We don't need him anymore. There's plenty of great leaders in the ecosystem. There could be a great leader that steps in Nvidia to bring that forward. Now, in terms of what the impact is to the ecosystem, they bailed CoreWeave out four different times. That going forward will not happen. That Bank of Nvidia aspect of this is unlikely to continue. And in fact, in the company's CFO letter, they wrote about all of those commitments and they put some language around the fact that, hey, if it doesn't work out, we're walking away.
Welcome to Prof G Markets. I'm Ed Elson. It is November 18th. Let's check in on yesterday's market vitals. The major indices fell for a fourth day ahead of Nvidia's earnings, as the Nasdaq hit its lowest level since April, and outages at major tech services also weighed on stocks. More on that later. Meanwhile, the yield on 10-year Treasuries declined slightly. And finally, Bitcoin fell as speculative traders readied their ammo for Anthropic's IPO set for tomorrow. Okay, what else is happening?
For the first time in the company's history, Nvidia reported earnings without Jensen Huang. Revenue came in at $115 billion, above analyst expectations and up 103% from a year ago. But it wasn't last quarter's earnings that investors got worried about. On Monday, the Justice and Commerce departments announced a criminal investigation into MegaSpeed, the Singapore company that Alice Wang once ran. They also introduced new required licenses for advanced chip sales. And yesterday, it was announced that Jensen Huang is set to appear before Congress on December 1st, and also that the SEC is officially opening a probe into Nvidia. Earlier in the day, Jensen Huang released his first statement since stepping down as Nvidia's CEO. He said he has seen, quote, an outpouring of calls for him to return, and that he, quote, would be honored to answer these calls and to return as CEO if the board decides that it is in the company's best interests. On the earnings call, however, the chair of the board said that Nvidia will not reinstate him. Shares were volatile in after-hours trading. So the question continues to be the question we have been asking all week: What is next for Nvidia, one of the most valuable companies in the world? Here to help us make sense of the earnings, we're speaking with Gil Luria, head of technology research at D.A. Davidson. Gil, good to see you. Lots to get into. Let's start with your reactions to the earnings. What did you make of them?
Well, considering everything you just said — founder and CEO resigns after 33 years because he slept with the Chinese, by multiple employees, there's an investigation by Congress, the SEC, the Department of Justice, and possibly others — it went pretty well. So we did have, we got good news, which was the first part: the chairman of the board said that they're not bringing Jensen back. They needed to be resolute about that. Had they waffled, it would have left a lot of room for drama. So that's one. And two, interim CEO Klet Press said that not only are they reiterating guidance for next year of 70% growth, they're actually increasing it to 75% growth, and they're doing so on a higher 2026 calendar. So that was better than expected. I wasn't even — I didn't think that they'd even touch guidance for next year. The fact that they increased it, given everything that we just said and everything that they know, that was good news. So considering all that, that was pretty good.
What does it say about what's going on at the company that they were willing to issue the guidance? Does that basically mean that they don't think that all of these risks and all of this turmoil is ultimately going to affect their ability to understand what the business is going to look like? I mean, what does it actually tell us?
That the demand for compute is much bigger than 70% more next year. And they knew that when they guided last quarter, that left them some room to say, 'Hey, we're aware of these investigations about our sales to Asia. We're aware of the new regulations. We're aware of all of that. There is so much demand that we're still increasing our guidance for next year.' So this is about the demand for compute being so big that even if they can't sell as much into Asia next year, they'll still find more than enough room to sell their chips.
I also saw that they are no longer shipping Hopper products to China, and it seems like this would suggest that they're about to get a lot more careful about the China relationship, which, you know, is a shift because they have been selling products to China. It's been a decent part of their business, but maybe that's all about to change. So what do we know about the extent to which they will be dependent on China revenue and also Asia revenue at large? Because I mean, this Singapore thing is the thing that really set this whole scandal off. What do we know about that business going into 2027?
So there's two separate things here, right? Those H20, H200 chips that they were allowed to sell into China have actually not been a big part of the business in any period. That's not what this is about. This is about the fact that their most advanced chips were sold into Singapore and other parts of Asia and then resold into China, and some other chips were deployed in data centers in other parts of Asia and provided to Chinese companies. Those would be violations of our trade restrictions on China. And if Mr. Wong or other people at Nvidia were aware of those things, they are still in trouble. That's not over. And that could be as much as 20% of their revenue. It could be a lot less, but it could be as much as 20% of their revenue. That's why it's so meaningful that Nvidia now knows that they may not be able to sell much of that, and they still think there's 75% growth next year. But let's be clear, those sales are very much at risk, and it reduces the long-term opportunity for Nvidia because China is a quarter of the market, and evidently they were still selling into that market, whether they were aware of it or not aware of it. They were still selling into that market, and that goes away going forward.
Did we learn how concerned they are about that 20%? It sounds like you're suggesting that maybe they're assuming that the 20% just goes away and yet still sales grow by 75%. But do we know how cautious they are about that Asia revenue? I mean, was that brought up? Did they acknowledge it?
They acknowledged that they're aware of the new restrictions, that they're doing an investigation, and with that in mind, they guided for next year. Yeah. So, and by the way, they know, right? They know how many of their chips made it to China. They may have plausible deniability. There may not be anything illegal here, but when you're selling that many chips and you know where it ends up or you know who's using it, they're very much on top of their business. And plausible deniability may have been good up until now. It is no longer sufficient with the scrutiny that they're about to receive. But again, they know they still got into next year. That's at least somewhat of a relief.
So Jensen Wong has been posting away on Twitter. He has talked about potentially coming back. As I mentioned, he talked about these calls to bring him back, and a lot of people want him back. A lot of employees want him back. A lot of shareholders want him back. I've been seeing these open letters and these petitions to bring him back. Do you think that that's going to happen? Is that actually on the table right now?
I don't think that it's a good outcome for investors. A company can function without its founder. That's happened before. Sometimes the handover is shaky. It will be in this case. But Mr. Wong clearly broke internal Nvidia rules, made some pretty big mistakes, and there may be some either criminal or national security issues with his future. So it's a bad idea for him to be in the race for the next CEO of Nvidia. That is not what's good for investors. Now, I know there's a cult of personality, and that's part of our culture, and the leather jacket, and he was such a hero, and he's so wealthy. We have to put that aside. The right thing for shareholders is to move forward. There are plenty of very good candidates that could take over this job and lead Nvidia forward and be successful. They're probably not inside Nvidia, though, because again, they probably knew as well. It's very likely that if something wrong happened, a lot of people internally will be implicated. So the right thing for investors is to bring an external candidate with a clean background that can come in, start institutionalizing their supply chain, the relationship with consumers. No more handshake deals, actual written contracts. No more investing in all of your customers to spread the wealth and raise the ecosystem. Focus on the business. They have the best product. It's the most cost-effective. All their customers want to buy more of their product. That's what they need to be focused on. So I really hope that this Jensen conversation goes away for the sake of shareholders.
Yeah, I 100% agree with you. A lot of people think that, you know, you mentioned the handshake deals, investing in these random startups, and only Jensen Huang had the ability or the power, I guess, the balls, would say, to just drop billions of dollars into these AI companies. And a lot of people say that he was kind of the central banker of the whole AI buildout. And to your point, there are a lot of risks inherent to that. And maybe it's a good thing to get rid of that and go back to being kind of more of a standard corporation. But I think there are some people who believe that maybe that was what was propping everything up. And without those handshake deals, without the circular financing, then something goes wrong with the AI trade, with the AI buildout, perhaps with Nvidia. I mean, something that without that, this is how the bubble pops. Do you agree with that? What do you think of that view?
I think that's all fair, but I just make a distinction. There's how we got here, and there's where we go from here, right? How we got here, Jensen Wong owns a lot of that. There is no Nvidia without Jensen Wong. Frankly, AI the way we know it, with its huge capabilities, with its huge usage that's growing so quickly, with all these very successful model companies, he has a big part of that. That's how we got here today. He deserves all the credit for that. Hopefully, he still goes down in history as a success story. It's the going-forward story that's different. We don't need him anymore. There's plenty of great leaders in the ecosystem. There could be a great leader that steps in Nvidia to bring that forward. Now, in terms of what the impact is to the ecosystem, do I think that people will still use a lot more AI and we're going to all spend a lot more on AI next year? Yeah, absolutely. Which means we're going to need the hyperscalers to have compute capacity, which means we need Nvidia chips. None of that changes. Things that do change are: there's CoreWeave, the special purpose vehicle that Nvidia created in order to diversify its customer base. They didn't just want to sell to Microsoft, Amazon, Google, Meta, and Elon. They wanted to sell to a lot more customers. And so they created this NeoCloud category. They propped it up. They bailed CoreWeave out four different times. That going forward will not happen. That Bank of Nvidia aspect of this is unlikely to continue. And in fact, in the company's CFO letter, they wrote about all those commitments and they put some language around the fact that, hey, if it doesn't work out, we're walking away. And I would not expect them to continue to step in in the same way that Jensen Wong did. Again, maybe he's responsible for bringing us to where we are today, but going forward, we would hope that this type of behavior goes away. It doesn't make sense. They have enough customers without creating new ones. Microsoft, Amazon, Google, Meta, and Elon will buy enough chips to provide the AI compute we need without all these special purpose vehicles that are there just to create a sense of competition that doesn't even need to be there.
I mean, on that note, a lot of people are saying, you know, oh, the AI bubble is going to pop. This is how it pops. And I'm sort of looking at a lot of these names and I'm like, I think the pop already happened. Like I look at, I mean, Nvidia, which lost a third of its value, more than a third of its value in a few weeks. I look at the semiconductor index, which is down 30% in a few weeks. I look at CoreWeave. I look at Iron. I look at Nebius. I look at the data center companies, the NeoClouds cut in half. It seems to me that, yeah, the bubble popped and here we are. This is what it looks like. And I guess the thing that I'm trying to figure out is, is there something else? Is there more pain that we're about to see that is going to bring prices down significantly lower? And I'm not so sure about that because I look at the multiples on every AI name that became exciting and hot over the past few years and they've all gotten crushed. And maybe it's, you know, less catastrophic than 2008 or 2000 because I guess we've got a lot of support throughout the rest of the stock market at the moment, and maybe because the big tech companies have stronger balance sheets and they have existing businesses that are working. But I mean, is there more to go, or is this kind of what it looks like when the bubble pops?
So I think we just prevented ourselves from inflating a bubble, because the valuations for AI companies were actually never very high on a forward earnings basis, and now they've come down to actually being pretty low. So because we were so worried about a bubble and because of everything that's happened, we maybe just prevented ourselves from inflating a bubble. Now there's other things that are happening that are preventing the bubble from happening. One is the capital markets have dried up. AI companies can't borrow money anymore. Criticism of data centers in the political realm has gotten to be so intense that we get more limitations on data center construction. So to answer your question, there could still be downside to forward estimates if we do indeed slow down the rate of data center construction. So again, the numbers we're talking about when we say Nvidia is going to grow 75% next year, that's a proxy for the whole ecosystem. Maybe we don't grow 75%. Maybe we grow 25% to 50%. That's still healthy. That's still growth. And that could still create the capacity of compute that Anthropic, OpenAI, and other model companies can provide us with the AI tools that we are again increasingly using and paying for. So yes, there could be downside to next year's numbers, but I would argue that the reason the valuations are so low is that that's in the numbers. Many investors and many stock prices are anticipating that the growth next year will be slower than originally anticipated. I'm arguing that that's not necessarily a terrible thing. I would rather we build 25% more data centers next year, and then 25% the year after that, 25% the year after that. That's much better than building twice as many data centers next year. It's better for the economy. It's better from a perspective that we won't borrow to build that. Remember, it's the borrowing that makes bubbles. It's the borrowing at 10% that creates bubbles. That's what we need to avoid. So yeah, the next step could be lower estimates for next year. It looks like that's already in the stock.
In terms of capex, I mean, I'm trying to think on what numbers we should really be focusing on going into next year. I assume capex guidance from the big tech companies is going to be a huge indicator of where the AI trade is going. Could you see a world where they cut capex, where they reduce their guidance, and if so, what kinds of impact might that have on the trade?
I think it's possible but unlikely. Because Microsoft, Amazon, and Google have had excellent results in their AI compute business. They all reported not only better-than-expected results but increased their outlook and increased their visibility. Amazon was willing to guide two years out. So I would expect these — now what they've said about capex is Microsoft said they'll increase capex, Google said they'll increase it significantly, Amazon said something in the middle. I think that means 25 to 50% more capex for those companies next year, which again I would argue is healthy. And if they just follow through and do that, because again the demand is there. They've already sold those data centers. The reason Andy Jassy can say 40% next year, 40% of the year after that, he's already sold all that capacity, not just Anthropic and OpenAI, but to every other company in the world. So I would actually expect those three companies to follow through on capex. Again, it's the NeoClouds that have run out of runway. Different story there.
Yeah, companies that really need capital, which makes me think about another company that's about to raise a huge amount of money tomorrow, which is Anthropic. Just as we wrap up, any predictions on how the Jensen Huang story will develop next? I guess one thing that I'd love to get your take on is who's going to be the next CEO? You mentioned that it should be an outsider. Who do you think it'll be?
So I think the leading candidates are Rene Haas from Arm, Charlie Kawwas, the number two in Broadcom, and Johnny Srouji, the head of hardware at Apple. I would say those are the most likely candidates. Again, internal candidates, probably a bad idea. Rene Haas is already CEO of a large public company in chips. He knows the company really well. He's got good connections with the whole ecosystem. That's what makes him the leading candidate. The caveat is, is he too close to Nvidia? Again, there's going to be things that come out. If some of those things come out and involve more people, is he too close? Which leaves the other two as good candidates as well. Decades in semis, very good things we're hearing about them, especially Johnny Srouji from Apple. We've talked to people that report to him that said that he's remarkable. He was a candidate for the CEO job at Apple. That tells you everything you need to know in terms of his gravitas. And then in terms of what happens to Jensen, we'll see how he comes out of it. He's testifying to Congress in a couple weeks. That's usually not pleasant. There's all these investigations. If anything wrong, if we find out anything wrong about him, he'll probably have to ride into the sunset, if he is eventually cleared of all wrongdoing.
Sounds like Elon has a job for him at the tariff. Final question, personal opinion question. Do you think that Jensen Huang knew? Because I kind of think that he did. And I know that we don't have evidence and we can't presume guilt. Like, I get it. But if you're having a relationship with this woman who, I mean, he must have known some basic facts about her. He must have known she was a TV reporter for Chinese state media, then becomes the CEO of this Singaporean shell company working out of these empty offices buying billions of dollars of chips one day. I just have to think that he could sort of connect the dots and be like, 'Oh, this is — there's probably something off here. There's probably something shady.' We even know that the Nvidia people from Nvidia went over to a lot of these companies and found that there were no chips. Like, where are they? So I kind of think that he probably knew, but I'd love to get your views. I mean, did he know? We obviously can't know for sure, but what do you think?
I don't think Jensen Wong knew he was consorting with a spy. I think he should have known, and I think that a lot of people in Silicon Valley are going to think twice about who they consort with going forward.