走进 Coatue：700亿美元对冲基金的 AI 与零售策略

The best companies we are seeing today that are going AI-native are winning. Before I worked at Coatue, I was working at Melvin Capital. Many of you guys have probably heard of Melvin as the hedge fund that was short GameStop.

We went from probably the best-performing hedge fund in the world to basically down 50% in 2 weeks, and the reason was we didn’t realize how powerful retail could be when they focus all their energy on a single stock. The way to source ideas now and come up with new stocks to invest in, a lot of that is coming from the internet. If you go on WallStreetBets, people are posting real work there, and now there’s just been this kind of proliferation of information.

There was a company called AppLovin. I had the CEO, Adam Forgie, come to our office. I messaged my boss at the time, and I said, “Hey, you have to get in here right now and meet this guy.” And he’s like, “I’m busy.” And I’m like, “Trust me.”

The first use case of AI truly is driving these advertising businesses to grow faster than you would have thought. Any job that exists in the U.S. where you work at a computer, at some point, can be automated, including my job. So I think as that starts to play out, there’s going to be a lot of revenue opportunities.

Michael, welcome to Sorcery.

Thanks for having me.

We have so much to cover today, but to start, let’s talk about you. Who are you? What do you do?

Well, I’m from Cincinnati, Ohio, and I live in New York now. I work at Coatue, which is an asset manager. We do both public and private investments, and my main focus is on public equities. Now we have a retail product that we’re also working on.

Sorcery has had a lot of fun with tech over the last year or so. One of the most fun podcasts we did in the last couple of weeks was with Keith Rabois, and this was when he had just rejoined Opendoor as board chair. The funny thing about that was specifically the cult sentiment behind it. How has the public market evolved?

1. Retail Investors Change Markets

If you look back maybe 6 or 7 years ago, the idea of retail investors was not a thing, right? What I love about the public markets is that anyone can invest in them. I would debate that’s actually how I got started. I used to debate stocks with my grandfather, and he worked in the plumbing industry, so he wasn’t a professional stock picker, but he loved investing.

He would invest in companies that he thought were long-term compounders. He loved Warren Buffett and the idea of value investing. Fast-forward to a few years ago, with companies like Robinhood and retail trading, and then just the internet broadly: more and more people have gotten into investing, and the impacts on the market have been huge.

Before I worked at Coatue, I was working at Melvin Capital. Many of you guys have probably heard of Melvin as the hedge fund that was short GameStop. I lived through this period where we went from, at the time, probably the best-performing hedge fund in the world from a return perspective—a single-manager, long-short equity fund—to basically down 50% in 2 weeks.

The reason was we were betting against GameStop, and we didn’t realize how powerful retail could be when they focus all their energy on a single stock. You’ve seen that same excitement with Opendoor. They’ve got a great team, and there’s been a lot of excitement around what they could do, and the stock was up 700% or something on that excitement.

The market dynamics have very much evolved, and it has created both new opportunities and new risks. On the risk side, the idea of a GameStop going up as much as it did because of the internet and Reddit and people getting excited was not something that existed up until that point. Any time when people were short a stock, there were squeezes, but it was always catalyzed by something.

Volkswagen and Porsche was a potential acquisition. This was just a lot of guys and girls on the internet deciding they were going to buy it, and it went up, and people had to cover. It completely changed investing and the risks that people think about.

How did that change your role in terms of what kind of data and information you pull from?

2. The Internet Sources New Ideas

I think one of the best parts about the public markets is that because anyone can invest in them, ideas can come from anywhere. What you’ve seen over the last few years is the emergence of all these different channels of information.

If you think about investing in the public markets 15 years ago, you would get quarterly earnings reports, 8-Ks, annual earnings reports, 10-Ks, and management would speak. But that was kind of it. Other than that, you were reading The Wall Street Journal and The New York Times, and that has completely evolved.

Now a lot of people, including all the retail investors, have opinions on stocks and are doing interesting analysis. If you go on WallStreetBets, people are posting real work there. Now there’s just been this kind of proliferation of information, with people like you having amazing guests on the podcast offering interesting insights.

We are tracking a lot of different data today. We look at how often stocks are mentioned on Reddit, we look at Twitter, and we look at how things are trending on the internet all the time—on Reddit and all these other places. The way to source ideas now and come up with new stocks to invest in or new analyses to do, a lot of that is coming from the internet now. That’s the world we live in.

In terms of Coatue’s fund, how big is the fund, and what’s the main portfolio that you cover?

Coatue as a whole is around $60 billion of assets under management. In public equities, we have around $25 billion, and then we have a private-markets business and a credit business, too. I focus almost all my time on public equities.

The nice part about doing both is that I also follow OpenAI and Anthropic and am very in tune with what’s going on in the private markets. One reason is that a lot of those developments are impacting public stocks, especially today. But also, when our private-markets team is looking at a private investment, there are often interesting insights from the public markets.

My knowledge of how digital advertising works might impact a business or how they think about it. Mainly, though, I focus on TMT investing in the public markets—trying to find stocks that are going to go up, and then trying to find stocks that are going to go down.

It’s internet, China internet, and cloud. We have a pretty tight-knit team, so we all work together—the core group of us.

Any particular names? I know Jack Griffin—thank you to Jack for the intro—but I know he mentioned that you found AppLovin for them.

AppLovin, yes.

Yeah.

3. AppLovin Shows How Ideas Start

It’s a pretty crazy story, and it goes into how you find ideas. What ended up happening was there was a company called AppLovin. I think at the time it was around a $20 billion market-cap company, and the name is amazing, right? AppLovin sounds like—it’s almost like a meme name to begin with.

This business does mobile gaming ads. Whenever you’re playing Candy Crush or one of these games—the best way to describe it is when you walk on an airplane and see everyone looking at and playing Solitaire or various games—they’re the guys who serve the ads in those games.

I had never heard of the company. I didn’t know what they did. A buddy of mine called me and was like, “Hey, you should take a look at this thing. It’s pretty small, but something’s happening here. It’s starting to grow really fast.”

I had the CEO, Adam Foroughi, come to our office, and I met him. I literally knew nothing about this company at that point besides that they did mobile games. I met this guy, and I will never forget this moment. I messaged my boss at the time and said, “Hey, you have to get in here right now and meet this guy.”

He was like, “You know, I’m busy.” And I’m like, “Trust me.” Within 5 minutes of meeting Adam, you knew that there was something really special here. This guy was the most locked-in person I have ever met.

After I walked out of that meeting, I was like, “Okay, we need to figure this out.” What ended up happening was that a lot of what we were seeing in the digital-ad market at the time was basically a pure play happening with AppLovin.

The idea is that AI is this big thing, and one of the places where we’re seeing revenues actually happen is at digital-advertising companies. What’s happened over time, if you think about Facebook, is that their goal is to serve you the right ad at the right time.

All of the AI learnings from LLMs and everything that we’ve seen over the past couple of years are directly impacting their ability to serve those ads better. When I first joined Coatue, I remember one of the first things I had to do was explain why Facebook could probably grow 10% or more.

They were going through this period where, with IDFA and Apple, they lost their ability to track. So there were questions around whether they could really grow above 10%.

Well, fast-forward 2 years: they're growing mid- to high-20s right now, right? And so that was an impossible thing to imagine at the time. But what happened was the underlying ad engines got better with AI.

And so you kind of knew that when you had met Adam Foroughi. The way he was talking about what they were doing—and that they had basically used GPUs on their advertising business—they were growing, I think, 15% before, and all of a sudden the ad business was growing 15%, 50%, 70%. The stock at the time was a $20 billion market-cap company.

And I remember I was like, “Okay, so if you kind of believe this to be true, and if you just listen to him and believed what he was telling you, and you put that in a model”—one of the things we do is make discounted cash flow analyses to try to see what a company's worth—you literally could not make the discounted cash flow analysis, in your worst-case scenario, be less than a 3x. It was the most remarkable thing I've ever seen. And so then we got to know him better and developed a really close relationship with him.

Sorcery is brought to you by Brex, the financial stack trusted by more than 30,000 companies, including one in three venture-backed startups in the US. Nearly 40% of startups fail because they run out of cash. Brex is literally built to help founders avoid that. Unlike traditional banks that let your money sit idle, chipping away at it with fees, Brex is designed to help you spend smarter and move faster. Their all-in-one solution combines checking, treasury, and FDIC protection into one powerful account. You can send and receive money globally at lightning speeds, get 20 times the standard FDIC coverage through their partner banks, and even high yield from day one. With same-day and even same-hour liquidity, access your funds anytime. Companies like Scale AI, DoorDash, ServiceTitan, HIMSS, Anthropic, Flexport, Robinhood, and Plaid trust and use Brex. Start today at brex.com/sorcery. That's B-R-E-X.com/sorcery.

Tracking these companies over time, especially one in the ad space, what does that also tell you about how AI is being implemented within the companies? Because ads are, as some would say, going to be one of the most disrupted areas. So how do you see that being affected with AI?

4. Advertising Powers AI Revenue

That's interesting. Yeah. Broadly, I think we're at this point in the market where there's been a lot of committed spending to build out this AI infrastructure, right? You're seeing new announcements every day: OpenAI doing a deal with NVIDIA or doing a deal with AMD, or there's some new data center build, and there's all this money going into it. In order to run these GPUs and do whatever the use cases are, you first need to build the infrastructure to do that.

We're at this weird point where the market's a little unnerved, worried that we're in a bubble. We're spending all this money—hundreds of billions of dollars—and yet, where's the revenue? ChatGPT charges, what, $100 a month or whatever it is? That's not enough to offset that.

But I think when you actually dig deeper, what we've kind of come to the conclusion is there's actually a lot of AI revenue happening today. Now, it's not enough. You have to believe there's going to be more. But a big part of that is advertising.

So the first use case of AI, truly, is driving these advertising businesses to grow faster than you would have thought. If you think about 2 years ago, what did I think Meta's revenues were going to be? Maybe I thought they would grow 15%. Well, they're growing 25%. So that kind of incremental revenue growth—that's AI. It's not generative AI, but that is GPUs accelerating machine learning to find and serve you an ad for a snowboard that you might not otherwise have seen.

We're seeing that happen, and it's happening across the board. AppLovin is growing really fast because of this. Meta's growing, call it, high-20s percent. Google Search, which people thought was going to grow sub-10%, is now growing mid-teens. So that's kind of the first impact.

What's also happening is the recommendation engines of all these companies are getting a lot better. You may have noticed this, but when you're on Instagram, the Reels they're serving you are more addicting. If you look at Instagram time spent, basically it was flat from, I don't know, maybe 18 months ago to 6 months ago. Now, you would spend 40 minutes a day, and that's gone up 15% in just the past 6 months because they basically took those GPUs and put them at the recommendation engine, and now people are spending more time, right? And that's also more ad dollars.

So advertising's a huge place. I think the big question—and it's what you're getting to—is what happens in agentic commerce?

Mm-hmm.

And what happens to these ad models when you have a shopping agent doing everything for you? My view is that it's early. It's very early. We had OpenAI announce the Shopify—

Yeah.

—and Etsy integration a couple of weeks ago. The product today is not at a place to really be that useful, but you can kind of see where it's going.

I think that from a shopping perspective, we are going to be in a world where the old world was, “I want to buy something, I go on Google and type in red shoes to go skateboarding in,” and it would come up with a list of results. The next step is going to be I go to Gemini or ChatGPT and say the same thing, but it knows a lot more about me, and it will suggest better products. In that world, there's probably no advertising.

And then the ultimate end, which I think is the most exciting, is a lot of the products that you end up buying—think about Instagram, right? More and more, I'm getting advertisements for things I never even knew I wanted, and then I click the button, and it shows up.

Tobi at Shopify made this comment, and I actually agree with it: those were actually not impulse purchases. I secretly wanted those things, but no one had ever shown them to me, right? I would never have gone on Google Search and looked up that interesting steak knife, but when it was shown to me, I bought it.

Now, imagine a world where you're in ChatGPT or you're in Gemini, and instead of asking it for something, it's just telling you, “Hey, you're going on this trip. I think you need a new ski coat.” Or, “Hey, I just found this interesting product because of a conversation that you and I were having in a separate chat.”

I think that's going to drive consumer spending for these goods a lot higher. And I think when you think about advertising, there might not be an ad per se, but from the merchant's perspective, instead of spending 20% of my revenue on marketing in the form of ads, and 2% of that is going to Shopify, that split probably changes.

That 20% that I'm spending on marketing to maybe Meta is now going to— that profit pool is going to be more directed toward Shopify or the actual agent players themselves, OpenAI or Gemini. But this is early inning. People are debating this literally every day.

This one hasn't come out yet, but I had Alfred Lin on. I think this will come out before it. He spoke, and Reid Hoffman also spoke at Forerunner's AI conference back in, like, the summertime, that Kirsten Green throws.

One interesting thing that Alfred said was, “Things that happened 3 months ago are not relevant today. Things that are happening today are not going to be relevant in 3 months.” Things are moving so fast, it's really hard to predict, but you have to be active. You have to be watching what's going on and gathering as many data points as possible to adjust accordingly.

And then another thing—and I'm curious about your perspective on this—was Reid Hoffman talking about how business models define different generations of technology. Advertising was the majority of the last one. We don't really know what the AI business model is yet. Do you have any idea?

5. Practitioners Reveal The Next Wave

On your first point about things changing, that could not be more true.

This has been the longest year of my life. I feel like we're at this point in tech where the implications of AI are going to be big.

Mm.

It's not a question of how big they're going to be. It's what is actually going to be impacted, who's winning from this, and who's losing from this.

Mm.

And that changes literally every day. A big part of my job is that I'm not making an investment, closing my eyes, and waking up in 5 years, right? Stocks are priced every single day.

I'll give you a great example: OpenAI had its DevDay a couple of days ago.

Yeah. Oh, my God.

This was so insane. They get up, and if you got named in the presentation...

By the way, you could argue that a lot of these companies that are named in the presentation to go and be part of this agent layer—it might not actually be a good thing. If everyone's using ChatGPT, and now you've just got an additional layer of maybe disintermediation...

It might not actually be good, but if your name got mentioned, bang, you're up 5. And the best was Mattel, the toy company, right? Not even a tech company. They got mentioned in this thing, and Mattel, the toy company, stock went up 6% in a second.

That just highlights where we are, because we don't know how this all plays out, and everyone's trying to figure it out. Any sign of you being an AI winner or AI loser gets priced into the stock very, very fast. Part of our job is to stay at the forefront of what's happening, figure out the implications in real time, and do analysis around that.

The way we do that—and I think it's a unique thing with Coatue, but I actually think it's underappreciated as the most important part of tech investing—is that the best way to figure out what's going to happen in tech is to actually talk to the practitioners of that tech. What I mean by that is we spend a lot of time not only talking to the company CEOs and the management of the actual public companies, and having relationships with the management of the public companies, but we talk to the private companies.

We talk to OpenAI and Anthropic. We talk to the researchers, because these are the people every day living and breathing this sort of tech, this AI that's going to change a lot of things, and they all have super interesting insights. But if you don't do that and you're just sitting at your computer trying to build a model or forecast the next 5 years or the next quarter, you miss these big waves.

Part of what makes Coatue really successful over the last 20 years and today is that, by being at the front of tech, tech goes out in different forms, whether it was the web, the internet, Web1, or Web2. There have been winners and losers and new markets in all these tech waves. We think the big one right now is AI, and that's not a hot take, but we think it's actually bigger than any of these previous waves.

We spend a lot of time focused on meeting the people in the industry, because that's where you're going to get these insights. They'll tell you what they think. The way Adam Foroughi at AppLovin was telling you, “Hey, this revenue's going to grow a lot faster than people think.” Often, those are the best tidbits of information to get, because these are the people doing this every single day.

The last thing I'll say on this is that when you have these big tech waves, every single time, when things are inflecting positively—think about when people got excited about AI and realized that, in order to do AI, you needed NVIDIA GPUs—even the most bullish person in the world about how big AI could be probably underpredicted the amount of GPUs you needed.

The same is true on the inverse side. When companies are getting disrupted, that pace of disruption normally happens faster than you think. If you can find those big trends and the winners within those trends, you can do all the modeling in the world and the valuation work and the DCFs and the analysis, but normally it ends up being better than you think to the upside and worse than you think to the downside.

A question I'm interested in is, in the early innings of this AI cycle—maybe in the last year or so—none of this really existed. I feel like now it's actually taking hold, and there are actual applications. But in the beginning, it was a lot of marketing hype, and I'm curious how you deduce who is real and who is not, and how you determine whether or not they're actually making progress or it's just a consulting presentation.

6. Real AI Use Cases Emerge

Totally. What happened was, you got to the point where, if you as a public company came out and didn't say how you were going to benefit from AI, no matter what industry you were in, people instantly were like, “Oh, they're behind.”

Yeah.

So you saw a lot of companies basically talk about AI before it was actually being implemented. I actually think you see the same thing even in the hedge fund industry today. But I think where we are in this cycle is that we're now actually starting to see revenues from these companies.

A year ago, you could point to—I remember there was a big AI scare in the summer of 2024, where there was all this build-out happening, and there was a moment where everyone kind of woke up and was like, “Okay, so we've got ChatGPT. What else do we have?” Literally, it became public discourse around the investing world: you couldn't really point to anything else.

These stocks were tanking. The power and utilities companies, the infrastructure companies, a lot of the tech companies—these stocks went down. Some of them were going down 15% to 20% in the course of 3 weeks. It's a disaster if you own these companies.

Nothing had changed. There were just nerves in the market, and you didn't have a lot of things to point to and say, “No, revenue's coming from there, and revenue's coming from there. It all makes sense.”

I remember in that moment, we had this really amazing conversation with one of the head engineers at xAI, and he said, “Guys, here's how I think about it. The tech today—so this was summer of 2024—the models today, where we are today, and where we are 14 months from now, is a very different spot.”

Where we were in July of 2024, the models were good enough to have a lot of applications that would generate revenue. The way he talked about it, he was like, “Each model is like a child,” right? With each breakthrough in the model architecture, or by training on more GPUs, the IQ level of that child goes up.

I remember at the time he said, “Today, I think the IQ of the child is about 100.” An IQ of 100 in this economy—there's a lot of work for a person with an IQ of 100 to do. But remember, it's a child, and the child can't work right away. The child has to grow up and figure out how to be used.

His point was that the tech was good enough, but we then needed to spend the time developing applications for that tech. Fast-forward a year, and you've seen that. There have been early breakouts of use cases.

The first one is coding. That's the big one, right? You have these companies that are generating a lot of revenue today: Cursor, Windsurf before it was acquired, and Cognition. There are companies generating a lot of revenue, and that's the first use case: agentic coding.

My view is that the reason that's the first use case is that a lot of the guys working on building AI in these labs code. What's the first thing they're going to try to figure out? It's going to be how to make their jobs better.

That's now starting to broaden out to a lot of other industries. We're now starting to see companies begin to generate revenue and products that actually look pretty good for a lot of other things, whether it's building an Excel model, going after the financial services space, or call centers.

My view broadly—and this is my view; I'm not sure this is everyone's view at Coatue—is that any job that exists in the US where you work at a computer at some point will likely be automated, including my job. As that starts to play out, I think there's going to be a lot of revenue opportunities.

One thing that we talked about before was positive and negative indicators of whether companies are not hiring anymore, whether they're doing layoffs, and whether AI is going to be automating more jobs. How do you view job automation when picking companies and betting on them?

7. AI Changes Work And Investing

There are 2 camps here. There's the camp that says AI will make workers 10 times more efficient, and therefore, you probably actually want to hire more workers because industries are competitive, right? If your workers are 10 times more efficient, you're going to hire more workers than your competitor, because then you're going to be able to do more things.

The other camp—and this is the camp I'm in—says you're probably going to get more efficiency, and it's orders of magnitude higher than 10 times. If you think about me, my dream with AI is that instead of having a few analysts work for me, I have 20 agent analysts doing the same job, and those 2 guys who work for me also have 20 or 30 agents working around the clock. I want to see that world happen.

In a lot of industries, what you're starting to see is that people aren't getting fired today or their jobs being automated today. Hiring is slowing. You've seen these charts of the college-grad software-developer chart.

If you think about the first obvious use case in the market of an AI application people are using for work, it's software engineering. My view is that this is a little bit of the tell for how this plays out.

So hiring slows, headcount growth slows. For the stock market, that's good because if you think about—let's take the Magnificent Seven, right? A company like Amazon or a company like Meta: if they stop growing headcount, these are companies that grew revenues 20% for years, and headcount grew in line with that. If they stop growing headcount and just make it flat, the margins are going to increase, profit and earnings-per-share growth are going to accelerate, and the stock is going to go up a lot. So the market will view that positively.

I think that's true across all industries. I think where it gets tricky, and the big question that people are asking, is: if jobs get replaced across these different industries, what are they going to do? What new industries emerge that they can work in? There are debates around that. And if not, Amazon's stock price might have gone up a lot, but who's going to buy the goods if there's unemployment?

I think people are still trying to figure out those debates. I'm pretty optimistic that this will happen a little slower than some of the fearmongers think. But there will have to be new industries for a lot of people to work in, or other ways of making money. I think that's one of the reasons—and you've seen this—that it's important for the average American to be investing in the stock market now, right? AI is going to benefit all these companies, and the stock market's going to go up. We might be at the front of a multiyear, amazing run in the stock market because these companies' revenues are going to grow faster. Their costs are not going to be as much as you would have thought. Your margins are going to go up, and the market's going to go up. I think it's the most exciting time to be investing in the public markets for that reason.

In today's high-speed business world, staying ahead means using the smartest tools possible, including the powerful capabilities of artificial intelligence. Meet Turing Intelligence. Turing builds customizable AI systems designed to solve your mission-critical challenges, no matter your industry. From expert guidance to tailored projects, Turing helps top companies realize AI that's more capable, more adaptable, and more effective. With Turing, discover how AI can accelerate your business growth. To learn more, visit turing.com/sorcery, spelt S-O-U-R-C-E-R-Y. That's turing.com/sorcery.

I covered Klarna's IPO, and Sebastian was really clear about when they were doing their turnaround, reducing headcount and freezing hiring. Now they're just waiting on attrition. And same with Opendoor. Well, Opendoor has to go through a lot more right now—

Yeah.

To resurrect themselves and do a turnaround. They're counting on severely slashing headcount and hopefully deploying more AI agents within. I'm curious: between all of this, how do you set long-term and short-term price targets in this environment?

Yeah, it's a great question. One thing on your previous point: I think the most exciting thing is that these companies are all going to be revolutionized with AI. It's not just the revenue growth—the actual workings of the companies. The best companies we're seeing today that are going AI-native are winning.

AppLovin's a great example. Adam Foroughi—I think they have the highest EBITDA per head of any company in the world, and he loves this statistic. His view is, first, everyone at the company needs to figure out how to use AI now. If you're not, you're fired. But I'm setting up the company not for what the technology is today, but for where the technology is going. Because it's changing so fast, I want to be ready in 2 years, when the models are significantly better and the applications are significantly better and can automate these different parts of the role or the workforce.

You're seeing other companies, maybe less aggressively, have this mantra, and those are the ones that are going to win. A big theme of Coatue is that we take a lot of the learnings we see from public and private companies and how they're implementing technology, and we do it ourselves.

Philippe and Thomas were early investors in the cloud transition, right? Coatue became cloud-native. All of a sudden, you had the best companies talking about how they were able to use all this data, get it into one place and do data science on it. So we built out this data science platform over the course of those years, and it's been amazing. Thinking about how this technology will transform Coatue gets accelerated massively in the AI world.

My belief, and something I spend a lot of my time on, is: how do we use AI internally, and how do we build a workforce and reimagine the workflows? Especially in a space where, as you know, most financial services companies, whether it's a hedge fund or a bank, are the last guys to adopt technology, right?

Mm-hmm.

We think that there's a huge moment where we're able to create the investment fund of the future, and it's happening now. My view is that—I tell my friends this, and they laugh—but I think that today, 85% of what I do basically can be done by AI. It's not a question of whether the technology is ready; it's how we implement the technology.

We're hiring a class of analysts to come in and help me with this problem, and basically figure out how to reimagine the workflows that we do every day. That goes from when I come in in the morning and check my email to see all the different sell-side notes. I spend 2 hours doing that because you have to read everything, but there are only a few important ones. How do I build a model with the click of a button? How do I take disparate data sets and bring them together? How do we do every single step of the investment process? How can we use AI to almost automate it?

If you can do that, those 6 new analysts we hire in 3 years are basically sector heads with 25 agents working around the clock. So it's a really exciting time.

Are you worried for your job?

No. The nice part about the hedge fund industry is that it's not that people-intensive, right?

Mm.

We don't need to cut people costs. There's just a huge prize for becoming exponentially more efficient at what we're trying to do, which is find ideas to invest in. The constraint on finding ideas is the amount of time you have and the places to look. I can only spend so much time looking and poking around different areas.

But if I have 25 agents able to do all that, working around the clock, I fundamentally believe we're going to be able to find better ideas faster. Even more importantly, I don't think other firms are going to adopt this that fast, and we're going to be light-years ahead.

The pitch I've been giving to the analysts we're trying to hire is: “Hey, we're going to teach you this investment process, but we're going to go and reimagine it together. How do we do this with AI?” In 3 years, when you're a full-time analyst, you're going to be exponentially more efficient, better at the job and significantly better than your competition because they're just going to start picking up these things.

I think I timed it perfectly where I'm not going to be replaced by AI yet. I just want to control it. I want to be the last—maybe the last analyst.

The AI captain.

The AI captain. Exactly. On your point about pricing stocks, though—

Mm-hmm.

It's really tricky because, on one hand, the main focus of Coatue is picking long-term winners, so investing on a multiyear horizon. What that literally means is, take Meta, for example. I have a model for Meta for what I believe they're going to do in revenue, EBIT, profit, earnings and free cash flow out to 2031 right now.

I'm projecting what they're going to do in the long term, what multiple I think the business will get assigned in that year, what that stock price is and what the return looks like. So you have your long-term view. The other way we do it is we literally build, like you maybe did in college, a discounted cash flow analysis. That means saying Meta's market cap should be worth today the sum of the future free cash flow that's generated, discounted back, right?

But as you know, stocks are moving all the time. So you have to have a long-term view of a business. What's going to happen in the industry? Are they gaining share? How are the margins going to evolve over time? How is the company's earnings profile going to evolve? But then you also better be damn sure you have a good idea of what's going to happen next quarter.

Mm-hmm.

What happened in the hedge fund industry is that, early on, when you think about Julian Robertson—and Philippe, my boss, was an analyst for Julian—

He sort of invented this: “We’re going to do fundamental analysis and invest on a multiyear timeline, and over time we’re going to be right.” Then what happened was you had guys come in who said, “We’re going to be more short-term focused. We’re going to focus on the quarters.” Data played a big role in that, right? All of a sudden, you could track credit card data. Early on, no one had that credit card data, so that was an amazing strategy.

Then the idea evolved further into having a bunch of different managers who are hyper-focused on their sector and, within those sectors, can pick winners and losers and really focus on the alpha piece. Then, as a fund, we’re going to control for all the other things—the factors, the shorts and the longs. We’re going to make sure we’re running market-neutral, and we’re going to squeeze this alpha out.

Now we’re at a point where I think the winning strategy is: how do you have a really good idea of who’s going to be a long-term winner and a long-term loser, but then marry that with a real focus on the short term? We spend a lot of time on the short term. I do, because my view is that the long term is simply a collection of quarters, right?

You want to make sure that you have an understanding of how we go from here to here, but also what that path looks like, because it creates great opportunities to buy a stock lower. Netflix is a good example. You knew what the end state for Netflix was going to be, but at every little hiccup, the stock might be down 20%.

You want to make sure in those moments that you’re not massively sized before it goes down 20%, because even if you’re a long-term investor, let me tell you, that is going to be an ugly day in the office. Then you need to know when these things have overcorrected and be able to size up in those moments when there’s a hiccup.

Because Coatue is concentrated in technology, how do you balance out these market cycles that are so favorable toward AI and what some would say is a bubble?

That’s something we think about every day. Broadly, we’re investing on the long side in tech. But even within that, if you think about the Nasdaq, the Nasdaq’s up, I think, maybe 17% or something this year. The AI trade has been a winning trade. But within that, if you pick the right stocks, maybe you’re slightly above the Nasdaq.

There have been specific sectors within that that have moved very differently. AI infrastructure—the build-out of AI, the data centers, Constellation Energy, the nuclear power needed to power these GPUs—those stocks are up 50%. You would say Microsoft is probably an AI winner, and Meta is probably an AI winner. Those stocks are up 25%.

Even in a moment where the market is going up a lot because of excitement around tech, you need to make sure that your book is sized appropriately, where you’re capturing the winners even within that, because that’s how you drive outperformance. That’s how you think about it when the market’s going up.

But even this year, there have been crazy moments. I remember I was so excited. I was like, “We need to take on more risk.” This is the 30-year-old me saying that. One of Philippe’s amazing qualities is that he is the best risk manager I’ve ever seen. He has a sense of when something is about to go wrong. It is incredible.

It’s really been great for him. In different moments during these drawdowns, he’s been able to—we call it cutting gross—go from, let’s say, 100% invested to 50% invested.

Mm-hmm.

So you’re sitting at 50% cash very quickly, and he gets the timing right. Then tariffs came. When Trump came out and put up the board with all the tariff prices—you remember that day?—I remember sitting there thinking, “Oh, God.” I think one of Philippe’s best qualities is that he understands how to bet on these tech trends, and he’s really good at picking stocks. But his single best quality is his risk management.

How do you get his buy-in on a new trade? What’s the process to get through?

I think there are a lot of people who can pick stocks, but what really matters at Coatue is that you have to be able to do the analysis. You have to be able to pick stocks, bet on longs that go up, and find shorts that go down. But the key piece is: how do you then convince Philippe and Thomas, his brother, and the rest of the group that you’re right, ultimately, to get that name in the book and to then have it play out, right?

There are a lot of people I’ve seen come through Coatue, and it was true at Melvin too. This is true at any hedge fund: they’re really smart, they’re really good at picking stocks, and they have great ideas, but they were never able to convince the person above them, who’s ultimately the decision-maker, to put that in the book. So this is a bit of an art.

The most important thing is, you spend 95% of your time doing all this deep work and all this deep analysis, but can you take that 1,000-line Excel model and all the expert calls and all the nuances around margins and growth rates and sequential growth and all those things, and can you summarize it and simplify it into a 3-sentence pitch? When he hears that pitch, he’s almost ready to buy the stock before even opening the model because the pitch is so good.

That is a skill that I’m still developing. I think that Thomas, Philippe’s brother, is probably the best I’ve ever seen at this skill. He can take something incredibly complex and get the idea down to 3 sentences where you hear it and you’re like, “That’s a great idea.” Then you go into the model, and you go into the details and show why that’s happening. I spend a lot of time thinking about, “How do I make a pitch very simple and get it in the book?”

Sorcery is proudly sponsored by Carta. Carta is transforming the private marketplace, connecting founders, investors, and limited partners through software purpose-built for private capital. Trusted by more than 65,000 companies in over 160 countries. Carta's platform of software and services lays the groundwork so you can build, invest, and scale with confidence. Carta's fund administration platform supports over 9,000 funds and SPVs, representing nearly $185 billion in assets under management, with tools designed to enhance the strategic impact of fund CFOs. For more information, visit carta.com/sorcery. That's C-A-R-T-A.com/S-O-U-R-C-E-R-Y.

Talking about AI and how value is accruing, it’s really interesting because there’s so much innovation happening on the private side. It’s OpenAI, Anthropic, all of these research labs, models—it’s a lot of things on the private side that are impacting the public side. I’m really curious because Coatue does both private and public. How does that inform your decisions? We’ll start there, and then I want to ask about valuations.

8. AI Value Spreads Across Markets

As I said before, when you’re investing in tech, you want to be talking to practitioners. One of the best parts about Coatue is that we do both public and private investing. We spend a lot of time with each other from a team perspective, but I also spend time talking to OpenAI and these various private companies to get a lay of the land of what’s going on.

I think that, at least in my 8 or 9 years doing the job, I’ve never seen a moment where the private companies are impacting the outlooks. Or, said differently, I’ve never seen a moment where a few private companies are impacting so much public-market cap the way they are today.

I think just having an understanding of really what’s going on in both areas helps you, A, be a better public investor, but B, be a better private investor. There’s also this idea that it used to be that you could be an investor in one specific sector. So you covered restaurants, and when you covered restaurants, the restaurant world wasn’t changing that much. Or maybe it was, but it was all within that ecosystem, right?

Today, you would need to have an understanding of the entire AI value chain to figure out what’s going on and who the winners and losers are. Meaning, I need to understand how many GPUs NVIDIA is planning on selling next year and who they’re going to sell them to, and have an idea of what that looks like, because those GPUs go into each of the cloud players’ businesses, and we’re at a point where your cloud revenues are 100% dependent upon how many chips you get. You get a lot of insights from seeing the entire ecosystem.

Where do you think value is going to accrue between all those layers?

Undoubtedly, there are going to be a lot of public winners. I think Meta is going to accrue—and they already are today—a lot of value. There are different offshoots. I think the biggest question right now is…

I was talking to a friend of mine who does reinforcement learning at Anthropic, and he laid out this case study of what you’re just asking: let’s take a coding agent, right?

Mm.

Let’s take Cursor. It is the most loved, most-used coding agent. They figured out this one specific area, and they’re crushing it. Well, then you have the labs. You’ve got Cursor here, and then you have the labs. OpenAI has its own coding agent, but it also has a lot of other things.

They’re doing coding, and then they’re going to do a lot of other different things. And then you’ve got Google, which basically is the labs plus Cursor plus its own cloud, its own little mini-NVIDIA with its TPUs, a search business, plus data on everything. So who wins in this? I don’t know the answer yet.

Mm.

But the funny thing is, if I just gave you that case study, you’d be like, “Oh, well, Google’s going to win because they have that plus everything.” Well, they’re also the slowest, and all the people love to use the thing on the far end of that spectrum, Cursor.

So I think this is going to be true for a lot of things. Is Cursor—and the next iteration of that for all the different applications or agents that you’d want to use—the winner because it’s so specialized and it’s gathered this adoption among the workers? Is it going to be the winner, or is it going to be OpenAI in the middle? Or is Google going to be able to take the vast amount of data it has and its ability to do things cheaper? Its cloud business—they own it. They don’t have to pay for it. Or are they going to be the winner?

I think this is going to be the biggest debate over the course of the next few years, but I don’t think we’re at a point where you need to answer that debate.

Fair.

I think all 3 win for a while.

I think it’s really interesting just seeing how much of a premium is added to these companies, even at the earliest stages. Carta reports that Series A companies with AI enablement in their name get a 30% premium. You’re looking at OpenAI; they just raised a $500 billion secondary. It’s insane.

I’ve had a partner from Altimeter on. He explained OpenAI’s valuation. I also asked Alfred about this, and maybe Elad Gil, who’s also coming on, and they all have different explanations for this. It sounds more justifiable for OpenAI versus these younger companies, and it seems like you can actually see the compounding happen there, along with the reliability and predictability of that revenue over the next couple of years, versus these smaller players.

Even at the family office level, we consider investments and we’re like, “Okay, do we think this new chip company has a chance of beating NVIDIA, or should we just do some more NVIDIA LEAPS? Like—

Yeah.

—what should we do?” So it usually just comes down to, “Oh, that’s less risky. Let’s just do that, and let’s just hedge that one.” I don’t really know what the question is on this one.

No, I mean, OpenAI—you can—we’re investors in OpenAI. The OpenAI $500 billion round makes sense to me. I get why there’s a lot of interest in doing that because, when I look at private investments, I look at them from a public-markets background, so I have a lot of analogies in the public markets.

Mm-hmm.

OpenAI’s got, what, 800 million weekly active users?

It’s gotten so crazy.

They’re spending, by my estimates, close to the amount of time every day that’s spent on Instagram. Meta is, I think, close to a $2 trillion company. OpenAI’s got a $500 billion round, and they’ve already got all these users. Could OpenAI go from a $500 billion company to a $2 trillion company, where Meta is today?

By the way, I think Meta is going to become a much larger market cap. I think Meta is going to be a 3x in 5 years. So if $2 trillion goes to $6 trillion, what could $500 billion go to? That makes sense because you’ve got users and engagement.

They’re building moats in real time. The more we’re talking to ChatGPT, the more information it has about us, and the better it can serve us products and ads. Sora was really fun. I don’t know—does it become a social company? Do they go and build a cloud? There are so many optionality plays with OpenAI that aren’t in the model you’re using. The model you have still works.

By the way, we haven’t added any of these additional opportunities to the model. Plus, you have such talent density there, and you’ve got a leader who’s going out aggressively, acquiring compute and infrastructure and building data centers in real time. They have the zeitgeist. That makes sense.

I think where it gets much harder is investing behind this proliferation of new companies. I say this with the caveat that I don’t spend my time doing this, so this is just a view from the outside. But I think it gets much harder when investing behind this proliferation of new companies.

One of the things with AI that’s great—we track this—is that it’s never been easier to start a company with AI. The fact that, with a coding agent alone, 2 guys in a dorm down the street can build software in a way they couldn’t have previously is incredible. You’re seeing a proliferation of new companies, and the prize is so big in any of these markets.

I mean, if you think about the TAM for AI, the easiest way I think about it is that there’s $20 trillion in labor spend. Software, I think, is like $1 trillion, so $1 trillion of the $20 trillion is software. That $20 trillion is up for grabs. The market opportunity is huge, but picking the winners and losers in that is really difficult.

Yeah. That’s a hard job.

Well, yeah, because every day you invest in a new startup, and then you’re just waiting for OpenAI’s new launch of something like n8n, right? You saw DevDay: “Oh, here’s our version.” And you’re like, “Oh, well, okay.”

Yep.

So there are so many examples of that. Yeah.

What is it called? Sherlocking or something?

Yeah.

I think that was with another one. I know we covered this a little bit, but I want to touch upon it again. For you particularly, what are the metrics that you track within these different companies to determine their success?

It depends on the industry, but broadly, we have this 5- or 6-year view of these companies. In the near term, what we’re tracking is broadly inflections. In the digital ad business, that’s an inflection in growth rates. In the cloud business, it’s an inflection in growth rates or margins, where you have a quarter that is better than people think, or worse than people think, and it helps prove out your thesis faster.

If you think about having a 5-year view of what a stock is going to be, ideally, we want the market to figure out as quickly as possible that that’s where it’s going. We say that IRRs get pulled forward. When you have a moment of inflection, that’s where your IRR can get pulled forward, and the stock reprices higher, more toward your view.

We’re tracking everything. I’ll give you an example: we’re tracking everything from credit card data to email traffic. My analyst sent me this today. Every Thursday, we sit down and have KPI tracking, using some real-time dataset or a mixture of them, for every single company we cover.

Even if I’m not looking at it, even if we’re not invested in the company, I look at that tracking every single week because it tells you something might be changing, and that might be a source of a new idea. It gives you an understanding of where we are in the broad economy.

Those are sort of table stakes and basic, but you take all that together—you’re looking at the ad market, e-commerce, and payments—and you have an understanding of where you are in the economy. Are things getting faster? Are things slowing?

Ads have been really great in Q3, but about a week ago, they started to slow. Is consumer spending slowing, or is that just a weird shoulder period in time?

Where I think the data science gets really interesting is when you can take differentiated datasets and piece them together to get a unique view of something happening that other people can’t see. One good example of this is one of the companies that we invest in: Reddit. We love Steve Huffman. We love the team.

We think Reddit is going to be a much bigger business over time, that it’s going to be this great ad platform, and that, really, in the AI era, there’s only one place where actual human-generated content exists. That content is super valuable.

It's really valuable because it helps train the models. If OpenAI wants to have a shopping assistant, right? All the reviews in the world are on Google. They're not on OpenAI, on ChatGPT today. So where do they go? They have to go to Reddit. What are they willing to pay Reddit to be able to use that data to ultimately build the shopping assistant that's going to take over the market? The answer is probably a lot.

But there was this moment where search, as you know, is being rearchitected, right? You have AI Overviews, and now you have ChatGPT. Reddit, at the time, was growing users, and then there was a little bit of a hiccup. The hiccup was related to AI Overviews being shown. If you think about the old world, you type in something on Google, and Reddit was one of the top links. Well, now you have an AI Overview that's taking up your screen, so Reddit's now down here.

Mm.

And you're like, "Okay, they just missed this metric." The market's freaking out because it's very easy to say, "Well, they were only growing because of Google, and now AI Overviews just took their entire slot. This thing will never grow again." That's how the public markets react. "This will never grow again," so the stock's plummeting.

But what we figured out was that we were basically able to figure out that, in the old world of Google, when you typed in a Google search, Reddit came up maybe—I'm just going to use fake numbers—10% of the time.

Mm.

And within AI Overviews, when they first started showing them, when AI Overviews were 5% of search, they were showing up 2% of the time. And you're like, "That's not great." Well, then AI Overviews became 50% of search in 2 months, and within that, Reddit went from 2% to 15%. So it's actually higher than in the old world, but because it went from 0 to 50%, that was the disruption.

The second we saw that, our takeaway was, A, those problems are going to be fixed, and B, that's proof that Reddit's actually more valuable in an AI world because they're showing it more because consumers want to see it. They liked seeing the answers and the citations from Reddit. So that's where you have the confidence with that data science to say, "I figured out the tech change, and now we like the stock even more, and this is how the user numbers are going to be fine now."

And now the narrative, as this gets out, is going to be not that they're an unclear AI winner or loser. No, they're going to be in the AI winner camp, and that means your multiple goes higher. The stock went up a lot when the market figured this out.

Reddit is one that I find fascinating to watch because I didn't understand why anybody was paying them that much money for their data. It just didn't make any sense.

Well, and the funny thing about that is that the thinking, even from Reddit, is changing a lot. What happened was OpenAI went out and basically trained ChatGPT on Reddit data, right?

Yeah.

And it sounds like they may not have asked for permission or done the right thing. At the time, it was just that you were going out and trying to build this model. So what Reddit did—and Google did the same thing in all of this—was basically say, "Okay, you guys, we're not going to sue you, but you took our data, so just pay us a licensing fee."

I think it's ballpark $50 million. Google pays Reddit $50 million a year to, A, have trained on it in the past, but B, have updated data. ChatGPT does the same thing.

So the view is, okay, Reddit's corpus of data is growing, but the incremental conversations that are happening are pretty small in comparison to the whole thing. Whatever that deal was in the beginning, it's not going to get better, right? That $50 million isn't going to go up. That was the view. I think that's what everyone thought, including the companies.

But then, as AI evolved, you started to realize that incremental data that happens is actually way more valuable. For example, if you want to build a shopping assistant, and Reddit is one of the primary sources on the internet where people are talking about products and what's good and what's bad, there were rumors that Mark Zuckerberg was hiring people for $100 million a year to build this model.

So if he's willing to do that, what do you think Google or OpenAI is willing to pay Reddit for the key piece of data that may determine the success of the entire shopping agentic TAM? My guess is higher.

I have 1 last question. This one is going to be really difficult. Are you ready?

I'm ready.

You ready? There is some confusion around the name Coatue. I know Philippe and Thomas are French, but a previous partner I used to work for would call it "Koatu."

Oh, yeah. That's wrong. That's just factually wrong.

Sorry, Mark. Sorry to call you out, too. Can you please explain to the class where the name Coatue comes from?

Yes, so I'm glad I know this one. Coatue is a beach in Nantucket.

Okay.

So it's a beach in Nantucket. I believe Philippe spent time there. What's funny is that, in my entire time at Coatue, I've never heard anyone talk about the beach.

The beach?

But actually, that's not true. We're redoing our office. We're basically building a 2nd floor because the firm's expanding. We need more room. People are coming up with names for the new conference rooms, and Thomas' idea was to name them after other beaches in Nantucket.

You kind of do need a beach and a hedge fund in Midtown. Why not?

Yeah. Well, as you know, Midtown is a stormy sea. Every day in the market feels like a stormy sea, so any beach would be good.

Okay. Well, that's a good way to end it.

Yeah. I appreciate you having me.

Thank you so much.

Thank you so much.

Of course. And hopefully you get a podcast studio in this new office.

We're—I think we might do that.

Do it.

If we do anything, we're going to have you on.

Thank you. I was going to just show up, but I appreciate the invitation.

Sure.

Thanks, Michael.

Awesome. Thank you.

Hey, it's Molly. If you enjoy our interviews, check out our newsletter, Sorcery.vc, where we deliver a once a week top deals and tech headlines email, and also go deeper on our podcast interviews. Subscribe to Sorcery today. And don't forget to subscribe to the podcast on YouTube, Spotify, Apple, or wherever you listen. Link in description to sign up.