely getting a lift from re-invent. There's no question about it. And as is Google from its announcement, but I'll tell you, Oracle and Dell actually are seeing some momentum in cloud, which is really interesting, but especially Oracle and IBM. So you're seeing these hybrid on-prem guys actually seeing a little bit of momentum in what the customers view as cloud. Remember, they think of cloud as an operating model, not a place. And so that's how they answer these surveys. And so, but Oracle is showing, just based on the preliminary ETR data that I saw is showing a meaningful, like a really measurable up-tech and spending momentum, and intended spending momentum. So I believe them when they say, based on the ETR data, I believe them when they say that demand is really strong, we're supply constrained with data centers. And we're finding out from all the conversations we're having and I just did some interviews with some of the people at Pure Storage and they have a flash solution, which is actually great for sustainability. And what we're talking about them about is that in Europe and other areas, but pretty much consistent across the world, the trend is there's not enough room to build data center. The constraint of power and cooling is gonna limit the net new data center. So the scarce resource of our future world is data center resource. As we used to say, the motherboard was the constraint for a PC, but put how many chips you can put on there, power and cooling will be the constraints for the AI world. In some cases, people can't afford to pay their power bill and to run their company. So, because they need expansion. So power energy savings is a huge factor. And it's not so much save the planet, kind of vibe, although that's kind of nice to say, hey, you know, save the planet, rah, rah. It's called save the internet. It's like save AI, right? AI uses a ton of power. And that's, again, not talked about. I remember Pat Gelsinger when we interviewed him at VMware, blockchain just sucks all that power. He's kind of right. And by the way, they used a lot of GPUs too. So blockchain use a lot of power and AI uses a lot of power. I think Rob Strechke and our research team was quoting that every four queries uses like a gallon of water on open AI. So, or some stat, I got the actual stat, but we'll get Rob on that stat. But his point was every three queries on chatGPT is using a significant resource. So imagine that times millions. That's the issue we're dealing with here, Dave. So Larry Ellison on the call said, I'm just looking at the transcript now. I haven't had a chance to dig in. He said, basically that there's virtually unlimited demand for their data centers. He says OCI is going to grow above 50%. He said, the demand quote is extraordinary. We can build the data centers relatively fast and I expect the OCI growth rate to be over 50% for a few years. Yes, says Safra Kats. We are not demand limited in any way right now. I mean, wow. That is, I mean, or it was kicking ass. There's no doubt about it. I think they got flat footed on the growth, but they see the visibility. And as you always said, they are a perfect private cloud company because it's engineered systems, right? So Oracle has a play. Oracle's play of having companies be their hardware, if you will, is going to be important. Now the question is can, this is going to be the strategic question for Oracle. And this is the so-called elephant in the room, as they say, can Oracle get over themselves and realize that there's multiple databases out there? That's going to be, to me, the big thing. And if they do that plus OCI as my private cloud or quote, hardware for AI, meaning it's stored somewhere, as Larry also said, servers are stored in the cloud, there's servers still. So, you know, there's a play there. As you said, Dave, always said, they have an engineered system. And if you look at what Amazon did at Reinvent with the GPUs and cobbling together the Grace Hoppers from Nvidia with the Interconnect, they're essentially building a system, engineering a system. So, you know, what gives? I think Oracle feels like we've got the best database chops in the business. We're the database king. We don't need no other stinking database. I mean, I would love to see Oracle on OCI start offering alternative databases. The question for you is, is that dogma legit or is that real or is it dogma that's going to hurt them? So far, it hasn't hurt them. It's so far it's worked out. How many workloads are running on Oracle? I know Cohere runs on Oracle, but is that a business deal? Well, and now, so look at, it's like we talked about last week. Real workloads. Okay, but it's like we talked about last week on, well, what do you mean by real workloads? But we talked last week about scale-up is coming back. Oracle is the mission-critical cloud. There is nobody in the planet that can run mission-critical workloads better than Oracle. IBM could be the one challenger there, but Oracle smokes everybody on mission-critical workloads. Trust me, this we know. We have so many proof points and data points. Oracle is the king of mission-critical. Now, they're not the king of business-critical. That's on a SQL server, right? And a lot of other databases. Mission-critical on Oracle database, not necessarily cloud. Mission-critical on Oracle database, but because they've got Exadata's running in the OCI, you don't see Exadata's running in Amazon, right? You do see them running now in Azure. So Azure, but Azure doesn't have, you saw Zia's Carrivales post on SiliconANGLE. He was kind of shitting on Microsoft's security posture, rightly so. And so to have mission-critical, you gotta have great security. Oracle has great security. They have great recoverability. They have great performance. If you want, that's why they're so expensive. They can get it. They're running the banks of the world. They're running kind of like the IBM mainframe. But anyway, I wanted to share another little quick tidbit from the earnings call. He said, let me give you one example to your point about constrained on GPUs. We got enough NVIDIA GPUs for Elon Musk's company, XAI, to bring up their first available version of the large language model called GROC. They got that up and running, but boy, did they want a lot more. They want a lot more GPUs than we could give them. We gave them a lot, but they wanted more. We're in the process of getting them more. So no question, GPUs were a big part of that constraint. I totally love that. Again, we'll see. Again, I was just kind of throwing the haymakers out there on Oracle, mainly to ask the provocative questions. They actually might be in a good position to sequence with OCI for some of the demand side things that they mentioned in their earnings that could, as the game shifts, look at, as you always say, Dave, the horses on the track, I always say, the cars on the track, when you have inflection points like this, the order, the pecking order can change very rapidly. The swing of a trend, who's got the tailwind? Who does the better marketing? Always a case. Again, we've been doing theCUBE for 13 years with SiliconANGLE, we keep on now called theCUBE Research. We've seen this movie. We've been through multiple inflection points. It's always the case. High quality, have the right tailwind, take advantage of that tailwind. That's the key to success, whether it's Oracle or a big company. And the winners that I see lining up right now are the ones that see the demand side quickly and get their company in line, get that boat pointed in the direction of the tailwind, just throw the sails up. I think that's what I heard on that earnings call was, we might have made some mistakes or maybe some supply constraints, but we're going hard at the demand side to fulfill it. There's a market, it's real, the hype is good, the sizzle's there, the stake will come. And that's how you figure out the pretenders from the winners, right? Whoever has the sizzle on the stake, win. I see people out there all the time right now, take about acquisitions, sizzle this, all the sizzle, but they have no big capabilities. So, I mean, I've always- If you go private equity, you start doing acquisitions, you're in a roll up, you're not innovating, right? It's like, I see, and that came up in storage. You know, you see the word private equity. I mean, Fitzgerald Fizzi or Charles Fitzgerald Fizzi, he loves to rant on the private equity signaling. Private equity's involved, company's dead. In this market, if you're in private equity, okay, you're not succeeding. Because you have no R&D budget and you have no innovation strategy. I've always had a lot of respectful oracles ability to turn R&D into product. And that's something that IBM, I think, for years was challenged doing. I actually am really encouraged with the sort of new IBM management under Arvin Krishna's direction and Dario Gill coming out of the sort of research facility of their intense focus on turning R&D into product, you know, like, you know, getting their mojo back. Oracle, I think, has always done that. You know, they're not, you know, my mind, they're not, you know, number one innovator in the world, but they take innovation and they apply it. And look what they've done with, look what they've done with MySQL. When they acquired Sun, they got MySQL, they got Java, they got all these kind of cool tools and platforms and they just let MySQL sit there for a long, long time. And then, you know, you had MariaDB come out. And so, but when you see what they did with MySQL heatwave, it's actually pretty remarkable. I mean, you see what Amazon's doing with no ETL, integrating transactions and analytics. Well, that's what Oracle did with MySQL heatwave. Now, they had no market share and so they really had nothing to lose, but they are really innovating at a very, very rapid pace with MySQL heatwave. And they've got the Oracle, you know, the God database, you know, the Oracle database. I mean, I look at them in the stocks that are nearly an all-time high. I mean, you can't, I mean, despite the recent pullback, you gotta have respect for that in my view. Let's move on quickly, more layoffs in media. Wired head layoffs, Fox meetings, other areas are impacted by the whole shift. LinkedIn is not moving to Azure and Hashi Corp's co-founder, Mitchell Hashimoto, is leaving the company to dabble in new areas and tend to his family's first child. He's been on theCUBE 2015's first time on. Hashi Corp Terraform, big success, one public, great company, a generational company, Dave. High quality people, great management team. The only problem that I have with Hashi Corp, they haven't done any business with theCUBE. We haven't covered any of their events since they've all been virtual. But we've had many conversations with founders on theCUBE and they just really did a good thing. And I think him stepping away is a bad signal for the company. Usually when founders stay around, there's a signal there. So, you know, my question's going to be, what's going on with Mitchell? Was there a falling out? Do you feel like he was irrelevant? Are the founders just kind of on the shelf? What's happening? Are they into competitive pressure? We know about the whole open source change of their license. Was that involved? I mean, just questions come up. And his goodbye note didn't really address those. So, again, we're in a really weird market. Companies are falling out of the sky. Startups are failing. Even some of the AI startups are impacted by these big moves by the big whales. The RAG and the companies are doing a lot of the retrieval stuff. Will they be relevant? Open source is gaining ground. So, Dave, you know, this is the, there's winners and losers. There, it's happening. So, layoffs show me that there's an old media guard. They're old companies that are on the wrong side of this going to have to lay people off. Like I said earlier, private equity is a signal of there's no there there. Let's get a buyer, liquidate or shut down. And so, that's an interesting signal. And the LinkedIn, which is owned by Microsoft, not going to Azure, I'm not sure what that means. Is that mean Azure's not adequate cloud? Or is it Azure's kind of all in to support open AI? Where's LinkedIn run today? They have their own data centers. And they have a very complex, they got Kafka for streaming. I mean, they got a well run operation. I mean, let's just be clear. LinkedIn is not shabby infrastructure. They have huge scale. And it's just, it is what it is. They just have their own data center. And they announced their migration in 2019 in great fanfare and it was suspended. They announced a migration to Azure. In 2019, LinkedIn's migration to Azure was announced. And they just suspended it indefinitely. So, they got really kind of good stuff over there. Generally speaking, I mean, LinkedIn's pretty sophisticated environment, started out pretty simple. It was just an online Rolodex and now it's this huge network. Generally speaking, when you do a migration, and I've had a lot of experience with consulting with companies that are trying to do migrations, when you do a big migration, you almost always have to freeze the code. And when you free, because you can't keep making changes, you'll never catch up. You'll have two parallel paths and they'll never converge. And so the question then becomes, okay, when we freeze the code, what's the business impact? You know, there's things are happening so fast in the business, if I freeze the code, I'm gonna fall behind. Not gonna be able to, you gotta fix bugs and you gotta make trade-offs. And so a lot of times the business case for migrating, really that's why CIOs hate migrations. The business case, a lot of times it's not there. So the question becomes, okay, what's the impetus to migrate? Is it, are we gonna re-platform and is gonna be kind of a whole new set of capabilities or is it just kind of a lift and shift? And a lot of times that lift and shift is not gonna get you the kind of business impact that you want. It might save you a few bucks, but it could cost your business on the other end. I mean, I've seen businesses risk their survivability, trying to get off of a mainframe, for instance, because they had old cobalt code and they were gonna save $2 million a year and it cost them $2 billion in market cap. Well, I just forwarded an article to you on the chat about Intel just so we can close that out. Gelsinger was quoted as saying that CUDA is shallow and most shallow. I'm like our AI software. Okay, his quote was the entire industry is motivated to eliminate the CUDA market. He calls it a moat shallow and non-sustainable. He's saying it's a shallow moat. He says, we think of the CUDA moat as shallow and small. Gelsinger went on because the industry is motivated to bring a broader set of technologies for broader training, innovation, data science, et cetera. So. It reminds me, John, they can't do that to our pledges, only we can do that to our pledges. That's our monopoly. It's so cool. Animal House reference for the younger crowd there. Just a good one. Well, food fight is definitely happening here in the chip industry, Dave, on that note. Great, great podcast. Do we have time for a quick rant? My rant is simply gonna be more of the fact that the chip wars are here. And my rant this week is about industry, news, media, commentary. I think my rant is that you're seeing a lot more vanilla content coming out, less analysis, less journalism, good journalism are being rendered down to a few good sites like silkenangle.com, our site, the information in the trade, tech target has still has some good reporting, Beth over there and others. Of course the mainstream media is still gonna be doing their mainstream thing, but you start to see the lack of good outlets and the ones that are faking it, the ones that are getting private equity, they're struggling and you're starting to see the shift. And I think the user behavior of the AI is gonna move to future value of content, meaning quality, flight to quality. So I think we've seen this movie before with the dot com bubble, the flight to quality. When mobile happened, it took two years, apps started getting better. You threw sheep for the first few apps, you played games and the apps got better. I think in AI we're gonna see a one or two year kind of like very elementary things and then flight to quality. And we saw that and it's gonna have an accelerated basis. I remember the mobile days when the app store came out with iPhone. It took about a year, the first couple apps were throwing sheep on Facebook, downloading simple games. And then it got better, Air B&B, Dropbox, real applications came on that were SaaS based. I'll see Amazon clouded a big part of that and that became that wave. I think we're gonna have a similar impact to what's happening now. So my rant is be careful of all the misinformation, all the bad data, the fake it till you make it, content, fake it till you make it, apps, security, phishing, all that good stuff and random. So be careful out there, that's my rant. And the regulation, I couldn't go there, but that's just like continuing, turning to you continue with some rant. But my rant is sort of far cry from the middle. And I don't, I don't, very rarely, if I ever talk about politics, you know, on the cube or in these kidder cupods, a little bit, we get into politics and that's not really our thing. But, you know, Sean Hannity Fox News is unwatchable in my opinion, he's just, he's there to divide. And I think, frankly, Rachel Maddow on the other side is just as bad, but somebody sent me a link of Joe Manchin on Hannity this week. And, you know, whether you like Joe Manchin or not, I gotta do more research on him. He's much more down the middle. And I think so many, so much of today's narrative is meant to divide as opposed to educate. And I think that I would really like to see a candidate, whether it's Manchin or others, candidates for president that are there to bring the country together, as so many have promised to do and have not. So I'm sort of really sick of all this divisiveness. I'd like to see some compromise. I'd like to see some positivity between business and government. The government right now seems to be just out to get business. The U.S. could have owned arm, you know, that they allowed that NVIDIA acquisition to go through. Now, maybe there were some antitrust concerns, but there were probably some things that could have been done with the industry to be able to get the U.S. to get that asset. Seems like every time I turn around, the current administration is trying to knock down business. And I just, I don't like it. And I'd like to see a much more moderate discourse in government, a much tighter private and public partnership, because I think now is the time to bring our country together and compete on a global basis. Amen. I agree. We got to get the politics out. By the way, on a side note on that, I saw Antries and Horowitz, a big VC firm, are gonna weigh in on political candidates for the first time as a VC fund. I mean, they're gonna endorse political candidates that align with their tech vision of the future. So you got accelerationists. You got decels, you got altruism, all kinds of weird stuff happening, Dave, with AI. And again, AI is causing a lot of conversations to happen. Again, let chaos reign and reign in the chaos. That's theCUBE pod this week. Episode 40, have a great weekend, Dave. We'll see you next time. Thanks, Sean. Thanks, everybody.