NVDA GTC、Wiz并购 / Goog 320亿美元交易、4月2日关税不确定性；Huawei Belt & Road；ChatGPT | BG2

If we impose high structural tariffs and allow this crazy Biden-era diffusion rule to stay in place, which makes it hard for us to export our chips, I literally think it's unilaterally disarming America in the race to AI. I think it's a very bad decision and positive for Huawei, right? Like, huge.

Correct.

It's going to lead to a Huawei Belt and Road.

Bill, it's good to see you.

Good to see you, Brad.

As you can see, I'm wearing the Florida Gator hat here as we kick off March Madness. I'm very fortunate that the school I went to, and the team that I rode the bench for, is at the top of the heap. We're the 4th seed according to the NCAA, but in Vegas we're number 1 on championship odds. It's an incredible team with a great young coach, super deep talent all the way down, 4 bigs who are all legitimate, incredible guard play, and an unselfish style. I give them a good shot.

Do you have them winning it all in your bracket?

Yes, but there's some bias. I will admit there's some bias in there.

Can you tell us who else is in the championship game?

I'd have to look at the bracket to make sure.

I have to say, it's good to be reminded that it's March Madness, but it feels like it's global madness in terms of the pace of everything going on in the world. I can't believe I was in Washington last week, and there's a lot of exciting stuff going on with regard to Invest America on Capitol Hill and at the White House.

1. Market Uncertainty

Congrats on that. I'm thrilled about it. The pace of AI and the pace of these changes coming out of Washington—much of which I think is really fantastic—are certainly unsettling the markets. The last couple of days, I've been down at GTC, the big NVIDIA developer event.

While there's a lot of uncertainty in the world, I'm super bullish on where all of this is headed. This is one of those moments when I feel like you have to hold 2 competing but simultaneous truths: things really are accelerating, and we're getting ourselves into a position for this golden age. But some of the caution is still warranted in the short run as we try to figure out how it all unfolds.

For those who listen to us regularly, they'll know that many months ago you expressed this concern that some of the changes were going to create disruption in the financial markets, that you were getting more cautious, and that played out. Everything moves fast these days, but that played out extremely quickly. The markets have corrected, and the Magnificent 7 have corrected. I think the cover of Barron's had the list of the Magnificent 7 and how far down they all are.

Any time you get a correction, people immediately want to know, "Now what?" Has the air come out that you expected, and is it something different now, or do you remain cautious?

I think that was on February 4 or February 6 when we did that podcast, and I said, "I think the golden age will come, but first we have to go through this golden age of uncertainty." We had maximum political uncertainty, maximum economic uncertainty, and maximum technological uncertainty. That causes discount rates to go up, risk premiums to go up, and multiples to come down.

The first leg of that, I think, is that the NASDAQ is down about 10% since that moment. A lot of the components of the NASDAQ are down much more—20% or 30% since that moment. I think that's just the gas of uncertainty being let out.

In the last week or so, we've started rolling over to the second potential leg of concern, which is fear of recession and growth. What I would tell you is that it can become a self-fulfilling prophecy. If people fear—

What everybody is now wringing their hands about is where we are today. I think it's a mixed message. Let me give you a couple of the data points that we're looking at and collecting on both sides of the message.

First, the negative data points have been gathering steam. What are those? Consumer confidence is worsening. The share of consumers worried about their jobs is up to a much higher level than it's been at over the course of the last several years. Business confidence is worsening. There's a record-high share of consumers who think business conditions are worsening.

Then we hear on CNBC from Delta Airlines first, then United Airlines, and then Frontier Airlines that consumer demand is really taking a hit. Airline stocks are kind of the bleeding edge. Am I going to fly to Vegas this weekend to watch the games? I'm a little bit worried about my savings or this or that, so I'll just watch it at home rather than going to Vegas.

I watch that very closely, and we track it. We have this really interesting chart that we'll post on the U.S. airline supply-and-demand balance, and we're seeing a real downtick in terms of consumer air travel. This is TSA passenger growth, so that's a leading indicator again on the negative side.

Are you able to tell if it's consumer traffic or business traffic? People have been talking about government traffic maybe being way down.

I think it's all in there, because it's hard to differentiate between a small business, a sole proprietorship, and a consumer. I think it's all in there.

If you look at the Atlanta Fed GDP tracker, which they call GDPNow, it's turned down pretty significantly. A lot of people, including Kevin Hassett and others, have pointed out the one-time issues impacting this around net exports, but let's just suffice it to say that it's going down.

The Fed just reported today that things are resetting. Literally, right before we got on here, the Fed released its new median dot plot for what it thinks about GDP, inflation, and so on. What they just told us is that the median estimates of all the Fed governors for GDP are lower. They took their GDP estimate for the year from 2.1% to 1.7%. They took their unemployment-rate estimate for the year up from 4.3% to 4.4%, and they took their median expectation for inflation up as well, which is slightly concerning, from 2.5% to 2.7%.

2. Globalism and Tariff Policies

All 3 of those things the Fed just told us are moving in the wrong direction. They're looking at the economic conditions and saying that economic conditions are worsening.

Now let's talk about what we're hearing out of the administration and the other side of that. One of the keys people have to get their head around is that this administration has a principled view about restructuring globalism. This is the political uncertainty I was talking about.

J.D. Vance gave an incredible speech yesterday, Bill, at the American Dynamism event that Andreessen puts on in Washington. He pointed to what he called the 2 conceits of globalism, and said this is the reason we have to have tariffs and restructure the world.

The first one he talked about was the view in the United States that we would just whack up the world. We would do all the high-value stuff, and then places like China would do all the low-value stuff. But, of course, they start with manufacturing, then it moves to precision manufacturing, and then it moves to design.

Before you know it, BYD is designing a better car, likely CATL is designing a better battery, and their vertically integrated supply chain around semiconductors is no longer dependent upon NVIDIA for inference. Now they're running inference at DeepSeek on the Ascend 910 produced by Huawei.

The second thing he said is that cheap labor is fundamentally a crutch to innovation, and that the United States has not innovated and industrialized the way that it should have.

My point here is that when you think about the administration's policies, that's particularly what is causing some of this uncertainty. This is very intentional. Scott Bessent, our Treasury secretary, calls it the American detox period: We need to reset the fiscal and monetary balance in the United States, and there's going to be some short-term pain for longer-term gain.

I don't think we've had a clearing event. We're going to have reciprocal and sectoral tariffs announced on April 2, and the market is churning back and forth, waiting and seeing, trying to figure out if this tips the economy into a negative light.

But what's the positive here? I would say 2 big positives, or maybe 3 big positives. First, Bessent said he's seeing data points from credit-card data and bank data that the consumer is really not slowing down. likely Brian Moynihan from Bank of America was on CNBC this morning and said that consumer-credit growth and bank expenditures from Bank of America were actually up 6% year over year. They had not slowed down, despite the fact that they're getting a little bit more concerned about the economy.

The real-time data does not look that bad to me. Secondly, when I talk to managers—and I've talked to a lot of long-only and hedge-fund managers over the last couple of weeks—they're all in their bottom quartile of exposures, Bill.

So they all kind of heeded the warning from the market and said, “I’m going to get out of the way here for a while.” But, interestingly enough, they all believe that the back half of the year and next year are going to be really great.

There’s this volume dial where they’ve all turned down the volume a little bit, but they all want to get back in before this plays through. So that’s the question: Are we in for another 5% to 10% leg down? Can we find a bottom for the market, or are we going to get a clearing event? I think it’s too early to know.

How do you square your airline data with those other data points?

Yeah. They asked that question to likely Brian Moynihan this morning on CNBC. They said, “How can consumer spending be good if the airline data is ticking down?” And he said they just shift their spending. They’re spending on local entertainment, restaurants, et cetera, while the bigger-ticket items are turning down.

He said, “But in the wash”—and they’re looking at $1.5 trillion of consumer spending per quarter—“in the wash, it’s up 6% year-over-year.” So again, when you think about the Fed’s median forecast, those are forward-looking forecasts. I think what a lot of people are saying is that the data right now is a little mixed, but most of the data indicates that the consumer is still holding up.

But I will tell you, talking to CEOs, CEOs are absolutely more on hold today than they were before. They want to see what these April 2 tariffs bring before they make big decisions.

Do you have any insight or perspective on where that lands? Because obviously, everyone’s talking about this, but there’s a big question as to whether Trump is negotiating or not. Will the bark be worse than the bite, and will what lands be simpler and less disruptive than what was broadcast at the beginning?

I have 2 strong points of view on that. One is that this is not a negotiating tactic by Trump. I think that diminishes the administration’s principled approach to restructuring globalism. You really have to listen to the speech that J.D. Vance gave yesterday. Whether it’s Scott Bessent or Howard Lutnick, they have a fundamental view that the middle class was hollowed out by effectively sending all of this labor offshore. They want to reindustrialize America. This is not just a negotiation to try to get a little bit more.

In that light, Bill, when you think about the tariffs coming on April 2, they’ve said there are about 15 countries. They’re going to outline, country by country, what they call the trade and non-trade barriers and what that totals. Is it $1 trillion? Is it $2 trillion? They expect a lot of those countries to come to the table. They said some of them have already come preemptively to try to cut a deal, and some of them will come after the fact to try to cut a deal.

I don’t see how, if on April 2 we lay this out, that can be the end. It feels a lot closer to the beginning of a negotiation than it does to the end of one. A lot of things will be made clear. One of the things we’ve talked about is whether there are going to be tariffs on semiconductors. There’s a good argument not to tariff semiconductors because we’re in an AI race. Why make it more expensive for US companies? But at the same time, if you just look at reshoring American manufacturing, you would probably say we need to put some tariffs on the import of semiconductors to incentivize the domestic building of a semiconductor manufacturing supply chain.

I will tell you, for what it’s worth—and I’m always open to changing my mind later and would love, perhaps, to be proven wrong so that I could adopt a different perspective—but based on all the learning I’ve done to this point in my life, I remain a big believer in comparative advantage.

I also think that there are people around the globe who want to work harder for less money than people in America could, and they have the ability to improve their lives on a percentage basis from a standard-of-living or wellness perspective more than someone here would for the same marginal effort. I think because of that, water runs downhill, and that’s where those jobs want to go.

Putting up the wall—I’ve said some of this stuff before—but putting up the wall, I don’t think it’s going to work. I don’t think there’s anybody who wants to build a $40 microwave in America. I look at the competition in the auto market, where China now has, what, 35% of the global market, and they’re producing better cars faster. Does anyone really expect GM and Ford to reorient themselves overnight? I don’t even think they could because of the presence of the labor unions.

I don’t know. You and I have talked about Morris Chang’s comments about why he went from Texas Instruments to Taiwan, and I guess one day maybe we get to full automation, which is a completely different world. But even if that’s what you’re encouraging in the US, it doesn’t bring the employment back. It just brings automation here, which has some value.

That’s the part that I think is a little bit of a disconnect between J.D.’s 2 main points. One of them vilified the loss of jobs, but I don’t think the approach brings the jobs back. He’s talking mostly about innovation and was a little derogatory toward cheap labor, which I don’t think is the appropriate way to look at people who are trying to hustle their way up the ladder.

I do think objectively, when you look at the facts, the bottom 40% of the country has not been keeping up. I think you make a good defense—and frankly, a defense that most CEOs I talk to are in agreement with you. I would say even maybe a majority of Republicans on Capitol Hill, Bill, are in agreement with you.

It’s basically an argument against higher structural tariffs. I do think the administration has a principled, structured, and sound belief that unfettered free trade has not been good for America. They have a view that fair trade means the US has appropriate incentives for structurally important industries—whether it’s advanced manufacturing, semiconductors, steel, aluminum, et cetera—to be produced in the United States.

Maybe it’s a cop-out by me, but I do believe there’s a middle ground here. If we just try to quantify it for a second, last year we had about $65 billion in tariff revenues generated by the United States. The real question is, what are we going to do? If I told you, Bill, that we’re going from $65 billion in tariff revenue to $120 billion in tariff revenue, I imagine you would say, “That doesn’t feel like that big a change. We can absorb that pretty easily.”

If I told you we were going from $65 billion to $1 trillion in tariff revenue, that we were going to erect $1 trillion in barriers, I think you would be more concerned about the impact that would have on the economy, not only globally but also domestically. I don’t mean to put words in your mouth, but would you agree with that?

I would just continue to say that when you talk about the hollowing out of the middle class in America, that happened—if you believe it happened—simultaneously with bringing 500 million people out of poverty in China. From a global perspective, is that trade? Did global poverty improve dramatically over that period of time?

Correct.

So you get into an interesting debate. But I also think whether it’s through tariffs or restrictions—protecting the US, and I would say this applies to cars and AI and everything else, by trying to somehow injure, disable, or hinder a competitor that’s in a different country—in the long run, that will make us weaker. I believe that fundamentally.

I don’t think this is about protecting us from global competition. If I believed that’s where the administration was going, I would make a much stronger case for pushback. I think this is about making the playing field more even so that we’re not the stooges on the global stage, where we’re effectively handing over our IP and our high-paying jobs to them.

We’re going to see on April 2 that there’s a firm belief that the Europeans, as an example, are harassing our companies and imposing standards on our companies. All these non-tariff barriers probably add up to trillions of dollars, would be my guess, and that fundamentally creates an unfair playing field.

You’re never going to sell a Cadillac in Berlin, but we’ve got tons of BMWs being sold in the United States. I’m all for making the whole playing field fairer. I don’t really have a great sense for the level of these imbalances, but I’m looking forward to exploring them.

It strikes me as reasonable that, over 40 years where we’ve basically had one-party rule—both Democrats and Republicans promoted free trade and globalism—there’s probably room for optimization. We’ll see if they overtilt on this.

I will tell you this: My belief is that if we’re targeting tariff revenues going from $65 billion to $1 trillion, I don’t think the US economy is prepared for that.

And I think that pushes us into a recession. If we're going from $65 billion to something like $150 billion just by making it fairer across the board—and Scott Bessent has been very clear: if they remove the barriers on their side, we remove the barriers on our side, and all tariffs will come down. If that's the objective of the administration, which, when I listen to their words, is what it sounds like, then I think it's a good policy and one the economy will easily absorb.

One thing I would say, and then we can move on and get into the AI stuff: if reciprocity is the goal, I really think the communication could be done better. That could be calmly discussed if that's the goal, and you can use math and show slides. You don't have to be combative to get to that goal, and I think you'd be more effective. Not that that's why we're here to guide the administration, but I really think that's a rational argument. Bringing a combative tone to that goal, I think, makes it harder to achieve and may create behavior on the other side that was unwarranted and unneeded.

Well, listen. I think the reset you've seen in the market—if you ask what the number-one thing that's striking fear into the market is, whether they're CEOs, Capitol Hill, consumers, et cetera—it's this fear over tariffs and what it's going to mean for the slowdown in the US economy. I think the Fed's median forecast just came down principally because of fear over tariffs and the uncertainty that creates in the economy. So we're going to know a hell of a lot more in the next 60 days.

That's why I think you're not going to see tremendous flows into the market over the course of the next 60 days, because I think a lot of market participants are going to wait, right? They want to see: Is this really just about making it a fairer and flatter playing field, or is this about starting a new big trade war? I don't, again, think that's the case.

One more comment on that, and then I will force us to move on. Zanny Minton Beddoes, who I've met, who's the editor of The Economist, was on this past week on Fareed Zakaria's Global Public Square. She's based in the UK, obviously, but covers all of the globe. She said the attitude among the European leaders is so angry and chafed that there's a likelihood they may come to trust China more than the US.

I just think that's an important reality to consider: that may be happening anyway. If you look at BYD sales into Europe and that kind of thing, boy, talk about picking a fight you didn't want—that would create the opposite of everything we're talking about.

I think for every one of those, I see an equal and opposite around the world.

But I would say, when it comes to more investment in the United States, I think the president's announced $1.5 trillion in new investment. We saw the $100 billion out of TSMC, including, importantly, by the way, building a new R&D fab—not just 3- or 4-nanometer, but an advanced R&D fab—in the United States. That's critical, and so I think we're going to end up in a better place, but we've got to get through that fog of war, like we've said.

So don't expect some big snapback in the market until we start seeing these. And remember, all the goodies—the tax cuts, the deregulation, the better business environment—and this is a good segue because we're going to talk about Wiz and the M&A market—all of the goodies of a pro-growth administration, right, will come when you get the reconciliation package passed. That's a little bit more back half.

3. GOOG Acquisition of Wiz + M&A Environment

So we're taking a little bit of this medicine up front: this uncertainty around tariffs, et cetera. And I think when it's balanced out with the pro-growth deregulation side of this, net-net, I think it's going to be a positive for the economy.

But let me shift gears here, Bill, because this is right in your wheelhouse. Something we've been talking a lot about is the M&A environment in the United States and the IPO market environment in the United States. We got a blockbuster deal announced this week: Google has announced it's going to buy Wiz for $32 billion in an all-cash deal. Of course, most people know what Wiz is. It's in the cloud-security area, which is a huge growth area. They basically monitor your company's workloads that are occurring in AWS, GCP, Azure, et cetera, and what code can be deployed on those platforms.

Rumors are that they're going to do something like $1 billion in ARR in this calendar year. So if you look at it, $32 billion looks like they're paying something just over 30 times forward revenue for the business. I think the business is probably close to break-even. I don't exactly know where that is. Google's total cloud revenue is about $45 billion. So when you look at this, this is about $1 billion on $45 billion, or you're buying basically 2% growth.

It certainly is not coming cheap, considering that Google trades at 5 times revenue versus these guys at 30 times revenue. They clearly think it's strategic to their business. But the more interesting thing: I tweeted this week and said on CNBC last week, the M&A market is back. They've called them off the beach. The corp dev teams are back in the office. They're looking for deals. And this one is going to be a key litmus test for this administration.

So talk to me about what you're thinking about from the logic-of-the-deal perspective and what this may or may not tell us about the business environment folks are entering into with this new administration.

Yeah. Those are both important topics, and clearly the backdrop, to remind everyone, is that it's been really slow-going for M&A, particularly with the perspectives that were held during the Biden administration, as well as the IPO market, which we've frequently talked about. So anything that starts to open that up would be seen as a huge, huge positive data point for the venture market.

I've thought a lot about this with respect to Google, and I think in order to understand it, you have to put yourself in the place Google's in. They're a huge, huge market cap, top 5 in the world, and clearly have been successful with acquisitions in the past. YouTube and Android came through acquisitions. So a company that's been successful at acquisition—and if they wanted to buy something with stock, they could easily do that too. But where can they spend it?

And there's massive pressure from the EU. As you noted, it's very unlikely they would allow Google to acquire anything that relates to search, maybe anything that relates to AI, anything that relates to YouTube. So those may all be off the table.

If you put yourself in those shoes and say you're on the board or on the M&A team at Google, in which of our businesses would we be allowed to acquire? Enterprise is one where they don't have dominant market share; AWS does. It's a more competitive environment with Microsoft and others, and you probably could get something done here.

Within that space, you say, how could we differentiate ourselves? AWS was first. I think AWS is considered a little more developer-friendly. They started in that place from the very beginning. They've been more flexible and open in what they offer.

Microsoft, I think, mostly leverages its corporate customer base and uses that to drive its business. So Google—how can they differentiate? Security, especially in a distributed world, seems like a very reasonable answer to that question.

To put it all in perspective, watching all that, it makes a lot of sense that they would come to this place. Now, they're still saying it might take until next calendar year to get this done. And I think they put up a $3.2 billion breakup fee that they lose if this doesn't get approved. So it's not like this is easygoing.

I remember when they tucked in Waze in 90 days. Even this is difficult, but I can very easily see how they got to this place. I'm not deep in cloud security, but I also have a lot of conviction that that's a good idea for an area of differentiation.

Well, I will tell you this. I hear it from every single company I'm talking to. They're going to be watching this one very closely because people want to get back to the M&A game. And basically, there's been no M&A over the course of the last several years.

One area I want to cover here is just the pace at which we're building monster companies, right? Hat tip here to Shardul Shah at Index, to Doug Leone for another legendary investment. This is a business that was started in 2020, okay? And they're selling it 5 years later for $32 billion. $32 billion—one of the largest outcomes in the history of Silicon Valley. Those guys are both repeat offenders at home-run, power-law outcomes, so hat tip to both of them. Greenoaks was in this.

Bill, I'll share a little vulnerability here about Altimeter and myself, because I've talked a lot about search—what can we learn from search as we go to invest in these model companies? I say, think about all the people who got Lycos and AltaVista and all these logos. They were right about the internet. They were right about search, and they didn't make any money. All the money went to Google.

Well, this is a case where Altimeter was an investor with Sutter Hill in a company called Lacework.

It was competing in a very, very similar space. This was the thesis of Lacework. Lacework raised a lot of money—too much money, I would argue, at the time. They went hard at go-to-market, but they didn't have the product-market fit and execution right relative to Wiz. I remember when Wiz did that Series B, and I said, “We've got a product issue. They're killing us in terms of how fast they're growing on product.”

This was a case where we lost money betting in this space, and some of these great, great friends of ours made a lot of money betting on the competitors. We've had plenty of wins here, but this was one of those humbling times: What did we get wrong in terms of team and execution? It shows how hard this business is, Bill. It is a razor's edge between the promised land and not making anything, because the winners tend to walk away with a disproportionate amount of the prize.

So, any reflections, Bill? You've been around this for a while. $32 billion of enterprise value created in 5 years.

Well, definitely. I would qualify one statement: You said they made a lot of money. It sounds like that won't happen until 2026, so let's not count the chickens just yet.

The one takeaway I have, which my partners don't think is fair, but I'm going to say it anyway, is that repeat entrepreneurs in the enterprise space are a golden ticket. There are so many learned experiences about go-to-market—mostly, I think, about go-to-market—that are repeatable. You can line them up and do it again.

I always think of the Workday example as picture proof of this. It's the team that likely built PeopleSoft, and they just came and did it on the new platform. The technology stack is always changing, and when it changes, it allows for a new entrant to come along and be disruptive relative to that. This has happened over and over again: from mainframe to minis, minis to client-server, client-server to the internet, and the internet to the cloud. It just happens over and over again.

If you have a repeat entrepreneur and you can establish a relationship where they're going to want to come back to you again—shout-out to my partner Peter Fenton working with Bret Taylor again, which is a perfect example of this, and likely Frank Slootman—the odds of success are way, way, way higher.

As you said, hat tip to Leone. He's been around a long time. Having a win this big at the end of your career is just super impressive. I think the key takeaway for me is that a repeat entrepreneur with a new wave in enterprise is probably the best way to increase your probability of success. Even then, you may run into what you did. It ain't easy, but it's a great outcome here. Great outcome for all these people, and let's hope it gets approved.

Yeah. So, let's talk about that a little—the approval process—because I've gotten a lot of questions. It's not only important to this one, Bill; it's important to the signal, right? Are we going to unleash animal spirits around M&A, or are we going to quash them before they get out of the gates?

Let's talk about the FTC for a second. The chair of the FTC, Andrew Ferguson—I think there has been a lot of misinformation about where he stands on M&A. The headlines that I saw over the course of the last couple weeks—in particular, there was a quote that I saw Bloomberg and a bunch of other people reporting, where he said, “We're not going to be deferential to the C-suite. We are going to be the cop on the beat for big tech.” People combined those 2 things and said, “Oh my God, this is a continuation of Lina Khan right here. We thought we were moving into a totally different business environment, but that doesn't sound good at all.”

I went back and read the transcript of a couple of his long interviews, and of course they pulled those 2 things out of the interview, but I don't think it's reflective of where he stands. Just a couple of data points that I would add to the conversation. One is, when asked specifically as to whether or not he differed with Lina Khan, he said really quickly, “I've written hundreds of pages of dissents with Lina Khan.”

He said the number one area he differs with the prior administration is that they lingered. They created a regulatory fear state where nobody wanted to do anything because, literally, they would bring an action and then nothing would happen. It would just die over time on the vine, and it was like purgatory for all these companies, so no deals got done.

He said, “I'm going to be a cop on the beat,” but he said, “Number 1, I'm going to go to court or get out of the way and let business thrive.” That is a totally different mindset: Go to court because you're clearly in violation of the law, or get out of the way and let the business thrive.

That sounds to me like we're going to get pretty quick action here. We're going to know one way or the other what they want to do. The second point he made was, “I'm not a regulator.” Think about that. “I'm not a regulator. I'm a cop on the beat. I'm looking for fraud, monopoly, or collusion.” I think those 2 things may well define a very clear and differentiated approach to the FTC, and I think it's going to be reflected here.

I think we have an unbelievable opportunity to unleash an incredible amount of economic growth and capital finding its best home and best place to grow businesses, with a lot of pent-up M&A. I will tell you this: When I look in the public market, I see great businesses that are growing 10%, 20%, 30% that are trading at 6 or 7 times revenue. This deal just got done at over 30 times revenue. There's ample room for buyers and sellers to meet in the middle if the United States is truly open for business.

Well, I would argue that, at least in terms of Google-Wiz, it passes that test. It's none of those 3 things, so hopefully it will move through quickly.

On the flip side, on the IPO front, there's talk of CoreWeave in the chute, rumors about likely Klarna—I think likely Cerebras is in the chute. It'd be great if we see some IPOs. Do you have any perspective or data on what we might see and when?

My data is—remember, I'm on the buy side. When Goldman or Morgan Stanley want to sell one of these things, our phone rings, and they want to check market conditions. They want to check what we're thinking, and then eventually the roadshow occurs, et cetera. I would say all of that: Our phones are ringing. There's a lot of activity, the ones you mentioned and many others.

So again, think about this: The Nasdaq is down 10%, Bill, in a month, and notwithstanding that, we get a huge M&A deal announced and we have the pipeline full on the IPO front. To me, this is a strong signal that people believe this is much more of a pro-growth administration that's going to be supportive of dealmaking.

If that stuff happens, then I will say, whatever you think about tariffs, whatever uncertainty is sitting in your mind about tariffs, there's a lot of upside from this activity percolating through.

One thing I would bring up, and that is pertinent to April 2, is that you and I just said that Andrew Ferguson at the FTC may in fact clear the way here. But the question is, what are the Europeans going to say? Remember, they've been the ones who've been a pain in the ass on all this stuff, right? These companies now have significant presence in Europe. If the Europeans continue to bite our ankles, I think this is one of the things the administration is going to point out on April 2 and just say, “We can't have this, right? You can't effectively thwart business activity between 2 American companies because you're going to sit there and impose these harsh standards from an M&A perspective.”

4. NVDA Market Position

That, I think, is going to get kicked into the domain of the trade negotiations and tariff negotiations. This would fall into the bucket, Bill, of a non-tariff trade barrier: We're going to punish American companies with excessive regulation that, in fact, hurts our ability to do business. So, this is going to be interesting to see how that one unfolds.

All right, now let's switch gears. I understand you were down in San Jose yesterday. Were you at the SAP Center? I hear there were 18,000 people there.

I saw a photo. His leather jacket, I think, has more zippers and buttons than the previous version. He's clearly the new Steve Jobs, I will tell you.

In fact, I was just down there this morning. I saw our buddy Will Danoff from Fidelity and Gavin Baker. Everybody is there.

I will tell you this: Whatever you think about the upside or downside of the stock—and we're going to get into that, we're going to get into their announcements—what I said to Gavin as I was on my way out is, “Hands in the air in prayer position to give thanks that this is an American company.”

This is an extraordinary company that's at the heart of our competitive advantage when it comes to national security, economic security, all things AI. Jensen is in full founder mode. They're executing brilliantly, and they're an American company.

By the way, we take that for granted. I don't think we should take it for granted, because we just got done talking about how competitive the Chinese are with BYD.

It could very well turn the tables, and we could be dependent upon Huawei for next-generation AI chips instead of the United States. Protecting that level of innovation here in Silicon Valley is important. But let's talk about what we learned.

Just very quickly, while you're on that topic, Jensen was born in Taiwan, and as many have highlighted, many of the founders and leaders of all these incredible American tech companies are either first-generation immigrants or second-generation immigrants. I do hope—I hearken back to when President Trump told the All-In crowd that he wanted to staple a green card to every diploma. I hope those activities are underway. I have not seen any visible activities on that front, but I hope those activities are underway.

No doubt about it. I will tell you, when you think about, in particular, the chip companies—Bill, Lip-Bu Tan, the new CEO of Intel, I think was born in Malaysia—I think every one of them has a CEO who was born in Southeast Asia, China, or Taiwan. I couldn't agree with you more. Smart immigration is a huge national competitive advantage, and we ought to continue to focus on that.

So, what came out of GTC that's the same or different from what we knew? Let's start at the top by saying the stock's basically flat, right? The stock peaked last year around $150 a share. It got as low as $105 the other day. It's back to $115.

On consensus numbers, Nvidia is trading at about 20 times next year's consensus earnings estimates. That's really low relative to the S&P. It's trading at about 24 times this year's consensus estimates. That's again lower than the S&P. This is not a demanding multiple. Why is it trading there, Bill?

Well, people are worried about the demand. They're asking, "Is AI overhyped? Is the demand going to continue to be there? What about DeepSeek? What about the competition from ASICs, et cetera?" So there is a big wall of worry about this. Even among the Mag 7—Apple, et cetera—all of those stocks, Costco's trading at 50 times earnings. So we have a lot of non-tech companies without a lot of growth that are trading at much higher earnings multiples at this point than Nvidia.

Okay. So a few of the key things that came out of this, and then I'll tee up a few of the debates. Number 1, we talked a lot a month ago about DeepSeek, and he was very clear on stage. He said there was a profound and deep misunderstanding of DeepSeek-R1. Remember, when it came out, there was a lot of fear that you could get intelligence on the cheap, that you didn't need as much compute, and that you didn't need to train these big models. He was vehement that that was untrue, and in fact, he doubled down on that, Bill. He said the amount of compute we now know that we need today is 100 times greater than what we believed to be true a year ago.

That was one of the things that was buzzing: What is happening to the aggregate demand, the aggregate TAM, if you will, for the market? I want to drill down on that for a second, because he showed an important slide that we'll have here, which was industry estimates on what the TAM is for AI data centers. He said, "Remember, we did this podcast on February 22nd, Bill, last year, where we laid out the data center rollout." He said that the amount of compute—the world was thinking it was going to be about $250 billion a year, so $1 trillion over 4 years. And he said, "No, that's wrong by half. We're going to do $2 trillion of capex over the next 4 years."

So, something closer to $500 billion a year of new AI workloads, right? Data centers supporting new AI workloads, and then replacement data centers. These are x86-architected data centers that are going to move to accelerated compute.

Well, he doubled down on that again at GTC, and now he's saying that it's going to be $1 trillion per year by 2028. We'll include this slide, but we reforecasted the slide we did last year. This is for both new AI workloads and all of the existing data centers that he believes will be rebuilt. When they get rebuilt, they're going to be more accelerated. So that's going to $1 trillion a year. The market is growing much faster than people think.

Now, why is this happening? I think he used a really great example. He said, as of today, you think about somebody who's writing software, and they're basically hand-coding the software. They can do that without accelerated compute, right? You're just working in a normal database in a normal data center. He said, but by the end of the year, most code is going to either be refactored or written from the start using machine learning. It's going to be coding agents that are helping to do that, and those all require accelerated compute.

If we listen to Mark Zuckerberg, Elon Musk, or Sam Altman, they're all saying the same thing: coding agents are going to come, and they're going to displace a huge amount of the handwritten code that was previously getting done on CPUs. That's the type of thing he showed—some queries and the reasoning models, and why that's multiplicative and leading to much greater demand. But those were, I would say, the high-level points.

He really made the case that demand is exploding across all of these different workloads. Literally, he said this morning that robotics and autonomous cars already represented greater than a $5 billion revenue run rate for Nvidia and were growing very fast. He said they have this incredible stack across synthetic biology, robotics, physics, all the stuff we see in consumer, and enterprise IT, across CUDA, which is multiple layers of software.

Finally, he said, "Remember, we're not in the business of building chips anymore. That ship sailed a long time ago. We're in the business of building supercomputers." Blackwell is 40 times more capable than Hopper was, and the demand is off the charts. Notwithstanding that fact, I think the market's checking them at the gate, and they're saying, "We'll believe it when we see it."

I think a framework for thinking about Nvidia at this point is that this is probably a company at $2.7 trillion that's going to grow at whatever the rate of its earnings growth is, Bill, right? You've seen all the multiple compression you're likely to see.

Yeah. I think, in that regard, they can grind a lot. My reflections—and feel free to push back.

1. As you already said, the market didn't move, which implies, if you believe in the efficient-market hypothesis, that there was full absorption of the data—neither positive nor negative—of what was said relative to what expectations were.

2. I do think there continues to be—and maybe we should just find a way to put a pin in it—this argument about scaling or not scaling. I think no one's in disagreement on the reality, but a bunch of people like to argue it in funny ways.

I do think most people believe that the pretraining scaling, if it hasn't topped out, has at least slowed materially. Even Nvidia is talking about inference being 99:1. The scaling is more on inference. Everyone seems to agree with that. When they say everyone got it wrong, I think that falls into the category of this nuanced discussion, where the bulls don't want to admit that anything isn't scaling, and the analysts are just saying, "Well, it shifted a little bit, and that might matter." Anyway, we can forget about that.

On the token-count thing, no one in the previous world—let's call it Intel's heyday—talked about the CPU units of performance, like how many times the CPU transacted. If they had, you would have had crazy exponential numbers. Also, it wasn't discussed that way, and no one's counted the CPU clocks that have happened over time. So I do find it kind of an ultra-promotional way to talk about things, to talk about the token count, just because we haven't done that in the past.

And on likely Rubin, I'm biased because Benchmark's an investor in likely Cerebras, but I would expect this in a more inference-focused world. Rubin, which is their next generation, has way more on the die, and the die is getting a lot bigger, which is the direction of Cerebras. That one kind of confirms to me that we're moving more toward inference.

The other thing that I thought was interesting, and I'd really love your view on, is that he got up on stage and talked about being the chief revenue destroyer. It was very provocative, and I don't fully understand it. I think he may have created a bit of anxiety in his customer base because he was implying that this next generation is so good that it makes the previous generation quite unvaluable.

The problem is that the companies he's selling to, within the past 3 years, moved from a 4-year depreciation to a 6-year depreciation on these servers. If he's going to kill them in 2 years, there's an economic problem for these customers.

Yeah, I don't—I was there for that. It was, I think, an off-the-cuff remark because the Blackwell generation, the GB200s that they're currently selling—not the future generation, the current GB200s—are 40 times more capable, right, when you look at TCO, than the Hopper generation.

So he said, "We're probably not going to sell a lot of new Hoppers for the cutting-edge stuff that people want to do." In fact, he showed a slide of the CSP demand—the top 4 cloud providers in the United States—and what the year-over-year order was: 1.3 million Hopper GPUs to the top 4 CSPs. Comparing like time frames, he said, "We've done 3.6 million Blackwell GPUs so far this year."

And I think his point really there was the pace of scale-up and scale-out, Bill, and how much utility that provides to the end customer. The point is, because we're investors in CoreWeave, we talked to CoreWeave or we talked to the CSPs. The reason they're elongating the depreciation cycle is because, the way in which they build these, all the software upgrades flow naturally through to Hopper. So Hoppers continue to get better over time, even though you've got the same chip deployed.

And remember, for inference, Hopper's great. You just wouldn't choose to buy more Hoppers today because the new chip is a better TCO than the old chip was. It's not that the old chip wasn't the right thing to buy at that moment in time. I would say the same thing's going to be true about Blackwell Ultra.

What I think is, when you're on the pace that he is, he's got 35,000 employees. They're executing blisteringly fast. They've got the best supply chain in the world, and I think it puts tremendous pressure on the competition. I think the gap between Nvidia and AMD, Nvidia and Intel, Nvidia and all these custom ASICs—in fact, he made a comment that I think roiled some people.

He said about the custom ASICs, “Don't you know that the vast majority of custom ASICs that get designed and get taped out actually never scale into any commercial production?” And we hear that Trainium is now cutting its cost; AWS is cutting the cost on Trainium chips to try to get a lot more customers. You usually don't cut costs on things when things are going great and you've got a really competitive product, right?

I think Trainium's not even competitive with Hopper, let alone with where Blackwell is today. And I think Jassy would be the first to admit they're buying as much Nvidia as they can get their hands on. So I'm not worried about what that comment means for their release cycle.

I think this is one of the few companies in the world that paints a picture like this. Can you imagine if Apple took the stage today and told us what was going to be in the next 4 iPhones? They'd never do it. They don't know, right? This guy is planning out where the next 4 generations of the product are going to be, and I think he's got an incredible path.

He introduced this new improvement around photonics, which I think is pretty incredible. If they execute against the plan, it's going to be a lot bigger business, and I don't see many people able to close the gap on them. I think they'll stay ahead of the rest of the world. So that was my take on things. My question back to you:

Yeah, go ahead. Go ahead. Go ahead.

5. Government Regulations & Impact on AI

You know, he said this morning that he doesn't think tariffs will have a near-term impact on our business. And there's a lot of talk, Bill. We've talked here about the Biden AI Diffusion Rule, right, that makes it much harder to sell these chips to, like, 50 countries—this convoluted system. We've talked about the fact that Chinese export controls are likely going up. We're going to get this announcement on sectoral tariffs on April 2.

Do you have any points of view or any thoughts about what we should be doing with regard to those things and Nvidia? If you were the president, would you be imposing higher export controls? Would you be getting rid of the Diffusion Rule or keeping the Diffusion Rule? How would you think about it?

Let me make one brief comment, and then I'll answer that question. I do want to agree with you on one point. I'm looking at the Magnificent 7, and I'm looking at them ranked by market cap. If you ask the question—I think you would agree with this—which of these companies is executing the fastest right now, and which of these companies is most exposed to global trends, Nvidia would win both of those, correct?

Correct. Correct. Yeah.

I mean, listen, it's a smaller position than it was the last couple years because all of our position sizes are smaller, as I took down risk at the start of this year. But again, I don't expect this to be one of those flash-in-the-pan situations where it's up 3× because multiples expand a lot. I'm assuming multiples stay largely the same, and I just get the benefit from the great execution and the top-line and earnings growth of the business.

And I will say, this slide that I showed you here, Bill: Wall Street consensus expectations are basically that Nvidia tops out at $250 billion of revenue. Somehow there's this ceiling there. If you believe that to be true, then their share of the market falls off a cliff because the market's growing really fast.

I don't think they're going to lose share of market. If anything, I think they probably gain share of market in a fast-growing market. But I do think, back to these Diffusion Rules, there are a lot of people who were around the rim at GTC, and they are very worried about tariffs. They're very worried about the Biden-era AI Diffusion Rule. They're really worried about Chinese export controls.

Well, look, I continue—I said it a few weeks ago—I think the number one risk on the stock is government action from D.C., flat out. A lot of people have been asking questions about the percentage of the revenue that goes to Singapore, and then they replied and said, “That's billing, not shipping.” And then people say, “Well, where is it being shipped to?” There are all these questions about DeepSeek and where are they trained.

I certainly think that it's impossible to stop a startup from any country traveling to Europe or Malaysia and running a model. I don't know how you're going to prevent that—or prevent them from moving the bits back. So I don't think it's a solvable problem, but I also think there's a lot of angst in D.C. about China.

I think a lot of it's misplaced and overly angry. I think your friends at OpenAI added to that last week by putting out an anti-DeepSeek paper that I thought was quite sad. If a company like GM or Ford were to put out a paper like that about BYD, we would look at them and go, “Oh, you just want the government's help. You must be uncompetitive.” So I don't know why we would think about that differently from one of our leading AI models.

But there, I do think that these people have a lot of power. I think it's a bipartisan issue, and I think it's the number one risk on the stock, flat out. My view on this is, if we impose high structural tariffs and if we allow this crazy Biden-era Diffusion Rule to stay in place, which makes it hard for us to export our chips, I literally think it's unilaterally disarming America in the race to AI.

It's a very bad decision and positive for Huawei, right?

Huge. Correct.

It's going to lead to a Huawei Belt and Road. Listen, I already think this has backfired against us, right? DeepSeek is running inference on Huawei 910s because they may not be as efficient as Nvidia, but they just throw a hell of a lot more power—which they got a lot of in China—at these chips, and they can do it.

And part of the reason China has been forced to build a vertically integrated domestic supply chain, literally from design to fabrication around chips, is because the United States made it super hard for them to get their hands on Nvidia chips. So we really have to ask the question: Did we achieve our mission?

There are a bunch of people arguing, “Well, we need to throw even higher export controls on China.” Listen, I don't want to do anything to make it easy on China. I'm fine trying to slow China down a little bit. I just think it's a task in futility.

There's no denying they already have frontier models. They're releasing them every day. Their frontier models are smaller. It's the DeepSeek moment now. Alibaba has one, ByteDance has one, right? So that horse is out of the barn, as we've already seen with BYD, as we already saw with likely CATL. China is going to have frontier capabilities.

The bigger issue is this, Bill: the Diffusion Rule, which the Trump administration ought to throw away and start over. This Diffusion Rule will make it hard for us to get chips to Saudi Arabia, to the UAE, to India, to our friends in Southeast Asia, and it makes it hard through this regulatory capture that some of the U.S. CSPs put in place—all these hoops they have to jump through.

And I just see Huawei walking right through that door and beginning to run the table globally. Frankly, they've done that before. Remember, they said this would never occur in telecom equipment, Bill, because of Nortel and all the highly capable telecom equipment of the United States. Huawei ran the table across the globe with Huawei gear because the U.S. did this. We ought not make the same mistake twice.

By the way, I think it can go further. Every time I've listened to ASML, which is, I think, a Dutch company—is that correct?—and one that has dominant market share, like 80% or 90%, when you hear them talk about the restrictions that the U.S. government has put on a Dutch company to sell into China, I always hear this kind of reluctance and a bit of being angered that the U.S. thinks they have that authority.

And then you take what I said earlier that came from Manny at The Economist. I could easily see us provoking Europe to the point where the ASML team just says, “Well, screw it.”

We don't care what you say. We're selling to China.” And I think that could easily happen. So I don't know. I think we've got to calm all this down a little bit.

Well, one of the things that maybe we'll wrap this little section here—when it comes to what we ought to be allowing China to participate in here, I totally agree, back to the tariffs issue, that it should be a strategic objective of the United States to re-onshore fabs. Right? And you and I have a bit of a disagreement on this.

You know, it gets back to Morris Chang. Morris Chang says it's too hard to do. U.S. workers don't want to do the work. They don't want to live in dorms. They don't want to do the 6 days a week. I don't think we need to do that.

I think that, with the level of automation that we now have in fabs, and given the strategic national importance of having these fabs, I'm happy that this administration has put the pressure on TSMC to build their next-generation R&D lab or fab in Arizona. I think it's achieving the desired results.

What I would hope they would do is maybe announce tariffs but make them conditional, Bill. They say, “If you don't do these things, TSMC, NVIDIA, et cetera, and prove to us that you're building manufacturing capabilities back in the U.S., then we reserve the right to hit you with tariffs, but we're not going to impose them today. We'll have a 2-year delay or whatever while we provide the incentive, but we don't want to slow down U.S. companies in the race to AI.” I think that would be a terrible mistake.

6. Consumer AI Demand

Hopefully, that's the middle ground the administration, guided by folks like Scott Bessent, David Sacks, and others, finds its way to. I think that would be a reasonable middle ground, and I think it is a great objective to get this stuff back in the United States.

But maybe to wrap here, we can hit on a topic that you and I talked about the other week, and this gets back to a couple of questions. One, just thinking about it: We had the launch of Grok—big bang—the launch of DeepSeek. Maybe we just check in a little bit. Gemini's had some updates on where we are in the state of consumer AI demand and, importantly, the topic that I've been reading about around contribution margins. Even if revenue is growing, are they selling dollars for 50 cents? Are they losing money on each incremental unit of business?

When I look in the App Store, DeepSeek jumped up to number 2 and was there for a couple of days. ChatGPT was at number 1 at the time in App Store downloads. Now DeepSeek isn't in the top 100, the last time I checked. When Grok jumped up to number 2 for a few days, it's now around 65, the last time I checked. Gemini never really jumped up; it's hanging out around 55. ChatGPT is still number 1 in the App Store.

The point being, I've said for a while now about consumer markets that all my pattern recognition from Google and Meta is that these tend to be winner-take-most markets, and that it's almost impossible to dislodge the inertia. You can't get there by being slightly better. At this point, you're going to have to be 10x better than ChatGPT to slow down that inertia. That was the learning from these prior periods.

Now, we've seen the reports of over 400 million weekly active users. I would tell you I believe that OpenAI is massively supply-constrained. I think they're building these 2 huge campus data centers, one in Abilene, Texas, and the other one in Denton, Texas—one with Oracle, one with CoreWeave. I think they have to build all that just to actually launch the products that they currently have. That's not even supporting future growth; it's just supporting the demand they currently have. I think they have 4 or 5 products literally sitting on the shelf because they don't have the compute to deliver them to customers. Any thoughts there, Bill? Have we seen—do we know the winner already in consumer AI?

Well, I mean, I think your argument is very solid. I don't think there are any data points that would suggest that there's an immediate threat to that. I would think that the powers that be at Google and xAI, Meta, and Amazon must be up all night trying to solve this problem.

And so I guess Netscape had Microsoft; OpenAI has 4 or 5 of the most powerful companies in the world aimed at them. And so you're right: They have a lead. It could be insurmountable. They have all the king's horses and all the king's men behind them. It'll be interesting to watch.

We've talked about the things. I think if someone crushed voice, it might give them an advantage to move fast. That might be particularly compute-intensive, which we'll get into in our next topic. But for now, I don't have any ability to take the other side of your argument.

Well, and you know the couple I didn't mention there? I didn't mention Meta, right? We know those guys are in full beast mode, but they've been unusually quiet. I'm actually shocked at how slow we've seen any change there in terms of a standalone consumer app. I know it's coming, but I've heard it's been coming for a long time—not even the integration within their existing apps. I'd love to talk with Zuckerberg about that.

And the other one is Apple, right? We've talked about that again, just not seeing any real change there. In the meantime, I'm seeing the stuff that we're going to see over the course of the next 8 to 10 weeks out of OpenAI, as they get these releases teed up. I think it's going to be pretty profound and continue to add pressure to that lead.

7. AI Unit Economics

One thing I've been thinking about: DeepSeek, by choosing to be more open than Llama, doesn't have these caps. Llama, Mistral, and a few others claim to be open, but if you get too big, you have to pay the piper. There's always been a continuum of open source; every company chooses its spot on that continuum and tries to see what it can get away with. DeepSeek is so prevalent in the enterprise right now, and it's been forked like 1,500 times on Hugging Face. For all the reasons open source works, that's why it's happening. I think it's an interesting moment for Meta to consider going left and getting even more open with Llama, which could be powerful for them. This probably won't happen, but even OpenAI could consider doing that. If you're a consumer product company, there's not much risk in taking that leap, and it would undermine anyone else trying to compete based on the quality of its proprietary model. Anyway, I wanted to bring that up. Where were you headed?

Well, listen, you've been appropriately tough in asking questions about gross and contribution margins for these businesses, right? I saw a report recently on Anthropic suggesting that they had lower contribution margin. So let's just first unpack what that is.

Contribution margin is basically what I have to pay as a variable cost to provide the service that I'm providing. In the case of Anthropic, most of their enterprise API, I believe, is sold through AWS, and they have to pay a huge customer-acquisition fee, probably to Amazon. In the world of Google, we would call it TAC. Then, of course, you have all the costs that you have to pay to actually serve the model, right? There's a variable cost associated with serving the model.

When you take those 2 things into account, because of the large percentage of their revenue that is indirect, they have very little or no contribution margin, right? So there's no net revenue really flowing through the P&L. Now juxtapose that against somebody like OpenAI, where almost all of their business is direct. They don't have to pay traffic-acquisition costs or revenue share to anybody, right? That obviously has to be the business model because the costs of serving these things are expensive.

Now, of course, both of those things ignore all your overhead and your training costs, right? So that's not truly fully loaded, but I'm just looking even at the contribution-margin level, because if you're not meaningfully positive at contribution margin, then you have no chance of covering the costs on a fully burdened P&L.

Here's what I've been thinking about a lot. I will introduce—let's just call this loosely, and in a first draft, an idea that I'll call the Gurley negative gross-margin AI theory. Okay, here we go.

My brain can wander in some pretty interesting places all by itself, but a lot of people have been comparing things to the internet revolution. I know why they do that. There's so much excitement, for all similar reasons, and so they want metaphors and comparisons.

In addition to that, I would say—and you just explained it—the unit economics are messy here. You have variable costs, especially if you're buying your compute from someone else. You may have CapEx. Then there's a question of how long you're depreciating that over, and whether you charge the unit economics against that.

Some of these players have credits that they were given by the large platforms as investors. Do you really take that into account or not? In addition, you have tons of VC money, and so if you're going to raise billions of dollars, you're going to have high burn rates.

These things are just part of the game. In addition, I think there is a mindset that most VC markets are winner-take-all, and so you can't lose market share, right? You just can't lose market share. So if you can't lose market share, how do you price? Do you price to market, or do you price to cost and value like you would in an economic textbook, or do you just price to market, meaning, “I can't lose because of price,” and so it becomes a buyer's market?

I'll create a theoretical example because I think there's also an issue of stacking. Let's say you have—and let's not pick on any company—a large model provider that's buying compute from a hyperscaler. Then let's say there's a startup doing voice for the enterprise on top of that model provider, and all these people are buying compute from the one below them.

If those second-layer and third-layer companies are both negative gross margin, the consumer is buying compute from the hyperscaler at a price that's lower than it would be if they bought it directly, because it's being subsidized by 2 players in the middle. On top of that, if this is true, you're triple-counting revenues, right? There's a transaction that a consumer is making against this model, but because it's negative gross margin, you're literally adding it up 3 times. The same thing would be true if you were in what people would consider a really shitty business—if you were a global distributor of retail products and you were taking 5%, although here it's negative.

This theoretically could be happening out there. You pile in tons of VC money, and you have the fact that everyone believes these markets are winner-take-all. The founders are likely unsophisticated financially—not, I'm not taking shots at them. They probably haven't had a lot of finance classes, and you don't have great visibility into unit economics.

It's set up to be messy, and you could have substantial resets as people get in touch with unit economics when and if they're forced to. That's a very unpredictable time window for when that would happen. This is a theory; I'm not suggesting it's 100% true. It's possible this is going on.

Well, listen, I think the one place I would absolutely agree with you, Bill, is that the unit economics and the ability to compare across these businesses are messy. That's number 1. Number 2, what I would agree with you on is that in no way are these business models proven yet to be anywhere close to as good as Google.

Remember, Google raised less than $50 million before they went public, right? Google was able to do what they did with very little upfront capital, and then their unit economics from the very start were profitable. This is a really important point: in the internet businesses that worked, the marginal unit cost to serve a customer was near zero. For Google, it was near zero, which is partially why it was so economically efficient. I think that was true for Meta also, right?

I will tell you, here's my takeaway: for whatever AI companies emerge with a consumer business model or an enterprise business model where they fully control the top of the funnel, where those people are coming to them at almost no incremental cost because their friend told them about it or whatever, there’s no marketing cost. Your only cost is actually serving the inference to support that customer. I think those will have unit economics that are exceptional and that, at maturity, look a lot like the internet companies.

But I will say that right now, you hear these top-line revenue numbers, and the first thing you ought to be asking yourself is, who do they have to pay to get that revenue? What are they loading in terms of their variable cost against that revenue? Time will tell, but my sense is there aren't very many sustainable business models here today.

What I mean by that is that to support this level of capex that these companies are undertaking, you're going to have to have tens of billions of dollars in high-margin revenue in order to support the reinvestment that you're making on the back end of these businesses. Obviously, I think the best-positioned one at this moment, subject to change, in terms of the independent players is OpenAI. But remember, Meta can invest. They have a printing press kicking out billion-dollar bills in the back room, so they can invest against this for a long time. Amazon can. Google can.

To your earlier point, Bill, you're not going to have a clearing event here where it's winner-take-most, because these guys will continue to throw money at it for a long time. Among the other independent players, Elon and Grok have a great product and have executed incredibly well, and he has a unique—very, very unique—ability to raise global capital for a long time along with xAI. So you absolutely can't count him out of the race.

Maybe we just wrap it there. I mean, it's a period of consolidation for sure. You've got a lot of uncertainty in the world. You've got NVIDIA trading at 20 times earnings. You've got a lot of questions about these companies, but we're going to turn over a lot of cards, I think, in terms of these products over the course of the next 8, 10, 12 weeks.

We're going to turn over a lot of cards on tariffs, on the status of the reconciliation bill and taxes, et cetera. That's what makes it interesting. We could find ourselves off to the races come this summer, or I think we could find ourselves in the middle of a trade war and everything slowing down. So there's a lot at stake here over the next 90 days.

Let me make 3 final thoughts on my silly theory. One, I think one of the reasons this can happen also is the steep price decline we've seen on models and older models. I think it's very easy for a founder or a team to say, “Well, it's okay that I have negative gross margin or that I'm pricing here, because look what happens: in 6 months, the compute for this will be a lot cheaper.” I think that's a very rational decision too.

I think that also leads to what everyone seems to be seeing: when people go to production, they look for cheaper solutions. This is why I think open source thrives, why DeepSeek is working in the enterprise. I think everyone, when they go to production, optimizes.

Third, just in terms of this theory, if we're going to compare to the internet era, if there are high variable costs, I think you have to think about e-commerce more than you think about Google. There were a lot of businesses—this gets back to selling dollars for 85 cents—that were growing by achieving that. So I'm very curious how it all plays.

All that said, as you know, I'm a huge believer in AI. It's causing all kinds of disruption. It's going to create all new kings and queens in the venture market. I just like getting under the hood and seeing what's happening on the margin.

Well said. Well said. We're going to see a lot more in the weeks ahead. Good to see you. Good luck to Florida.

Appreciate that. Go Gators. Take it easy. Bye.