Letter #332: Alex Sacerdote and Leon Shaulov (2026)
Founder of Whale Rock and Founder of Maplelane Capital | Intelligence as Infrastructure: How AI is Rewiring the Economy
*KG Note
I am in San Francisco for the next week, then New York 6/15-6/17. If you are around and would like to try and grab a coffee/meal, go for a walk, or play tennis, please reach out (email; twitter).
Intro
More on this newsletter here.
Today’s letter is the transcript of a conversation with Alex Sacerdote and Leon Shaulav on how AI is rewiring the economy at the Sohn Conference.
Short Bios
Alex Sacerdote is the Founder of Whale Rock Capital.
Leon Shaulav is the Founder of Maplelane Capital.
Full Bio, Summary, and Related Resources below paywall
Transcript
Host: This conversation is going to be a really good one, a very timely one. The topic is “Intelligence as Infrastructure: How AI Is Rewiring the Economy,” and there have been a lot of thought pieces lately on this very subject. So, it will be good to get your perspective, what you are seeing, talking to the companies both inside the AI ecosystem as well as outside. Leon, when we spoke earlier, you said that this may not be an issue now. This is something that you are looking at being an issue for unemployment in the future in terms of AI working its way through the economy. How do you see this all playing out?
Leon: To approach AI and its impact on the economy is such a broad topic that I could be here for three hours. I am going to be looking at it from a point of inflation and what the causation is. I think long-term, this is a highly deflationary force. It is as simple as we are going to get a lot more for a lot less. Just take the healthcare industry as an example. It is 18% of GDP, 10% of an individual’s income. You are going to get a lot of it just through all these LLMs and things like that. It is really going to be disruptive.
Host: What is the switch? Because we haven’t really seen that yet. There are some companies, and there have been a lot of announcements lately, of companies that are laying people off and they say it is due to AI, but it hasn’t really manifested in the broader economy yet.
Leon: There is a time lag. In the short term, you have to be a little careful with the deflationary call, because in the short term, you can actually end up in an inflationary move. If you look at it, the pricing of CPUs, memory, and just the infrastructure alone, it is skyrocketing. That is an input cost. On the labor market, the labor market is actually quite robust. If you look at software engineers, you would think this is the one area that would be highly disruptive. You would just fire all these guys. That is not what is happening. Last month, there was an 18% increase in software engineer hiring.
I have been thinking about that more and more. Why is that? I do not think it is happening on the tech side, because they are actually laying off or being much more prudent about it. I think it is happening in the old economy. As everyone is trying to put these LLMs in place, and learn how to use them and how to implement them, they have to have someone help them. This is causing some of the hiring among these software engineers. Product developers are highly in demand.
So, right now, the labor market and a lot of your old economy businesses are slower to fire and be disruptive about it. So, there is a time lag first to get inflation; the labor market is pretty strong. Then, each one of these models gets a tremendous advancement every three months. As you go, even from a software engineer standpoint, as prompt engineering advances and you can just prompt these models to tell them exactly the task you want to do, there may not be so much need for these guys. So, first you get sticky inflation, and long-term this is a highly, highly deflationary move. If you put robotics on top of it, the labor market can look very different a few years from now, which will make the Fed’s job challenging. I do not envy being these guys over the next few years.
Host: Yeah, because the tools at their disposal could be pretty limited. Alex, do you agree with this timeline?
Alex: I have to say, I think Leon has really come up with the answer here. There are super smart people on both sides that are saying it is going to destroy the job market, and others are saying it is going to be a huge boom. But I think it is dead on, and for the first time we at Whale Rock, we want to be hiring. We are looking to hire Claude ninjas, and we know we need help to build these amazing things. So, you need to do a little bit of hiring before. Coding is the one area where it literally replaces labor, but then that is allowing people to build software where they were never going to build it before. So, I think it is still a very hard question what it does to jobs. It is going to be incredibly powerfully productive, but it takes time.
Really, we are in the first batter’s box of AI. We all have been using AI, but it is just AI 1.0. It is a search engine on steroids. But now we see what business AI is going to be, and it is Claude Code or something like that plugged into all your data sources. Then you can build skills on it, and then you can build agents that actually go out and do things. There is just a tiny, tiny percentage of the white-collar population that is using AI in that very advanced way, maybe 10 basis points of the one billion white-collar workers. So, of course, we haven’t seen any major productivity gains. But if you look carefully at what these people are doing, it is astonishing and astounding.
If you think of where we are in this whole AI story, that 10 basis points of people, Claude Code has 14 million daily active users (DAUs), only 14 million. These are the people who are really using it for business every day, but they are not the 10 basis points. Those 10 basis points are burning a thousand times as much compute and tokens as the rest of the people. So, you are going to see that 14 million DAUs go to 500 million DAUs, and then you are going to see the portion of people that really use AI with 14 agents running things increase. That is all happening straight up. At Whale Rock, we talk about S-curve adoption. This is an L-curve, straight up, and it is just starting now. All this CapEx that we have been putting in place, and the reason these chip stocks are going up, we have half of what we need from a compute standpoint right now before it has even started. That is how it is going to play out from that perspective.
Host: So, how do you think about investing in an L-curve?
Alex: Nobody has ever seen anything like this, ever, in our entire careers. We were there for Internet 1.0. My first stock at Fidelity was Amazon. I remember at the time in 1998, there were only 100 million internet users and only 2 million e-commerce users. I said, “It doesn’t even need to grow for this stock to be a buy.” This one is moving faster. The revenue growth that we are seeing at Anthropic, going from 100 million to a billion to 9 billion, and then it is already at 45 billion. It is going to be maybe 100 billion.
Host: That’s run rate?
Alex: Run rate. Last month annualized. So, it is not for the full year, but it is growing so fast, it is a good metric. 10x-ing at major, major scale. No one has ever seen anything like that. We think the foundational model layer, AI is a stack with the chips at the bottom, the clouds in the middle, the foundational companies above that, then the applications on top. I think the two places that capture the most value in AI are the foundational models, where we own Google, OpenAI, and Anthropic, and still at the chip layer. Because, as I said, we are in a dramatic undersupply of chips, and there is dramatic growth ahead, but also it is the golden age of hardware, where there is now so much innovation.’
[Random interruption]
Host: I think the models heard you and wanted to also participate in the conversation.
Alex: What did they say? I couldn’t hear them.
Host: I don’t know. I think they liked your thesis of the oligopoly of LLMs.
Leon: One of the Claude agents.
Host: Yeah, exactly. What do you call them? A Claude ninja?
Alex: A Claude ninja.
Leon: I think what is so powerful here is that between Anthropic and OpenAI towards year-end, you are going to be looking at $200 billion of revenue. You can break it down any way you want to. You can make a guess. What is more interesting about it is the margin profile of these companies. Because they have been able to lock up compute, and they were some of the first, and they have it already locked in for the next several years, this is going to be enormous incremental margin. On a fixed cost basis, the pricing per token, everything is rising.
There was this huge debate a year ago, even two years ago, even six months ago, about where all this CapEx was going, what the ROI was, what it would all look like, and if they were just wasting money. You are going to look at profitability at Anthropic that is staggering. So, you could be looking at something that is like 18 times earnings. That argument would be put to bed, and the L-curve of the adoption is so enormous that it is – I have traded several tech booms since 1998, that is when I started. There has never been anything like this.
Host: So then, what do you make of some of the more legacy tech industries? A lot of people have been describing semiconductor moves as being parabolic, is the term that I keep hearing people say. The sector is down today greater than the market, but is that something that you think is the best way to play perhaps in the public markets right now? And how do you decipher within chips?
Leon: Look, it has definitely gone up a lot. As I said to you the other day, this was much easier a month ago. But in some of the stocks, you probably have certain things that have gone 100 miles an hour in a 60-mile-an-hour zone, so there is going to be some accidents and someone is going to get pulled over. But most of it is just on an incredible trajectory. AI demand is driving so much compute demand. First, we started with GPUs, then we went to memory. Now it is CPUs, networking chips. That is just creating massive supply constraint.
If you think about the semiconductor industry, maybe just take the last decade, they have gone through so many boom and busts that most of these companies have gotten pretty disciplined about CapEx. They just haven’t spent. Since the last foundry and memory down cycle, no one spent, and it was really one big spender, and that is Taiwan Semi. And you can even break down their spend; you can make a very good argument that they have underspent significantly. And you can use metrics like profitability over CapEx, or revenue growth rate, acceleration, over CapEx. They’re all very anemic. You can actually see it now that you have seen some of these announcements from Intel and Samsung and some of the lower-end stuff as far as the foundry competition that maybe Taiwan Semi made a mistake. They will have to rectify that.
Now you look at the industry today, and it is not just Taiwan Semi. You have memory companies like Hynix, Micron, SanDisk. What is the last time we talked about a NAND cycle? It must be a decade ago. It is a powerful cycle. The profitability of these businesses is absolutely enormous. If you look at forward CapEx indicators, historically how profitable the customers are leads to forward CapEx. Intel is now – this is a company that was dead for years – and its foundry business is starting to pick up customers. So, you have to look at this landscape and you say you have gone from one spender, and by the way, in that environment, they probably had all the power in negotiating with semi equipment companies. You are going to multiple spenders, all of which underspent. The forward metrics suggest they were going to have to spend a lot.
I actually think people think it is like $120-130 billion of WFE (wafer fab equipment). I think you are going to reach a $300 billion mark over the next three to four years. Semi equipment companies, all of the customers now have 70 to 80% margins. The memory guys are 80%, Taiwan Semi is approaching 70%, semi equipment is at 50%. So, you have pricing power on top of it. These stocks may not screen as the cheapest things in the world right now, but I think the estimates are like 50 to 70% too low.
These businesses will also look less cyclical because a lot of the customers can now sign long-term agreements. Memory guys are signing LTAs right and left. So, you now have much more visibility on the kind of CapEx you can put forth over the next three or four years. So, they will do it. It is just not as cyclical to them, so they are going to do it. That could be reflected in the semi equipment multiples also. Pristine balance sheets, they can do M&A, they can buy stock. I do not know what the next 10% is, especially when you have had this kind of a move, but I am guessing the next 50 to 100% is up.
Host: Alex, you look like not quite buying.
Alex: No, no, I fully agree. He articulated a lot of great points. In addition to AI just being the most compute-intensive thing we have ever seen and shortages as far as the eye can see, we are in a golden age of hardware. For the last 40 years, hardware hasn’t changed. It has been an x86 server that costs $2,000. 20 or 30 companies can make it. Every little part in that server has been commoditized: the networking, the PCB, the heating system, the cooling system. Compute basically grew 30% compute demand, the bits. That is good, but Moore’s Law is going 30%. So, there was really no growth. Everything was commoditized.
All of a sudden, AI hits. Elon calls it a supersonic tsunami, and it really is because it is 10x-ing every year with no end in sight. That is putting pressure on old compute, which cannot do these things. So, you have to innovate at every single layer of these $300,000 massive server racks that are now highly complex machinery. A printed circuit board, which used to be a total commodity, now there are only two or three companies that can do this properly, and you have to upgrade every year. For example, the networking speeds. It used to be 1 gig, and then seven years later, you upgrade to 10 gig. Now, you are on 400 gig, next year it is 800 gig, next year it is 1.6 terabit, the next year it is 3.2 terabit.
So, the people that are selling into that, there are only a few of them who can do that, that are innovating hand in glove with Google and Nvidia. So, there is less competition, there is higher margin, there are higher ASPs every year, and you have tremendous visibility. All these companies that nobody used to ever pay attention to are now golden, wonderful businesses. That earnings algorithm is units growing 50% (that is the end user racks), your ASPs growing 20 to 100%, your gross margins rising 300, 400, 500 basis points, and your visibility is three or four years out. You are growing earnings 100% for the next four years, not to mention we are in short supply of everything you are making for the next three or four years.
I have never seen anything like it, and so the moves that we are seeing are justified. It is going to be bouncy, but AI is a compute problem first and foremost, or it is also a model problem, but I think it is a phenomenal way to catch it, and the multiples haven’t caught up. It has all been earnings. In some cases, we have seen multiple expansion, but if you do a next three or four year kind of earnings analysis, it is really powerful.
Host: There is no concern that AI gets more efficient and the computing problem goes away?
Alex: I think there are always going to be innovations. But in general, the tokens are growing. Tokens are the unit of compute in AI. They are growing 14x every year, and the chips basically get better 100%, maybe 200%. So, you are growing your CapEx, and you are adding that to the base. Your computing estate can maybe grow two or three times as efficient with these innovations, but your token demand is 12x-ing. So, maybe you will get some efficiencies that can push up beyond that, but it is still – it is not going to be able to keep up.
Host: What does this all mean for software? Software sold off in the first few months of the year; it has rebounded about 20%. IGV, the ETF. Over the last month, it feels like an inflection point where people are trying to figure out if this is a zero-sum game as it pertains to AI versus software, but I am curious about your perspective.
Leon: I think it is too broad to say “software.” There are different verticals within software. If you are a horizontal application layer, there are some troubles there. There are a lot of troubles there. But I think if you are a data-driven business or infrastructure software, you can succeed. And you can be very successful. We can debate some of the multiples that are being paid in the market for winners versus losers, but now I look at a company like Datadog, that is a unique asset. We can again debate if it is 30 times or 40 times; the market will get that, but they are at the heart of actually benefiting from everything that is happening.
So, I think just saying “all of software,” that happened in January, February. You just had anything that had the word software attached to it. Now we have seen a lot of separation over the last month and a half. People start to go through the rubble and figure out which ones are which.
Alex: I think that the decline is largely justified. Basically, the old way of doing code is pen and paper or horse and buggy. The new way of code, it is not a car, it is not a jet engine, it is the transporter from Star Trek. It is such a massive change in how software is getting sold. The good news for software owners is that software tends to be very sticky. So, it probably will take time. Nobody wants to rip out their existing system, but in the back of your mind, you are thinking, “In one, two, three, or four years, could that really change?” And maybe it could.
In the near term, they have a problem in that software used to be at the top of the CIO’s list. Now AI is at the top. Everyone is spending all this money on tokens, and so that is taking budget away from software. The software companies themselves, we thought they would be able to build great AI applications and then sell them and get money for that, but that has been a fail so far. Maybe it is just a matter of time, but maybe it is a culture thing. They do not have the right people. It is very hard. It is a different sales process because you are selling a service, not software, and it is a different business model. So, I do not think software is going to be bouncing anytime soon, but we are watching it really carefully because we might see a few software companies actually develop and benefit from AI. Leon mentioned Datadog. A lot of the big model companies like Anthropic are using Datadog’s tools. So, that is also a pretty good tell.
Host: In the remaining time, let’s talk stocks. What do you think are the best ways to play this - what did you call it - a supersonic...
Alex: Tsunami.
Host: Tsunami.
Alex: I will just start with two: a small one that you haven’t heard of, and a big one that is easy to buy or think about.
The first one is TTMI. They make printed circuit boards, which used to be the biggest commodity of all time. But as these AI chips and servers are growing, demanding more power, needing more signal integrity, running much faster and hotter, they need more and more printed circuit boards. There is a tremendous unit growth story, and then the printed circuit boards themselves are getting much more complicated. They used to just have 10 layers, and now they are going to 20, 30, 40, even 120. That is causing ASPs to rise, and there are very few companies that can do these highly complex printed circuit boards. TTMI is one of them, and they make them for Google, Nvidia. They just won Nvidia, and they also do them for other AI companies. Then they have 40% of their business which is defense. There is a huge up cycle in defense. They have won business with the Iron Dome contract, and you know how defense is getting so electronicized.
The second one is just Google. It is simple. They have won AI. They are the only public company with a foundational model. Their Google TPU chips are phenomenal. They are now powering Anthropic, and other people are using them besides Google. Search is actually getting accelerated, and they have got so many other assets like YouTube and Gmail and Google Sheets. They are going to infuse AI. The stock is very cheap, and we are going to see revenues accelerate at Google. So, it could easily be up 50%. I do not see very much downside.
Leon: I actually like analog, I like semi equipment. I brought it up earlier. Lam Research happens to be my favorite because I just think they have such high exposure to memory, and that is where the market is still skeptical, and there has been a ton of lack of spending. I think this is more of a middle of 2027-2028 story where I think there is just going to be a boom in spending, and these guys, I think the Street estimates may be like 50 to 70% too low. I think they are going to do $55 billion of revenue, and margins will go significantly higher. That is kind of one of my favorites, and I actually think the analog semiconductor sector is quite interesting.
There is a decent chance that this could look like memory from a pricing standpoint, of how tight things are. I think someone pretty smart pitched Infineon earlier at the conference. I like that one. I think Texas Instruments is very good. I think Renaissance in Asia is quite interesting. I think it is just going to remain really tight for a while, and I think if you find ones with the AI power angle attached to them, you are going to have significant upside.
Alex: So, my mother is in the audience. I think Leon’s mother is in the audience, and I bought the SMH, the semiconductor index, for my mom a while back, and she is going to keep holding it after what Leon said.
Host: See, that is what good sons do. They buy their moms ETFs. Happy Mother’s Day, by the way. All right, thank you guys so much. Really appreciate it. Thank you.
Summary, Full Bios, and Related Resources below paywall
