The Six Five Pod | EP 286: NVIDIA Earnings: Market Reactions, AI Investments & the Future of AI Infrastructure

On this episode of The Six Five Pod, hosts Patrick Moorhead and Daniel Newman discuss the tech news stories that made headlines this week. Some of the handpicked topics for this week include:

1. Saudi Arabia's Tech Investment Surge: The hosts dive into the recent surge of investments from Saudi Arabia into U.S. tech companies. Their core discussion focuses on KSA’s $1.3 trillion investment plan to diversify its economy beyond oil and position itself at the forefront of the AI and technology landscape. Analysts discuss how this massive influx of capital is set to reshape global tech dynamics and highlight the strategic alignment with key U.S. companies, including Nvidia, AMD, and Cisco, to secure advanced technology resources.

2. Anthropic Investments: Microsoft and NVIDIA’s latest investments in Anthropic cement Claude as a strategic “second pillar” alongside OpenAI, with roughly $30B in committed Azure spend and NVIDIA securing another major model customer for its GPUs. The move lets Microsoft hedge its AI bets and ease antitrust pressure, while Anthropic follows OpenAI’s path of going multi-cloud—diluting Amazon’s perceived exclusivity and underscoring that in AI, compute access and distribution are just as critical as model quality.

3. Microsoft Ingite 2025 Highlights: Microsoft Ignite focused on full-stack enterprise agent architecture, spanning Foundry/Foundry IQ for models and knowledge, Agent 365 for governance, and a Windows-based runtime to operationalize agents at scale. The standout was WorkIQ, a semantic layer that learns workflows and coordinates human–agent collaboration across Microsoft 365. Security, governance, and a unified AI-first cloud framework were key themes in the agentic vision that was presented.

4. Yann LeCun’s Departure from META: Citing philosophical and organizational divergence—the former Meta exec has long been skeptical of LLM-centric AI and increasingly sidelined as Meta brought in new leadership layers focused on commercializing generative AI. Pat & Dan note Meta’s shift toward massive CapEx, applied AI, and product-focused execution that’s no longer aligned with LeCun’s academic, research-driven orientation. He’s expected to start a new company focused on “AI in the physical world,” with Meta likely investing, echoing past talent migrations seen at Google Brain and OpenAI.

5. Cloudflare & X Outages: Cloudflare experienced a bad configuration push that propagated across ~300 data centers in seconds, taking down major services including X and even impacting ChatGPT—highlighting how fragile the internet becomes with single points of failure. Many companies rely heavily on Cloudflare for DDoS protection and performance. Hosts speculate that competitors like Akamai and Fastly were likely making calls following the incident. This event underscores the irony of an industry racing toward AI-driven automation while outages are still triggered by basic config errors.

6. The Battle of AI Giants: Google vs. OpenAI 🤖⚔️The ongoing rivalry in the AI sector is heating up, with the hosts exploring whether or not Google can effectively unseat OpenAI as the industry benchmark. Following the recent launch of Google's Gemini 3, they analyze its current capabilities and how the model stacks up against OpenAI’s ChatGPT. While OpenAI has established a potent brand presence, Google holds significant structural advantages, particularly in infrastructure, proprietary data access, and financial resources. This crucial rivalry is broken down for developers and enterprises seeking clarity on the future of AI development.

7. Earnings & Reactions: Explaining the market volatility 📉📈
We take a closer look at the recent market reactions following Nvidia's earnings report. Despite a strong performance, the stocks took a downturn. Pat & Dan discuss the factors contributing to this volatility, including FedSpeak and the impact of crypto liquidations.

‍

Be sure to subscribe to The Six Five Pod so you never miss an episode.

‍

Or listen to the audio here:

Disclaimer: The Six Five Pod is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we ask that you do not treat us as such.

‍

Patrick Moorhead: Six Five Pod is back. Episode 286. Are we jacked? Are we jonesing? We will see.I had a really good workout this morning. It was rough. Bench press. chest, back and arms, just kind of a combo, combo upper body deal. And you know, Dan, I do need to get those legs going. I did have some people on X kind of coming after me and my, my teeny tiny 34 inch inseam legs, by the way, that do have a lot of muscle on them, but I get it when the arms are popping so big, right? I mean, it's going to make the, uh, going to make the legs look tiny.

Daniel Newman: I did hear you're bigger than me now, many times this week.

‍Patrick Moorhead: That is the weirdest thing, because anybody who's actually met you knows the reality behind this. I mean, I get you're like 20 years younger than me, and that gives you an advantage. But you know, Daniel, you never took a break. And in some areas, you're probably twice as strong? Twice, three times. I don't think three times in anything, but we will look at that.

‍Daniel Newman: Four times maybe. Hey, I was happy and I learned something about angles. When I take photos, no more straight square photos. I got to shoulder up a little bit and turn.

‍Patrick Moorhead: Yeah. Hopefully the producers will flash up that infamous picture that got everybody. I mean, I got more responses on that picture than all of the Nvidia earnings. So it just shows what's going on.

‍Daniel Newman: That's because LinkedIn kind of sucks, to be honest.

Patrick Moorhead: No, LinkedIn is the worst. I feel like I'm forced to post there because some people just don't get on X. It kind of blows my mind. Like who, who in tech, who is, is really rocking it is not on X or Y. I just don't understand that. Yeah. I posted about that too, like five or six reasons why, why people don't do that. Um, but, uh, I mean, when the trillionaire club is rocking it on X and having real conversations, not having their comms teams, you know, doing it for them, that's when it really lights up and gets interesting. But we've got a great show for you. You know, we've got the Saudis, you know, doubling down on their trillion dollar investment here in the United States and all the companies that come along with that. More investments in Anthropic out there, more circular. No, just kidding. I don't think this one's circular. Big tent event for Microsoft, for its enterprise customers, Microsoft Ignite, Google Gemini 3 launch. Wow, setting everybody on fire. We're going to debate whether Google has taken over an AI in the flip. And we've got a couple other topics that if we have time that we will get to. So, and after the flip, we debate whether Google has unseated OpenAI as the new benchmark for AI. We're gonna go into bulls and bears. Daniel, I don't know what you said, with your stock stonks pumping out there and your analysis, but dude, freaking market just absolutely laid an egg after.

Daniel Newman: Incredible. All going to zero people sell everything. Not financial advice, but I would just sell everything and go just go to go to paper money.

‍Patrick Moorhead: Yeah, you know, yeah, I was thinking gold, but gold is pretty high right now. So anyways, we're gonna hit Lenovo earnings and Nvidia earnings. And if we have some time left over, we're gonna talk about some Amazon doing bond sale raises and Databricks looking at going out at a $130 billion valuation. So let's start off with the decode where we decode the biggest news of the week. All right, Daniel, we've got. Saudi's kingdom showed up at the White House, pretty much a who's who, they're rocking deals. You've got Nvidia, you've got AMD, you've got Cisco, you've got Adobe, you've got Qualcomm, you've got XAI. What's going on here? We struck out in China, so let's light up the kingdom.

Daniel Newman: I mean, look, it's remember they're investing in infrastructure. you know, U.S. based. The target's very simple. The Middle East realizes that the next foil is technology and A.I. They don't want to be without. They realize that to win over this administration, a massive investment in the U.S. will serve it well. From what I understand, chip licenses started to go out soon thereafter. You know, approvals to start bringing some NVIDIA in. But I mean, look, they apparently have this 2030 vision that looks something like 1.3 trillion that they want to spend on real estate and infrastructure. And by the way, none of this has, it's basically non oil, just to be clear. So it's non oil. Yep. Okay. So, you know, this is, you know, all about aligning with U.S. companies, buying the most advanced technology. And as far as I'm concerned, this is, you know, a positive in terms of injecting liquidity and keeping the investment cycle going. It's a strong non-circular deal opportunity where this is big investment that's gonna go to all those companies that you mentioned. They're going to be building data centers. They're going to be putting U.S. technology into them. They're going to be doing this for the foreseeable future. And, you know, Pat. If I was in the Middle East, that kind of diversification would probably make sense given where we're heading with utilization of oil, different energy, with nuclear, with renewable. I know we're not huge on renewables for powering data centers, but it isn't just going to be turbines forever. At some point, nuclear is going to come up. So they're going to need to invest into all these different things to make sure that they're able to stay competitive. And they want to be a commerce center for the world. And, you know, you actually work closely with some of the groups over there. And this is a big step in that direction for them.

‍Patrick Moorhead: Yeah, totally. And, you know, pulling out right. Saudi Arabia has made most of its money off oil and it sees an opportunity. It has the energy, it has the cash. And with China getting the getting the hand, you know, I don't know, blockade. Yeah. What's that called again? The Heisman. Yeah. Thank you. Thank you. My brain is actually working this morning, but, uh, just had a little glitch there. So, um, Let me just talk about what the different companies are doing. So you've got Cisco, AMD, and Humane, a special JV, right? You've got Epic and MI 400s into Sovereign Clouds. And I think this is just a perfect combination here. I like it. I like it a lot. And I need to do the drill down to see if Cisco is actually doing the compute here. And Humane is really emerging as this sovereign AI integrator of choice. They're kind of the glue layer between these US Silicon folks and KSA Capital. You've got Qualcomm, right? We talked about hybrid AI for a long time. And Qualcomm is opening an engineering center that boots on the ground in Riyadh to create infrastructure. So another move, another move by then. And XAI, right? First tenant inside of these data centers, which I think is pretty cool here. And the U.S. Commerce Department greenlit this after a lot of hand-wringing. and it's not opening up to everybody, but it did talk about some very specific security and observability and authentication checks. It's not impossible to track what everybody's doing, but when the lights dim and somebody's running a frontier model, you do have to go in there and see who is going in there and using it. And I know that You know, the next version of DeepSeek is going to be trained on my aura ring in like 12 seconds. But, you know, the reality is that if China wants to do leading edge frontier models, they probably need to use the latest and greatest. So we will see that the hundred billion dollar plus targets and the capital injection is I guess we'll have to add up. what those numbers mean. Okay, let's move to the next topic here. We have got another round of investments inside of Anthropic. We've got Microsoft and NVIDIA investing money in, and Claude committed $30 billion to Azure. Classic frenemy, right? Is Microsoft hedging against OpenAI? NVIDIA is securing another whale here. Completely natural, right? Everybody's hedging their bets. There is not going to be one winner. And when it comes to AI infrastructure capacity, you have to take what you can get. And most people are willing to pay the price. And NVIDIA is just going all Switzerland on this, which, again, you have to expect. So I think it also minimizes this regulatory optics here. Microsoft is softening potential antitrust heat that goes against that from their, quote unquote, control over open AI. And, you know, it'll be interesting to see how Claude runs on, you know, Maya chips were brought into this conversation as well. I don't know if anybody missed this, but, you know, Claude on Azure would be super interesting, particularly that second generation. The first generation ended up being more of a pipe cleaner here. So, you know, net net, Microsoft, NVIDIA got insurance policies.

Daniel Newman: Yeah, I mean, My read is, you know, it's not, they're not going as big or as crazy as the open AI deals, but you are seeing a bit of that follow suit that compute will be the currency of the future, not just intelligence, the currency, having enough compute. This is why I think the head of infrastructure at Google came out this week and actually said they have to double compute capacity every six months. in order to stay competitive.

Patrick Moorhead: Crazy.

Daniel Newman: Okay. So the whole AI is over thing. I don't know. I think, uh, I think we may be getting ahead of ourselves because these companies know what they need to do. Um, what I will say is, is Anthropic is been a little more conservative. They've spent a little bit less. They're trying to get to profitability a little bit sooner. I think they came up with like a 350 billion valuation though. So they are, marching right behind OpenAI in terms of these exuberant valuations for companies that have no revenue. I think NVIDIA did invest $10 billion back in Anthropic. So that did happen this week, if I'm not mistaken. So all those people that are kind of saying they're investing in and basically giving the money right back for compute, But at the same time, I'll just say this, Pat. You were in corporate strategy and probably worked around corporate venture. What Amity is doing, yeah, you could argue that they're investing in companies that are going to buy their stuff, but isn't that always how corporate venture works? Does corporate venture ever invest in companies that are like, yeah, we're going to use your competition stuff? Again, it's not explicit. You don't write the agreement and say, hey, you're going to do this and now you're going to do that. That's not how it works, but it's implied in many ways.

Patrick Moorhead: Yeah, it is for sure. Also, the companies that invest in there also know that if it's best for their investment to diversify, they happen to be open and the smart people who receive the money won't get locked in to that. There has been a ton of venture investments have gone south because they saddle themselves too closely to one person.

Daniel Newman: Yeah, so I mean, I think that this was, I don't know, my read, it was a wise set of investments. They are a combination. They are, I think you said they're hedging or diversifying. And I also think to some extent, increasingly OpenAI and Anthropic are taking parallel paths, not directly competing on all things. I mean, Thropic has sort of been the de facto code, cloud code, and that's where all their business and revenue is coming from. And OpenAI is much more focused on language and content. And not to say the others don't do the other, but I'm literally thinking that both are massive scale opportunities. And I think NVIDIA and Microsoft want to be attached to both. Probably the biggest sort of at risk in this deal is Amazon though. For Amazon, it's had a certain amount of, you know, seeming or apparent exclusivity around Anthropic. You also see Anthropic following the OpenAI path in that way, where OpenAI made deals with everyone, now Anthropic's doing the same thing. And some of that exclusivity is being somewhat, you know, maybe wiped out. I think Amazon, again, hard to tell why companies took hits this week, but I know on that news, Amazon went down. Yeah.

Patrick Moorhead: All right, let's move to the next topic here, Microsoft Ignite. This is Microsoft's big tent event for enterprises. We were not able to make the show this year, but we had our teammates there front and center, and we were watching from afar. We did get a bunch of pre-briefs, which was really nice too.

‍Daniel Newman: Yeah, it was a big set of announcements, Pat, you know, really about building frontier organizations or frontier companies. The idea that Microsoft is really pushing is, you know, how to build a company that is quote unquote frontier, that embed AI and agents in every part of your workflow. And you saw that a lot across the entire stack. So, you know, they had this, you know, these kind of different IQ layers. They had the work IQ, which is basically able to get into all your data, your emails, your files, all the meeting work. It actually starts to memorize work style and then basically helps you align agents to work. It's kind of creating this new interesting discussion. Pat, we had this a few times at Open Text this week in different interviews where, how does an organization with agents look in the future? They're really starting to tell this story of humans, agents working inside of an organization. And by the way, do you keep your agents in Workday? Where do they sit? Are they employed? Do they get a W-2? Yeah. Yeah.

Patrick Moorhead: It's some fascinating conversations.

Daniel Newman: But they, you know, they added voice imprints or they're adding voice imprints to copilot. So new semantic mechanisms for how we interact. We know this It's kind of a subtle cue as to what devices in the future might start to look like as we move more and more away from always typing into a keyboard and starting to interact with these things more semantically. You saw Agent 365, so they delivered a new control plane that helps you basically manage, govern, register, create access control, monitoring. All these things are really interesting, Pat, and what I think a lot about is like, you know, when do you sort of free agent, meaning like you have an agent doing something, but right now I think all of us would agree human in the loop, very important, but really to get full scale out of agentic more and more, you're going to want to figure out when do I let the agent do the work, you know, really all the way to completion and what's going to build that layer of trust to do that. So that's something really interesting that I know was a focus. They had new announcements, I'm just kind of looking down the list of things I picked, you know, Azure ecosystem, AI first cloud, you know, this is more infrastructure driven, but you know, how do you unify your data, your apps and your infrastructure all as one, still a challenge. Microsoft made a number of announcements around that. And then of course, given all this, they did put quite a wrapper around security and governance. You know, how do you secure agents? How do you secure the data that they have access to? How do you build secure workflows? These were all things that I think were really in focus. And then there's some other ties, Pat, to things like how this lives inside Windows, device, M365. But again, it was a really large one. But the one that caught me was definitely the thing I walked away most interested in was WorkIQ. I really think Microsoft is trying to take the lead in terms of helping set the pace, the standard for what work with agents in a future looks like. And I think they're starting to really tell a story that you can visualize. And then they're putting the wrapper of all the technology required to implement.

Patrick Moorhead: Yeah, a couple of things stood out for me. So I guess, actually, let me start high level, right? This is about Microsoft shipping a full agent stack to enterprises. So models and tools are foundry, the knowledge part of it is called Foundry IQ, and then the governance on Agent 365, and then I'll call it the runtime on Windows, Windows 365 for agents. Maybe the producers can flash up a classic layer cake, birthday cake chart that shows how these work. So I would say right now, this is the clearest story from any hyperscaler so far of this enterprise agent operating model. that I've seen so far, you know, you and I go to reinvent, um, uh, in two weeks and we will see what, uh, AWS says. So, um, it's funny just, just when I'm like, yeah, the layer cake is complete. We keep adding layer cakes. Uh, the one thing that keeps coming up when I have conversation with businesses, Daniel is, I have agents that are in other public clouds, I have agents in my sovereign cloud, and I have traditional infrastructure, let's say a VM. VMs that I want to light up with AI, how do I manage agents there? And the right answer, the preferred answer is not, oh, I'm going to have a different agentic stack for every single cloud. That's just an absolute non-starter. So do a little bit more research to get underneath that. But NetNet, a very good showing from Microsoft. You know, it's funny. I had somebody, I heard somebody say Ignite, Microsoft proved that startups are building toys, but Microsoft is building the factory floor for the AI economy. That sounds like a Dan Ives statement, but I thought it was very apropos. All right, let's go to the Google Gemini 3 launch. You know, Daniel, I remember when Bard one came out and it was so bad that it, you know, uh, they deleted the video that came out of France and then, uh, Bard two came out and it was doing some guardrail nonsense. Okay. That, uh, that was just everybody pretty much hated. And then you had Gemini 1.0 come out and it's image generation. How would I say this? Introduce certain characters as historical characters that just didn't look appropriate. And here we are today, and Gemini 3 comes out. It's multimodal native, and it's faster, and huge context window. It's winning a lot of the, racking up a lot of the awards out there from artificial analysis. Uh, people are going nuts over a nano banana image generator, which by the way, isn't just for, you know, fun images like, you know, Jensen and I working out together and flexing in the mirror, you know, producers, if you want to show that one, you can, uh, but you actually are bigger than Jensen. I don't know. I don't know. We'll see. He has lost a bunch of weight since I recorded him on the 65 pod. He's looking really good. But a nano banana isn't just kind of fun and games. You can actually put, you know, I put a research paper in there and asked it to suggest some cool graphics that would make the paper better. It is absolutely fantastic and you can access Nano Banana through Notebook LM, Gemini 3.0, a lot of different places. Now, Let me just tell you about the downside. This thing hallucinates like crazy. I had two agents that would kick off. I had one agent that prepares me for the day, that goes out to my email and my calendar and out to the web and LinkedIn to prep me for all of my meetings. And that broke. OK. And I mean, it's actually accessing Gmail and it's accessing Google Calendar. OK. And it gave me literally nonsense. It made up a company called Nova and a marketing meeting, a marketing launch meeting that that I had. Second of all, I've got another agent that you're going to like this, Daniel, at 5 a.m. on Friday morning. It goes out to our Google Sheets with all the notes and it pulls back and it suggests certain talking points. for the pot and key areas to drill into. And that said, it couldn't find the Google sheet. So I don't know if this is growing pains or them, quite frankly, folding under the competition of, uh, you know, feeling

Daniel Newman: Yeah, this is all fluid though. They launched something. Every one of these models gets launched, has some gremlins, gets better, but there has to be a point where they turn it on. We're in a sort of continuous improvement and CICD sort of pipeline, agile development. this stuff isn't hopefully, you know, writing your final doctoral thesis is for you. It's not, you know, hopefully, you know, creating your final artwork or whatever else you're trying to. But I think it's got a lot of potential. I think Google continues to get better. You know, they're seeing definite uptick in their use of generative AI. We've seen that in a lot of the numbers, largely because they put those sort of AI generated insights inside search, which has sort of kept maybe some of the Google crowd, boomers, and myself from necessarily always running to chat GPT to do everything. But, you know, I think it's early. I think you got to give it a, you know, give it a few weeks and see how it gets, you know, the bugs and the gremlins get worked out. Because I did notice some of the results were very good, too. You know, I did see some very good high performance. And I think there were some metrics. I actually tweeted those out that show the improvements from generation to generation. And there's a lot of upside in this next model. But like everything, Pat, these things get launched, people beat the crap out of them, they prompt inject them, they try to do crazy things and they find what the QA folks miss or couldn't get to.

Patrick Moorhead: Daniel, all I wanted to do is work with the other products that it actually publishes, right? I'm not asking it to do some crazy thing, right? And then everybody's like, well, you're not using the API. I'm a small business. I'm not going to do that right now. Okay. That person was my son, by the way. who if he's not hitting the API from Gemini or chat GPT or cursor, it's like it's trash.

Daniel Newman: Yeah, lightweight. Well, I you know, I trust Patrick, other Patrick.

Patrick Moorhead: No, I appreciate that. And he appreciates you. All right, Dan. Yann LeCun, the godfather of AI, who we interviewed at Davos last year, is leaving meta because, you know, he basically doesn't believe in any of this LLM stuff. We have hit the limit. What's going on here?

‍Daniel Newman: I think he sees a more holistic AI view than just LLMs. I think he's seeing a meaningful shift. I think there was a lot of talent evolution there when they brought all those big money people in. I think they put Alexander Wayne, the guy from Scale.ai in charge, which from what I understood, put another layer between Jan and the leadership of the company. And I think the company, with the kind of investments it's making, is not as focused on theory and research, which is what Jan was very into. He was a professor. He was a scientist. He was a researcher. He was an academic. And I think the company is very focused on making money with it. And I think that all created a bit of a potential drift between. Now, I will add, Pat, to that, that I do believe he's gonna go start a new company, is what I was reading. And I do believe that Meta is going to invest in it. And it's like a AI in the world, as opposed to just the kind of limited purview of AI through language models. And so all those things are, you know, potentially impacting the decision. But I think it's probably a bit of both. I'm sure he felt a little sideline in this deal. And I'm sure as the company is going to double down on trying to justify these massive CapEx investments, he might not be the right face for that next phase. So this one to me was pretty straightforward. I think it's just a different phase in its life. And, you know, By the way, we want to hear what Jan had to say last year in Davos. Pat and I had a chance to sit down with him there. It was a great conversation.

Patrick Moorhead: Yeah, it's interesting. A couple of things going on here. First of all, when you're at a company for a long time, you grow and the company grows. And if you're not growing in the same direction, it's the time to leave. And in other circumstances is you have a general disagreement with the way that things are being run, or maybe even that you feel slighted because Zuck did bring a bunch of people over him to basically run his lab and he reports to new, very junior, but very brilliant people. And, you know, if, if there's any historical parallel here, uh, there were a ton of departures at Google brain, uh, right before the transformer, uh, explosion. And I don't know if that means that his company or whatever he's working on is going to, um, explode, but, um, yeah. And then, you know, there were a lot of departures at open AI, uh, where, you know, they try to coup and, And then I think half of them went to a startup focused on physical AI, kind of doing what MBZ UAI announced a couple of weeks back and what Fei-Fei, her startup, that's really all about the physical world and how you attack that.

Daniel Newman: I don't become a quantum guy, Pat, with quantum being so unique for kind of solving problems in the real, in the physical world, along with AI and the, you know, and sort of the, and it's interesting.

Patrick Moorhead: Yeah. Hey, let's hit one more decode topic here. And that was CloudFlare went down and X and ChatGPT went down. And just another reiteration of a single point of failure and what happens. We saw this with AWS. We saw this with Microsoft in the same week, right? And hey, it was a bad configuration push. It was an automation error to 300 data centers in five seconds.

Daniel Newman: So we're all like, yeah, man, we're going to automate all this stuff, right?

Patrick Moorhead: IT op is at the peak of AI and automation, and then stuff blows up in our face. So a lot of people go to Cloudflare for DDoS protection. The Netflixes of the world, they use it to get better responsiveness and better latency. The Akamai and Fastly are the two competitors in there that I'm sure are going to making some sales calls this morning. But we did recover. But I got to tell you, you and I had a lot of conversations, Daniel, about, oh my gosh, X is down. What are we going to do? I didn't know how to work out.

‍Daniel Newman: I didn't know how to work out without being able to tweet my workout.

‍Patrick Moorhead: Well, we even said like, are we gonna have to talk to each other at breakfast?

‍Daniel Newman: And that sucked.

Patrick Moorhead: I never want to do that again. That would just be horrible.

Daniel Newman: Usually I just respond to you on Twitter or an ex. You know, I still can't get over Twitter though.

Patrick Moorhead: Like it's still a tweet to me. It's a tweet, man. It's not going to be X. It was a dumb name. And it, it was all Elon Musk ego who, uh, when it was X and PayPal that merged and they took the PayPal name, but he really wanted X to be the super app. And he just keeps leaning, uh, leaning into that. So anyways, Hey, great, great decode. Check out our Open Text World 2025 videos that we have and our Super Compute 2025 videos that we did in St. Louis. But let's get to the flip, you know, where we Debate, take polar opposite, pretend debate, just like a lot of older shows like Crossfire on CNN. I think there's a show on ESPN as well, but hey, let's... Let's go in. So the topic, as I said, is can Google unseat OpenAI as the new benchmark of AI? Let's flip the coin. All right, Daniel, you are for that Google is going to unseat OpenAI as the new benchmark for AI.

‍Daniel Newman: Oh, yeah. I mean, look, this one's so easy. I'll keep it short, Pat. OpenAI has models. That's it. Google has everything. Google is the full stack. Google has the data. Google has its cash flow, so it doesn't have to borrow money and earn cash and lose a fortune for the next five years. Google has its own infrastructure. It has TPUs, seven generations. It's being able to build successful models on its own. It's got its research teams, just like OpenAI does, side to side, and it's got and much better operating leverage than OpenAI does. So, you know, I kind of hit that in the first part, but this one's pretty straightforward to me. You know, OpenAI is the beneficiary. They were the first out of the gate, but they will not be the biggest. They will not be the best. They do not have the tools and the keys to be successful in the long term. And what are those tools and keys? Money, infrastructure, money, data, So you look down the list, Pat, and to me, it's basically very simple. Open AI came out first, had the benefit of an early mover advantage, but in the end, the performance, the model improvements are slowing. In the end, the amount of capital intensiveness that requires to continue to build this stuff is exponentially higher. In the end, the ability to continue to raise capital that they're going to need to raise to be able to get to profitability is now five, six, seven years out. You hear Google's head of infrastructure say that every six months you have to double compute capacity. That means that these incredible trillion and a half dollars of spend that OpenAI is going to need to make may even be more. And they're going to need to raise even more money. And if you look at the way the economy is, if you just look at the way the world is looking at what OpenAI has done, the commitments it's made, it basically does not believe it can get there. All the while, you've got Google, the most profitable company in the world this year. They will make more cash than any other company in the world, including Apple. And they will be able to put that cash to work to be able to hire build infrastructure, expand data centers, invest in LLMs and research. And by the way, they've got massive estates like YouTube and search that continue to provide the world with the most important data. And they run enterprises. Google Cloud has enterprise data. And as we know, a huge part of the unlock is still having access to inside the enterprise and all that data that is not available to Sam Altman and OpenAI no matter how bad they want to steal it from the world, use it to train models and not pay anything. So the TLDR here is, it's all very simple. You can be the best LLM, but that gap is getting closed with every new generation of model that comes out. That's not just Google, that's open source models, that's Anthropic, and that's everyone else. They're fighting battles and demons from every side now. They've got Amazon after them, they've got Microsoft after them, they've got Google after them, they've got Anthropic after them, they've got Quen after them, they've got DeepSeek after them. And they don't have all the other tools in the toolbox. Right now, if I'm going to place my bets, I want it on the one that has the cash flow, that has the operating leverage, that has the customers, and it has the ability to basically continue to build with much less risk. And within three years, you will see Google is greater than.

Patrick Moorhead: I mean, Daniel, that was a nice, simple argument, probably one of your best, one of your better ones. But, you know, you're you're completely wrong. Open AI remains the king. First of all, when any person on the street, a student thinks about AI, they think about ChatGPT. I'm going to use ChatGPT, okay? It's like Kleenex now, right? Nobody's saying, hey, I need to go Gemini this or You know, maybe they say Google search when they want to search, but when they want to AI, it's chat GPT. And I still think that chat GPT or open AI has reasoning leadership. I mean, oh, one strawberry pretty much proved. that OpenAI is still a generation ahead on thinking versus predicting. And we'll see with 3.0 what happens here. ChatGPT has a 73% market share. And yeah, oh my God, it's down from 100%. Right? Let's panic. Panicans. I get that. But you have this company that is 5% of the size of Google with 73% market share. Something is going on here. And I think what that shows is developer loyalty. Let's talk about enterprise trust. Listen, Google Cloud is awesome. It's still the number three, potentially going to be the number four player when Oracle ramps up. But when it talks about enterprise trust, it's through Microsoft. And Microsoft is OpenAI's proxy. They own the CIO relationship. Enterprise security teams trust Azure and OpenAI. And they still view Google Cloud as a distant number three for mission critical apps. My final point, which is probably the better one here, the best one is OpenAI and ChatGPT are just too big to fail. You have AMD, you have NVIDIA. What happens if OpenAI fails? Gemini does nothing with GPUs. Google was very clear on that. Gemini 1, 2, 2.5, and 3. did not use GPUs at all. Jetson did bring up GPUs, which was interesting and related to Gemini. But yeah, so maybe a little bit of confusion there. But Google itself is saying it was all TPUs. So when you have undoubtedly the most powerful and influential company in the world, which is NVIDIA, that needs OpenAI to succeed, where do you think this story is going here? Okay. And then you've got AMD and the alignment and the investment, um, uh, across companies, uh, through there as well. AMD has to, uh, see, uh, open AI as, as successful. So now with that, Daniel, uh, good, good try, but I win. Well, there you go. There you go. Did you win? I don't know. It's good to see you so excited today, Daniel. It's just top of your game, feeling good. We're really excited for this. I love you, buddy. Is that sarcasm?

Daniel Newman: Possibly. I don't like sarcasm. Don't make me shitpost you. All right. I have another side to me, and it's much, much more vindictive.

Patrick Moorhead: Gosh, you're mean today. Go back to bed, get a nap. I was actually thinking I was funny, dark and funny.

Daniel Newman: You look so serious though. I know, I'm sorry, man. All right, good simulated debate here. I love you though.

Patrick Moorhead: Of course I love you. By the way, I won that argument. Dude, not even close.

‍Daniel Newman: Dude, have you seen the Three Stooges? Yeah. You know the one where Open Eye gives money? NVIDIA gives money to OpenAI, and OpenAI gives money to Microsoft, and Microsoft gives money back to NVIDIA, and NVIDIA gives money to OpenAI. So I don't know if it's true, but that's going to be the reason I won the debate is because some people think that's true.

‍Patrick Moorhead: All right. Good simulated debate, folks. Hopefully, we turned up the contrast ratio.

Daniel Newman: Send some replies and let us know who won. Seriously.

‍Patrick Moorhead: Put Dan in the comments. I gave you, Dan, you need to get free peptides for all those who vote for you. Hey, more on that soon, people.

‍Daniel Newman: We're gonna be whoring out peptides on this show. Oh my gosh, you got another one? Not another one. First one. Aren't you done? I'm trying to be. I'm really kind of hoping I can just shitpost in the podcast.

‍Patrick Moorhead: What about do broadcast TV?

Daniel Newman: I mean, if our podcast gets awesome enough, I wouldn't even do that. I like that creature on news, right? You know what I mean? Like, I think that would be more fun because plus it's like, you know, we have more time to talk. Let's structure. You know, we can just shoot the breeze. Imagine doing this all day. We just stay live and people just watch us.

‍Patrick Moorhead: Yeah, that'd be pretty awesome.

‍Daniel Newman: Kind of weird.

‍Patrick Moorhead: Anyway. All right, man, let's go into bulls and bears. I want to do this. We're talking about the latest investment news, the latest earnings, the latest meltdowns, the latest... It's not financial advice. The latest... Yeah, it's definitely not financial advice. All right, baby. Let's dive in. Daniel. Daniel. NVIDIA just absolutely beat, beat, raise, commits to the $500 billion. Everybody raises their CapEx weeks beforehand. And then the market absolutely tanks. What is going on?

Daniel Newman: Dude, there's actually a watcher.guru, you follow that account?

‍Patrick Moorhead: Yes.

‍Daniel Newman: Well, says Jensen, NVIDIA CEO, Jensen Huang says the whole world would have fallen apart if they missed earnings in Q3. I don't know where that quote came from, but. Feels like the whole world kind of fell apart anyway, man. You and I were doing our victory laps, you know, me more than you, fairness, on the way home from supercomputing this week. And, you know, thinking they did the business, futures skyrocketed, everything looked really good, only to see all of those gains in just a matter of a few hours fall apart, basically on two macro pieces of data. One is FedSpeak. A couple of Fed chairs came out questioning whether we needed to continue to cut rates. That quickly brought the excitement and rally to a halt. I'll talk more about what's happening there in a second. And then the second one is there's a clear, hard, heavy selling of crypto. And a lot of people need to understand that crypto, so much of this is bought out leverage. And so when crypto has these super volatile moves, people get liquidated. And when these people with big money and crypto get liquidated, what they tend to have to do is they tend to have to sell equities or other assets to cover their losses. And oftentimes that's a kind of a trickle effect. So I've been saying a lot. Probably if anybody reads any of the stuff I post on X, I talk about kind of crypto is the key. And I say things like that a lot because you can actually get a pretty good read on what the market's going to do by watching what crypto is doing. If crypto is selling hard, you'll generally see the NASDAQ follow. Not all stocks. You'll see like the NASDAQ, the growth names will follow. If crypto is ripping, a lot of times growth will go with it upward. So while they're not technically correlated, like crypto has nothing really to do with AI. What it does have to do is with how growth investors are positioned and how the market tends to, where money tends to sit and where liquidity tends to come from. So that's been really interesting. And there was also this thing, Pat, I don't understand this that well, but there was a couple of segments yesterday on CNBC, Tom Lee, Brian Sullivan covered this on Power Launch. But we're talking about that there were some glitches that was actually causing like liquidations to take place and they were being exposed by market makers to liquidate people's accounts.

‍Patrick Moorhead: 3.6 million views, I think on that.

‍Daniel Newman: I don't know how true or how accurate this is, but like this is crazy because it didn't it literally broke the whole financial system. Because like I said, it's a chain of events. It's not just the liquidation of these crypto assets. But, you know, bottom line was I stand by my assessment that the AI trade actually held up with NVIDIA and you saw that reaction after it. But you see the correlation between so many other parts of the market that if crypto collapses. And by the way, there's a lot of companies and startups and big tech that have crypto on their balance sheet. You've got these interesting leveraged companies like MicroStrategy and Bitminer that have Ethereum and all these things create this domino effect. In this domino effect right now, you're seeing in the sell off of a number of different growth assets, which have come down 30 or 40 percent a week. a lot of them. So this is huge balls. And these are, in some cases, highly speculative companies. In other cases, these are good companies with good earnings that are just, you know, have had high growth and have just been absolutely pulled back. So that's kind of the macro thing. The Fed thing really quickly was yesterday, a couple of speakers came out kind of dovish or sorry, hawkish. And then today, by the way, and I just shared this because yesterday I just need to do a little victory. I was OK if I just do a quick one. I said, we still need to cut. And despite the fact that these comments came out that the Fed wouldn't cut, I said, they're not going to telegraph it. I said, I would eat shit if I'm wrong, but then in December they will still cut. So I will not actually eat shit. I want to be clear about this. I will only. do some sort of public apology for being wrong, which takes a lot for me. But this morning, a new Fed speaker came out, Pat, and I'm not kidding. Yesterday, it was the probability of a rate cut. When all this selling was happening, they said it was 28%. It had fallen from 50% to 28%. A Fed speaker came out today and said he thinks we need to cut rates. Guess what? The current probability of a cut is now in December, from yesterday to today. Take a guess. Yeah. No, I asked you a question to take a guess. Dude, I have no idea.

‍Daniel Newman: It went from 28% yesterday to 70% that we're going to get a rate cut today.

‍Patrick Moorhead: Wild, right? Shouldn't that make the markets go up then? It reduces risk?

Daniel Newman: Maybe, but this crypto bug that apparently is still happening. And just as a header there, Bitcoin's gone from 70, under 81, to almost 87 and back down to 82. So there's still something really weird going on with Bitcoin. Even for Bitcoin, this is extremely volatile. So I need more capitulation as much as I hate that because it's not good for me.

‍Patrick Moorhead: I tell you though, I want to do a victory lap on something, which was, and I might go find all of my broadcast appearances where I said this, when they asked me months ago, hey, what is going to be, basically, what is going to be the milestone where investors are going to be looking at all this investment a little bit cross-eyed? And I said, once companies started looking to the bond markets for financing, And then boom, Oracle comes out. Boom. Meta comes out. Boom. A fricking Amazon is doing a, doing a 15 billion, a billion raise. And then you've got, of course, you know, XAI with the special instruments and stuff like that. So yeah. And that is when you're not just using free cashflow, which by the way, I got from Oracle because They started late in cloud and didn't have that cash machine going like all the other hyperscalers. And that made sense to me. The weird one was Meta, right? Because they seem to be the AI darling in the market. And, you know, we were patting them on the back. Daniel previously in our shows because they were showing revenue and profit uplift via AI. So, hey, Dan, I'm going to call an audible here and I'm going to go next topic being NVIDIA earnings. Um, and you know, you and I both, uh, did a lot of broadcast. I did three shows, Fox, CNBC and Yahoo finance. Uh, you're the king of broadcast and you crushed it, but, um, you know, what are your thoughts? I mean, beat, beat, beat, um, it was up and then it was down 3%.

Daniel Newman: Yeah, I mean, my thoughts are well documented, but if you haven't read them, seen them or heard them yet, you know, what had to happen for the AI trade to be proven intact was going to be data center growth and beating the data center number. It got strong, mid 50% growth year on year, 25% quarter on quarter. So the growth was good. It had to show margin expansion. There was concerns about the impact of inflation, supply chain risks. They said they were working back towards 74 percent, 74 percent people. And they got from 73.4 to 73.6. So a strong move in the right direction there. And then the guide and the guide on all fronts was terrific. So, you know, they beat on all fronts. They showed margin expansion and they were able to validate the next quarter. Basically said every GPU they can build is sold out through 26. And it seems like it's even really beyond that path. So my read was this was tidy. This was great. And again, I really think the factors I talked about the beginning of the segment have a lot more to do with what's going on. I mean, there's still gonna be these concerns about customer density, but I actually published something. I think that's, that's just stupid. I think when you have four customers that sell to 5 million customers, and then your next 10 customers sell to another 10 million customers, you know, they've got a great channel. Why would you want to go sell to every enterprise directly when you could sell to four massive hyperscalers that have millions of customers? They're all by the way, good for it. Some of the stuff about growing inventory, Pat, in my last comments, you have a chance to chime in because I know we're trying to get to the end here. Sorry, people. Pat and I have real work to do. We've got meetings to hang out on. I've never seen a company grow this fast that doesn't have inventory growth, ever. So when people are saying this whole inventory thing, I just don't buy it. Inventory is a snapshot in a moment. They've got a certain amount of inventory. Inventory may be like even racks that are not fully completed and ready to be shipped, but they still can account for everything that's in them. And that inventory needs to go on a balance sheet that needs to be adjusted on a quarterly basis. And that snapshot needs to be taken at a moment in time when the period closed. So like that rising number does not bug me very much at all. The rising, I think there was like five days in aging that was added is like, look, these are, that's a small risk to look at, but when they have the customers that they have, I don't think a slight extension in aging necessarily signals any sort of massive fraud scheme. I think it's more likely just slightly slower pay, maybe due to the size of these deals, the cash flow when things fall. It's something to look at. But again, I just thought that was kind of nonsensical.

Patrick Moorhead: Yeah, I liked on the call, they went right after the bearers, they hit demand, they hit financing, they hit depreciation. And I got hit all three times with depreciation, uh, uh, schedule and, you know, all a series and each series are still in use. We can debate, uh, the, the amount, you know, Daniel, I just look at, uh, the, uh, the, the $500 billion six quarter read, uh, plus hyperscaler CapEx 25 to 26 going up another 40%. Um, the, the, the demand is there. I really don't think we have any risks until 27. I think 27, it's how do we power these new data centers, and then we can get into more of the debate on downstream, who is going to pay for this in the end. But good stuff here. Hey, let's hit Lenovo real quick. Revenue beat, EPS beat, really good read on the net income. They did raise guidance for 2026. Initial stock reaction was up 5%. ISG, finally turning a profit. Actually, they had turned a profit before. The whole AI server business with Neptune is winning deals. Non-PC revenue, 46% of totals, that's SSG and infrastructure. SSG service-led transformation is working. Another double-digit growth here. And over 22% operating margin, it's 18 consecutive quarters of growth, albeit on a smaller base, but it's super impressive that there's not a blip there. It's funny, they're cranking up, quietly putting up in video-like, numbers just without the AI multiple, right? And it's something that I think the company could do a little bit of a better job on related to infrastructure and services on how they be more forward look on that. And I'm not saying the PC and the phone business just because I think we all know where that is and the software has to catch up with the hardware at this point. What potential advantage For Lenovo that could gain on its competitors, this sounds really nitty and tactical, but with memory prices going up 50%, I see what Lenovo did during the pandemic. And they racked up some serious market share because they could ship and others could not. They have a stellar supply chain. And they're also vertically integrated on a PCA and PCB across multiple product lines in the company.

‍Daniel Newman: Yeah, I think you hit it pretty well. I mean, the big inflection, ISG turning profitable. Diversification more and more. We know that the company being able to show that it can be a powerhouse the way Dell has in infrastructure and implementation of big AI is going to be a big important shift for the company to get. kind of that credit in the market that it wants to get. But performance has been stable, Pat. I mean, I think we had some conversations with executives and they were giving us a bit of the readout, which said it's been so consistent, it's almost boring in parts of the business. But that's credit to a well-run organization. Um, you know, there's, there's strength in the devices business kind of keeps it sort of as capped. It's upside in some ways, because people definitely see it for what that is with the devices. But that's also another like kind of kudos because great consistent performance, market gains, share gains. Um, and so this is a company that's firing, it's operating well, it's, you know, pushing to grow profitable in all the right places. And, uh, you know, I think the credit will come eventually. especially if they can land that AI narrative beyond the device. And I think that's the big move that Kami has to make and hopefully will make over the next several quarters.

Patrick Moorhead: Yeah, short term, you know, their one Neo cloud went away from a rocket ship here. And we saw this with with Dell. And then, you know, this slow building of their enterprise capability, the sales, the marketing, and the the AI proof points is is important. And it's also important to know that They were a little bit late. They had a little bit of lack of market fit on the on the AI Neo cloud side on a specific platform and also the deployment services. They're working hard to improve. I do think that they will get a Neo cloud win. I mean there are they're in pretty much all the U.S. clouds or most of the U.S. clouds, just not for these big AI servers. So, hey guys, that is the show. Thank you so much for tuning in. Dana was giving me the hook here, but really appreciate you tuning in. Tell us what you think on the internet. and tell us whose arms you think are bigger, Daniel and I, just based on the picture that you see there. Are you taking those fat burners this morning? Might have. Okay. Anyways, thanks for tuning in. Hit that subscribe button and have a great week. Bye-bye.

‍

‍