Citrini Research 拆解：智能体、“幽灵GDP”、消费支出｜Figma 财报超预期

1. Anthropic Security Product Wipes Billions Off Public Markets

It is so good to be back. I love this. We're going to start this week with some news on Anthropic. Surprisingly, its security review feature wiped $20 billion off cybersecurity stocks. Obviously, Anthropic released its latest security product, and it massively hit some of the biggest players—Cloudflare and CrowdStrike, to name a few. Is this a dramatic overreaction from public markets, or is there underlying truth to this?

Jason Lemkin

Well, I'll tell you the interesting thing to me. This kind of shows where the markets are and where our mass panic is. Although I think our panic is well grounded; I think we should be panicking. But most of this already exists.

You can go to Claude Code today, and I literally did this on the plane flying back last week. I did it inside of Replit, and you can say, “Run a detailed security audit on my code,” and it will already do it. It will already run a total static code review, and if you're on Replit, Lovable, or Vercel or something, it will actually do penetration testing and everything for you. Literally, it will do everything.

I think it can already do better security audits and testing than a mediocre engineer will ever do. So what I mean is, obviously, the pace at which things are getting better is—I mean, each week, we can't keep up on this show. But it's also interesting that, conceptually, just like Claude Code Review dropped IBM 10%, it already existed. We're panicking about features that have already been in Claude for months, in many cases.

Is this a sign that it will go more aggressively into the space? Of course it is. But I just want to point out to folks that aren't doing it: so much of the stuff that we are panicking about, which I think is worth panicking about, is a total nothing burger in the sense that, if you're paying attention, this is months-old news in some ways.

The first general comment is that Anthropic markets itself as the nice AI company. For a nice AI company, it sure creates a lot of damage and kills a lot of stock portfolios. Maybe the Pentagon is wrong and they need to buy more Anthropic and just point it at the enemy. It'll bring China to its knees, right? This is an astonishingly destructive company.

Incredibly distracting. Incredibly destructive—to have a press conference clip $20 billion off something. But let's do security first, because in this case, from a technical capabilities perspective, no surprise.

Second, from an adoption-in-the-enterprise perspective, I don't see Anthropic eating up CrowdStrike's business. I just don't see that. No, people are going to have a security layer. There was a great—I think it was an HSBC report out today, after the Citrini madness, which we'll talk about in a second—that basically said, “Software is the means by which AI will diffuse into the enterprise,” which I thought was a wonderful quote, and I think it'll be true for security also.

In other words, these capabilities, like code scanning, will diffuse into the enterprise probably by means of companies like CrowdStrike and other companies like them bringing it to bear. So at that level, A, no new news, and B, it shouldn't be such a big panic.

But—and this is the big but—unlike some of these other sectors, CrowdStrike and the security companies were effectively trading at a price that assumed nothing could ever go wrong. Even on Friday after the correction, I went in to look and I was stunned: 22% revenue growth projected, 31% cash margins, and it was trading at 16x revenues on Friday after the first hit.

The thing is this: when you are priced for perfection, anything less than perfection will be a kick in the nuts. That's what's going on here. It's entirely separate when you're dealing with the companies that were priced at 6x revenues and they've gone to 4x. That's a separate discussion we'll have later.

But I think what's happening on some of these super high-multiple stocks is the occasional reminder that, when you're priced to perfection, literally any little thing—any increase in tail risk of you not being the winner—logically corrects pretty substantially. So I think that's all that went on here. I don't think CrowdStrike is obviated by this. I think it'll totally be fine; it'll continue to be a business.

But when things are priced at extraordinarily high prices, it doesn't take a lot to knock the narrative off-kilter. Would you say that CrowdStrike, with the repricing, is now fairly priced or underpriced?

Jason Lemkin

It's actually a very good question, and to make it more general: do you prefer the stories where the AI disruption is perhaps a little remote but the values are still pretty lofty, like CrowdStrike? Or do you prefer the ones where some of these stocks have corrected to 6, 7, and 8 times 2026 EBITDA, where you go, just on a value basis, that's wrong?

I was thinking about companies like Toast. I was thinking about one of our own, DocuSign, at 7 or 8 times revenues. These things are absurdly cheap, whereas CrowdStrike is still not yet cheap. It's just marginally less expensive, but it probably has more clarity on the way the world is going.

I probably err toward a bit of value. In other words, I just say to myself, buy the things where the cash flow alone makes it easy. Would you rather have your money in CrowdStrike, which is still priced well, or Monday.com, priced at 1.5x revenues?

I'd like a basket of companies at 1.5x revenues, not an individual stock, because I think at the level of an individual stock, it's hard to say; it's very idiosyncratic. At the level of the basket, if you buy 20 stocks at an average of 3x revenues and 8x EBITDA, I think you'll do just fine.

I don't know, man. Just last week, you kind of mocked me, and I can take it; it's good. You mocked me for saying Shopify is at partial risk of disruption. I said it's not at total risk of disruption, but the fact that whoever builds the agentic layer will have more and more value accrue to them means you only have to have a partial deceleration in your numbers.

You only have to have a partial risk of disruption. You only have to see more of the value of HubSpot or DocuSign flow to an agent for these stocks to do worse. Look, I'll give you an example for e-signature. I know this space. DocuSign, right? It's a great company—

Jason Lemkin

We both know it, and a lot of folks were saying, “Oh, this is going to be destroyed by Claude, because all you're doing is creating an image of a signature.” No, this is a very complex enterprise workflow system that also is a partial system of record.

Someone has to say these contracts are true and valid, and someone has to route them through 100 steps of transformation and negotiation and all this stuff. So that is not going to be destroyed inside of a Claude chat, but some of that can be done by an agent—an agent that truly auto-contracts for a business.

Okay, let's say it does all of your commercial transactions. That could take enough of the value away that these companies are maimed. And I think—and I even think it's logical for CrowdStrike—obviously, Claude Code Review is not doing endpoint security, but can anybody be maimed by Claude?

Claude just launched an entire enterprise agent solution today. Do you think anybody can't be? I think anything can be. This is how I think about it: Claude is like inverse Fortnite, the game where you fall out of the sky onto an island and it keeps shrinking. I used to play it; everyone plays the game, the video game where it keeps shrinking.

Your territory keeps shrinking. Claude keeps consuming more and more of you, and you're stuck on a smaller and smaller island that you have to own more and more market share of. For almost everyone that's public, your surface area is shrinking because of Claude and AI, and the question is how much.

I don't know that we've seen the bottom of these impacts. I think the impacts are accelerating, and that's why Anthropic today had to publicly say—its head of AI said, “Listen, guys, on the one hand, things are accelerating, so it's bad for software. On the other hand, we're best friends with software.”

On the day they launched the enterprise agent, we're best friends. We're enabling them, but the pace of change is so fast.

So, I want to believe there are these safe islands. But I believe the agents are going to own enough of the value that just owning less of the value in your space can create terminal decline.

Jason Lemkin

Terminal decline.

Give us the rest of the year.

Jason Lemkin

You're going to come around on me on this one. You're going to come around. Just give me the rest of the year.

Literally, on the plane last week, I did a security audit in Replit, which is using Claude Code. Some of that is Replit—we could talk about it, but I don't know the details. It's great. I literally sent it to the Replit team—the entire technical team—because I'm like, "Do you see how good this is?" And they're like, "We didn't even know it was this good yet. We didn't even know our security audit had become this good last week. It's so good and so much better than before." And they didn't even know.

So, to think that we're in some sort of static world for the rest of the year, I don't even think so. Literally, we could delete some of our podcasts from 4 weeks ago. They're so dated.

2. Do Agents Turn SaaS Incumbents into Valueless Databases

So, let's try not to be like that. A comment here: intelligence will infuse all software over the next decade. That intelligence is generated by foundation models like Claude and OpenAI. I think that's a given.

Jason Lemkin

Agreed.

So, what we're really saying is: how much of that do they do themselves? That intelligence has to get to the enterprise, and there's probably 4 ways it can get there.

The first is they buy it from Claude directly—everyone buys all the software from Claude. The second is that they build it themselves; every enterprise constructs its own agents. The third is they buy it from existing incumbents who integrate AI. The fourth is they buy it from gazillions of new companies, all of whom are leveraging foundation models. You founded SaaStr, right? You have to figure out which of those scenarios you believe in.

Actually, the post I read—and I agreed with it very much—is that I think it's 3 and 4. I don't think enterprises are going to digest or build their own systems on top of Claude directly, and I don't think Claude's going to build all these focused systems for everything. I think there's going to be this software mediation layer between them, and that's the opportunity.

Jason Lemkin

Here's the thing. If you look at all the publicly traded B2B companies, there's only 1 that has a competitive agent. It's Palantir. No one else has seen a single ounce of revenue acceleration due to their AI agents.

And yet, the companies we talk about each week have jaw-dropping acceleration. It's not just Anthropic, because they have built the agents that matter in their space. It is not sprinkling AI dust on top of their analytics software. There's not enough value there.

But let's talk about that 4th category. You would also agree that there are many privately held, recently founded companies exploding in revenue as well. Do you think it all goes to the model companies, or do you believe any of these companies—I mean, you mentioned you guys talk about Replit and Lovable a lot—do you believe that's defensible?

They might not even be defensible at the rate of change, because Claude Code just last week, since we did the last show, launched the ability to see apps inside of Claude Code. So, for a lot of product people, you don't need Replit or Lovable anymore as of last week. Now you can change Claude Code and visualize your app inside of Claude Desktop and inside of Claude Code. You don't need to do that anymore in a third-party app.

Jason Lemkin

And so they're all being disrupted. But my real point is—sorry, Harry, you're the boss—I’ve talked to 3 founders over the weekend of public and near-public companies, and this is the advice I gave them: your agent is not great. You're being disrupted by the agentic layer.

Now, I hope that ServiceNow builds these great agents, and I believe Agentforce has a shot, and I believe others do, too. But they are all being disrupted in real time, and that's why the folks Harry interviews on 20VC are stressed as fuck. They are stressed as fuck because, no matter what they say, they know they do not have the dominant agent in the space. No matter how many LLMs they stick inside of a feature.

When I see what many leaders of top public and private companies and our 100 nine-figure unicorns are saying, it's lip service. I see it on LinkedIn all weekend long: "We're adding AI to our email feature. We're adding the ability to process emails more efficiently." You're going to go out of business, and you're not going to fail because your customers are going to renew, but your growth is going to fall so far that you become irrelevant in 2 years.

And that's why I'm going to buy my 4 stocks, don't get me wrong, but I'm already changing my mind since last week, because things continue to evolve. The panic is overdone and real at the same time.

You made a very clear statement that we have not seen any of the public providers make great agentic use cases work and have a meaningful impact on revenue. I'm very naive: why? Tobi's a brilliant CEO of Shopify. Marc is trying with agents. Why have they all failed so far?

Jason Lemkin

I'll give you 2 reasons if you want. There's a long list of reasons, right? I'll give you 2 practical reasons. One is it's a lot of work, man, and no one wants to do this at these companies. No one wants to.

Every agent—here's the problem today, and this will change in the next 2 years; it is not true today—every agent is essentially custom. Every agent needs to be trained. Every agent needs to be onboarded if you want it to be great. Every agent needs its data cleansed.

This is a vast amount of work for organizations that already think they're overworked and working too hard. There is huge institutional momentum to overcome. The second is what I just said is true: you need a massive amount of forward-deployed engineers and trained workers that are technical enough and smart enough to train and deploy an agent.

The workers don't exist in most companies. The average customer success person who shows up with a green, yellow, red-light dashboard cannot train and tune an agent. And then there's a meta challenge for the Shopifys, Mondays, HubSpots, Toasts, and others, which is you can't afford the human to do it.

At a niche level, we're seeing it with startups; we're not seeing it with public companies. Hyper-niche agents work really well because they have a small set of things to do. As soon as you get to spaces like Shopify or, even worse, Monday, where you have 100 verticals, it's very, very hard to build a very specific agent automatically that does everything that churches need, basketball courts need, and refrigerator businesses need. They're not all the same needs, and so the agents aren't good enough.

In fact, a lot of these leaders that I describe, if you look at them, they're in beta. They have 6 people using them, 60 people, because it's too hard for them. So they're going to get killed by the startup that does it. They're going to get killed.

Killed. But that's the key: a startup that does it.

Jason Lemkin

What it's not going to be is the foundation model directly selling to the church, directly selling to the thing. They will provide the raw intelligence, but there will be software opportunities to build compelling software companies in most of these verticals, just as there was in SaaS. So, maybe we're more in sync than we think.

Anyway, I think—let me just say clearly—I think intelligence-led applications are the only applications that are going to sell and grow quickly over the next 10 years. Non-intelligent applications will at best be flat to mild growth if they're not obviously disrupted by intelligence, and at worst be down.

3. Why No Public Company Has Created a Good Agent Product

Right. Yeah. I would argue, for example, I'm picking on Toast because I'm not an investor, so I have no emotional connection. I think that's a good one because I don't think there's a ton of agent work to be done, and it's a lot of payments and restaurant organization, which is fairly durable. We can argue, but yeah.

Whereas something like Monday, it's very knowledge-work-y. I can see a much bigger disruption story there. That's the incumbent, but your point—and what I think we're now in sync and saying—is that those opportunities can be grabbed by standalone companies, perhaps built by companies started and founded in 2021 and 2022, leveraging directly on top of the foundation models.

What you're not saying is that all that revenue just accrues to the foundation models, correct?

Jason Lemkin

No. I think people are going to be maimed even more than they think if they don't own the agents in their category. You've got to own the agents in your category. Whether those agents are owned by a startup or whether some version of that agent can be done inside of Claude, they're going to maim you. And it's accelerating.

Jason, I want to be very direct with you, and you'll give me a direct answer. You're better at direct than Rory's sometimes nuanced answers, which people love as well.

Will Claude Code make Replit and Lovable weaker within 12 months? Do you think they will meaningfully enter their space and take market share?

Jason Lemkin

I think it will do everything that it can do. To Rory's point, anything that can be done either inside the browser or inside the desktop, Claude will do. That's what we've learned this year. Anything it can do.

So, right now, if you go to design something in Claude Code, you'll laugh.

All the crappy Claude Code websites look the same. They have the same artifacts, the same icons, and the same purple color scheme. You can laugh. But can all of the design—all of the parts of Figma that are design—be done within the code? Of course it can.

So I believe it will aggressively attack parts of Figma this year, even though they're key partners. Replit and Lovable—it continues to do more of them. Now, will Claude Code want to host entire websites? Will it want to go into production?

The only thing that ultimately will protect them is that this is all their teams do, right? I don't think they want to build databases, build production websites, and host domain names. But if they change their mind, there's no reason they can't license Supabase or Neon, or fork their own Postgres.

They're pretty good at that company. They can build their own database. They've already got plenty of servers; they can spool up a few more to host websites if they want to. They can. Fortnite's the game, you know: at the end of Fortnite, the circle gets smaller. So if I'm at Replit, Lovable, or even Figma, I would be worried. The circle's just starting to shrink.

I think at many companies, that circle is shrinking. I have 2 investments I made last year that I love. Those products no longer have a reason to exist today because of Claude. I made 2 investments—not enough to bankrupt the fund, and they've tilted and evolved—but these were standalone investments last year, and I'm not going to go into them, that were great, that blew up in the early days, and they just have no reason to exist today in that prior form. Just no reason.

When we turn around and, all of a sudden, you can preview your entire app inside of Claude, which you couldn't do last week, if I'm any of these companies—Figma, Replit, Lovable, Vercel—all of them, I love that the Fortnite circle is shrinking. So you've got to do something about it, right? And will it stop? At the end of the game, it does get pretty small, though, right? It's stressful.

4. Anthropic Secondary Sale Makes Hundreds Decamillionaires

But the flip side is, if you nail the agent, look how much revenue these guys did building essentially an agent on top of Claude Code. They built $1 billion of revenue building the agent that didn't exist. So that's the flip side. That's our job: to build these billion-dollar agents. If you can do something that is extremely high-value that could not be done before, you can close millions in revenue in your first week. It's never happened before in the history of software.

Before we move on, I do just want to stay on Anthropic. They've lined up $5 billion to $6 billion for an employee share sale at a $350 billion valuation. Obviously, there are people queuing up out the door for this. This follows OpenAI doing the same a couple of months ago.

Have we ever seen liquidity at this scale when we look at the number of millionaires minted from OpenAI, soon to be Anthropic? When we look at the NVIDIA millionaires that exist already, is anyone going to be able to buy a house in the Valley? There are a lot of things that seemed silly in AI 12 or 18 months ago.

Jason Lemkin

I'm going to call bullshit here.

Okay, great.

Jason Lemkin

I love Noah Smith's descriptions on Substack. I follow him in economics a lot. He called it basically scary bedtime reading.

I think the way you need to break it up is to figure out—because I knew we were going to have this one, and I saw your 7 points, and it's too much detail, Harry—the big picture. I would suggest we approach it in the following fashion.

The first thing you have to figure out is, at a micro level, for each of the things he says are going to happen, do we believe they're going to happen? In other words, is AI going to replace coding? Is AI going to replace DoorDash? Is AI going to replace Amex?

The wonderful thing is, I think that's something this group is well equipped to do, because we're all investing in venture companies that are at the tippy point of the spear in terms of adoption. I'm allowed to have an informed opinion on question 1, which is the micro level: are these changes going to happen?

Then the second big-picture question, lumping all the other things together, is this: what are the macro consequences of this? In other words, if you assume that there's a high level of AI adoption over a short period of time, then everyone—and that's, as I say, what we can talk about because we understand it—gets to pontificate on global macro, which is what you were starting to do there.

I'm fine coming back to that and revisiting this in a few minutes. But I think he was hypothesizing a 2-year adoption cycle of almost everything, such that everything from ServiceNow to DoorDash to Amex gets rolled over in 2 years. If you believe that's going to happen very quickly, then you do graduate to the global macro question. I still think he's wrong about that, but if you don't think the adoption is going to be that quick, you can literally ignore the rest of the piece.

5. Will DoorDash Be Replaced by Agents

It totally does. Why don't we start with the micro, then, where you think we are seasoned and responsible enough to have an informed opinion? Let's break it up. He points to software development, SaaS apps, companies like DoorDash, and then companies like Amex—in other words, interchange, right?

Jason Lemkin

Well, yeah, let's break it up. Let's do the stupid ones first. DoorDash, right? The idea was you're going to want to delegate to your agent the purchasing of an item, optimizing for—there are going to be 6 different versions of DoorDash, and your agent will choose between them. The only reason that you don't do this today is because of friction, and if it was automated, you'd let the agents order your pizza.

I just call bullshit on that. At the consumer level, I don't know how you feel when you're buying pizza on a Friday night. You're talking to your wife, and you say, “Honey, you know, we had the freaking kimona[?] last time. I say we go with the pepperoni.” She's like, “No, I don't like it. I told you we want to stick with the 2 salads and the pizza.” It's not something we want. No one wants to delegate to an agent how to decide what food they get and then have the thing come up: “Good news, I saved you $2. Bad news, you like the high-end pizza, but I got you the crappy little pizza.”

Rory, can I add just one thought? Here's Andy Fang, CTO of DoorDash.

Guest

“We strongly believe agent commerce will be transformative to our industry.” He believes this. I believe it. I have a large investment exposed in the space, and I can see agents and AI ripping through it.

I don't mean to interrupt, but I think the examples we think are safe—this is DoorDash saying, “We need to earn the right to service customers and agents end to end: discovery, ordering, delivery, and support. We need to earn the right in the new world.”

So to think that these spaces aren't threatened—it only has to be maimed. It does not have to be the idea that we're going to vibe-code our own DoorDash. That's stupid, right? It is stupid.

Jason Lemkin

Yeah, so good.

And he's trying to get people millions of views, right? So the core idea he started with is stupid, but keep extrapolating, right? Now that we've agreed that an agent can decide for you, the CTO says it's real. That's the threat.

Jason Lemkin

But he didn't say, “Okay, you caught me. I'm going home.” If you want to make more automations around a recommendation, say, “Hey, there are 3 different pizzas.” But do you really think it's going to be that way? I just think there's a level of customer inertia for consumers to move to this vision of 5 different competitive DoorDash companies being enabled in this world.

I just don’t see it. DoorDash is a combination of a huge amount of logistics, a huge amount of customer aftermarket service, and a huge amount of signing up restaurants, right? The only thing that’s changed is that they beat out 4 or 5 other big competitors to now have some kind of stable oligopoly with Uber Eats and one other smaller player in the US, right?

What in software is going to allow a new competitor to emerge and take market share in a high-fixed-cost business like this? Talk me through the new competitor emerging and taking market share.

Jason Lemkin

Well, the most simple one is that a new competitor can decide whether Uber Eats, DoorDash, or direct is the right thing. You have 3 options in the US. Nothing else really exists, right? The agent may make that decision.

In fact, I would prefer that because I don’t want to figure out which one to use: DoorDash, Uber Eats, or direct. I would prefer that the agent knows my favor—I'm just picking one example, but this is a real threat today. This is what Andy Fang is saying: it’s a real threat that we no longer need to decide which is the best place for us.

The agent decides which is the best deal between these options, which is the best source for me, which is the best for my family, and it makes the decision. It only risks disruption if the agent is who we go to, as long as the agent is who we go to rather than the first party. It doesn’t destroy the company. It doesn’t destroy it.

Yeah, it doesn’t destroy the company. Okay, now we’re tracing through the unrealistic statements there. Let’s go to an existing example today: Netflix recommendations.

Jason Lemkin

But I’m right about DoorDash because Andy Fang said the same thing. Why are you dodging the one that everyone thinks is free from disruption when the CTO says the ground is shifting underneath his feet? He literally said it this week—the CTO.

Okay. No one’s going to say, when they’re the CEO of a public company, “I don’t believe that stuff.” They’re going to say, “We’re on it,” because that’s the message you’ve got to give, right? But I don’t think they’re saying, “Oh, my God, 3 more companies are going to displace me.” Our job as investors is to analyze the facts and try to come to our independent conclusions.

So I’m just going to take the idea. I think 2 of the most personal things at the consumer level are the food you eat and the TV content you watch, right? The good thing about the content you watch is that we’ve had 10 years of AI already. The Netflix recommendation engine, let’s be clear, is a massively useful tool to them because, at the margin, it helps them predict what people want.

I do agree there is core value in knowing people’s preferences in aggregate, right? But what percentage of your content viewing do you base entirely blindly on the recommendation engine when you sit down at Netflix—5%, 10%? I think it’s light.

Jason Lemkin

Sorry, but DoorDash is barely in B2B. What’s the Netflix point? I’ll answer your question if you want the answer, but I’m missing the point. Do I think AI can disrupt Netflix? Netflix thinks AI can disrupt Netflix because we’re all watching short-form content.

As of the last 45 days, you can watch an incredible short on YouTube that was entirely AI-generated. You can watch Star Wars stories that are better than the crappy last 3 movies, and they’re AI-generated. That’s utterly disruptive to Netflix. Utterly, utterly. They’re so panicked, they have to buy a studio.

I’m breaking it apart into 2 separate things, right? If you can generate content using AI, that’s very disruptive. I was trying to focus on recommendations because the idea is consumer preference: you will entrust your decision-making to an agent who will, quote-unquote, know what you want. That’s what you’re saying about DoorDash, and we’re trying to prove something that hasn’t happened yet.

I was making the point that content is another thing that’s quite personal. Netflix has had this agent running for the last 10 or 15 years—the recommendation engine—and, at the margin, it does a good job of predicting in aggregate what people want. But if you had 2 choices—1 program that gave you exactly what they recommended, and the other program that allowed you to pick—I’m going to tell you, you’re not going to go with the recommendation engine.

You’re going to say, “I didn’t like that recommendation last week. I’m done.”

Jason Lemkin

But, dude, YouTube is the number-one way we consume video, and it is entirely based on a recommendation engine. It is the best recommendation company on planet Earth. There are no channels on YouTube that matter anymore. Followers don’t even matter anymore. Nothing matters.

Every day I log into YouTube, and it gets better and better at knowing what I want to watch. It is epically—epically—better than anything else on the planet and utterly disruptive to how we view a few things.

It’s a good argument. I’ll give you that.

Jason Lemkin

Everyone is stressed today. It’s a good thing. Everyone is stressed because the further you go—and again, I don’t mean to—the further you go, the more folks are at risk of being maimed by AI. Just maimed. Except George Kurtz at CrowdStrike—he’s fine—but everyone else is at risk of being maimed by AI.

Even if your growth goes from 31% to 20%, that’s a big deal. That’s a big deal.

Agreed. But do you really believe that you want an agent to recommend your food?

Jason Lemkin

Well, listen. I already know the answer is yes. It’s not my opinion. I have data from over 10,000 restaurants. I know the answer is yes.

Right. If your agent had all the historical context on every pizza order you and your wife have made—so it knew the price point, the location, and the estimated delivery time—and it could also analyze every TikTok and Instagram review and the latest food trends, it could tell you about the latest option within that price band, with that crisp topping your wife likes.

It already knows that because you left a review or because you said it in a WhatsApp. Say it’s got OpenClaw and is able to deliver that to you. I think most people would.

Jason Lemkin

I’d love it. But, Harry, I think that, at the margin, yes, I’d be sitting there on DoorDash going, “I get a 5% extra satisfaction rate and selection rate from this.” But do you think AI is so disruptive that it can warrant the creation of an entirely new company?

Basically, what you’re saying is all the investment you made in restaurant relationships, logistics, the app, and the consumer, right, is as nothing because this new thing is disruptive enough—

No, valuations are fine. No one said they’re nothing. I don’t think anyone said they’re nothing.

Jason Lemkin

So it’s not. All these SaaS apps become dumb databases, and Toast just becomes a POS system. They’re not going away, but they become more and more commodified. More importantly, they don’t capture enough of the incremental value. It’s the incremental value we’re investing in.

6. Will "Ghost GDP" Soften Consumer Spending Power

Jason, I wanted to deal with the easy ones first, right? Do you really believe that there will be a direct competitor to DoorDash enabled by AI?

Jason Lemkin

No, listen. We all agree that that is clickbaity, and so is all of it. But let’s talk about Ghost GDP for a minute. I admit I’m living on the bleeding edge. Let’s concede that I’m a laboratory.

We’ve gone from 12 people to 2 people on my little team. That team not only does investment; it generates 8 figures a year in revenue. That is Ghost GDP. The value from those folks who are gone—the profits that are left—go to 2 people. Well, 3 people.

Define Ghost GDP for me.

Jason Lemkin

Ghost GDP is the idea that this productivity is not going to human workers who then spend it. That’s the fear: that we’re creating this productivity, but it is not going to anyone. There are no humans to spend the money.

It’s great that I can spend more money, but I don’t think that’s great for the economy if I get a little bit richer. We lost 8 people on our team, right?

Okay, stop. First of all, I agree we’re now doing macro because—and in this case, we can do macro because we agree we can only do macro once we’ve conquered micro. In this case, you’ve conquered micro. You’ve said it has happened: “I had 12 people; now I only have 2.” 10 people no longer have jobs.

What does that mean? Well, at least they have other jobs, but the value that we’re creating—this 8-figure value—is accruing to fewer and fewer people. There aren’t as many people to buy handbags, shoes, T-shirts, and Netflix. Even just to buy Netflix, there are fewer people, right?

That was the point of Ghost GDP in this inflammatory, annoying article. But I don’t think these things are wrong. I think he’s just trying to claim everything’s going to happen in 18 months, and it’s not going to happen in 18 months.

Jason Lemkin

What we’re basically saying—but implicit in that, and you have to be logical, implicit in that—is that productivity gains, which have been the engine of growth for the last 200 years, are bad. There’s some implicit statement here that it is bad that Jason is now able to do something with 2 people that he had hitherto only been able to do with 12.

Right? And I’m going to say something. Across the arc of the last 200 years since the Industrial Revolution, productivity gains have been good, right? Because the other 10 people who used to be wasting time writing slop for Jason can now do other things, and the sum total of human achievement will contain the extra work that those people do.

In the long term, I don’t think you can argue but that productivity gains are good, right? We used to have 80% of people working on farms. We now have 10% of people working on farms. Actually, 4% of people working on farms. And we sell so much food that we’re all fat, right? Those other 85% of people are doing other things.

Across the scope of history, productivity is freaking awesome. It’s the only thing that’s made us rich. So, I want to say that so clearly because then all these macro arguments that I was hoping to ignore are basically some version of: even though productivity is amazing in the long term, because you can’t disagree with that, something bad is happening in the short term.

And then the onus is on you to say what that is. What is it? What’s so bad about constant change and the fact that jobs go away and new jobs emerge? What’s going to happen that’s so bad?

Jason Lemkin

A softening of consumer spend, with a concentration of wealth among fewer people. There are fewer people to spend money across different parts of the economy, and that impacts a large number of people. And to your point on how productivity actually leads to a worsening in the economy, you only need to look at Japan in the 1990s and the need for Abenomics.

You saw massive productivity increases in the 1990s, with massively improving mechanical infrastructure that they brought in, and actually it didn’t disperse to a huge amount of the Japanese population. So, there’s very recent precedent, actually.

If you want to see a dystopian version of this, go to Japan and meet with B2B founders. I was at a dinner last November with all founders from IPOs, $20 million and up only. It was a VIP dinner put together. It was great. It was the best of the best, okay?

And they’re all talking about how, inherently, their seat base shrinks. This is not just the AI topic we’re talking about today, about seat-based risk. This was last year, all talking about how each year their seats shrink because their economy is shrinking, right?

It is, just to Harry’s point, a structural headwind to people buying things. I think in the short term, this is all great for us as investors. It’s terrific for us to get more productivity. We’re going to make money out of it, and we should put it in the bank and flee to Miami or Monaco, because I don’t know, but I’m not sure it’s good for everybody.

Jason Lemkin

I’m going to call bullshit on that. There are 2 or 3 different things with disaggregating and the seat-based comment, right? Are you making a comment on Japanese depopulation, which I don’t think we can blame on AI? It’s been a trend for 30 years. I just think there’s a loose parallel to this Ghost GDP idea Harry brought up, of depopulation.

Well, there are different ways you can depopulate a workforce even if the humans are still there, right? But it is a similar structural headwind to folks buying stuff.

Our 10 agents at SaaStr generate millions of revenue, but they buy nothing. Our agents buy nothing. They work all weekend long. Repley[?], Art, Qualie[?], and Monty—they’re good kids, okay? They create a lot of noise; they’re a lot of work, but they buy nothing. Nothing except tokens. That’s the only thing they buy: tokens, and they buy millions and millions of tokens, for real. That is a little different from the past.

Productivity increases are only good if the consumer wallet is dispersed and people are able to spend money. If that shrinks, that is not a good thing.

Yes, but what do you do? Ban productivity increases? Good news: we’re all doing fine. Bad news: we’re all in 1790 and we’re all one bad harvest away from starvation, but yay, we’re all fine as long as nothing goes wrong, right? It’s not a credible argument.

Look, I do agree that at a very micro level, you could argue—and this is why we actually have to go back to the micro—that if the disruption happens extraordinarily quickly and people don’t have time to adjust, then in the short term you will have some element of structural dislocation that will result in some form of, definitely some form of, recession or GDP slowdown if those folks can’t be digested into new jobs quickly. So, I do agree.

My point is this: in the short term, you can articulate a thesis like you’re just saying, right? If all the 45-year-old programmers are let go at the same time, and there are 6 million programmers on the street and there’s no other work for them, and it happens in a month, then in the short term there would be this GDP hit. While I’m still correct that over the medium and long term, GDP growth bails us all out.

So that’s why it does go back to the micro: do you think it’s all going to happen so quickly? Do we think that all these things are going to be displaced extraordinarily quickly?

Jason Lemkin

I don’t know. When I look at Andrej Karpathy talking about the evolution of how much of his work has gone to AI in the last 6 months, I do question it.

Okay. What do you mean by that?

Jason Lemkin

I think more and more of labor will be replaced by AI. We will see the concentration of value among fewer people, and fewer consumers will have money to spend in the economy, which will lead to problems and a shrinkage of that economy.

There are 150-odd million people working in the US, right? So, what’s your estimate for displacement? I’m just trying to get a sense of it.

Jason Lemkin

It depends on the time horizon, but I’m not feeling that great about it. If we look at the most obvious areas, which are customer support, legal, and bookkeeping, that doesn’t look great. If we want to add in Waymo and what it’ll do for self-driving within a 4-year period, gosh, I think you could see 30 to 40 million.

I think, again, Waymo was a good example. You look at the projections 10 years ago when people talked about self-driving, and you can find all the Citrini articles from then saying it’s all going to happen in 4 years; it’s over. Here we are today, and even though I think Waymo is amazing, they’re doing $350 million. I think they have single-digit thousands of automobiles in a few cities, right?

Steady rollout has started to increase, probably. I don’t know how many years you’re talking about before it gets to further mass scale. So, it’s a very long—almost every diffusion takes longer than you think, right? I think we massively overestimate the pace of adoption here, and you massively underestimate humans.

Does it ever take shorter? Because everyone always says that about the pace of diffusion, and they use the Industrial Revolution, where you had to buy machinery, transport it, and train people. When it’s Nano Banana Pro and it removes an entire industry—well, I think it’s shorter, much quicker.

Jason Lemkin

In my life. It’s my life. And the thing that is stressful about this—and it’s a way to make money, as in, the stressful thing—is that, again, I don’t want to endlessly talk about vibe-coding apps, but I have so much experience. I’m shocked everything is faster than I would have ever expected and better than I would have ever expected.

We started this about security, right? If you told me when we started this podcast that today I could just talk to a vibe-coding platform and it would do an A-tier security audit while I was on an airplane and I didn’t have to do anything, I would have said, “Guys, it deleted my whole database. There’s no way it’s going to do a freaking enterprise-grade security audit.” Yet here we are.

And then we turn around, and everything Anthropic can do this year is faster and better and bigger than we thought. I’m not saying at some meta level everything isn’t slower than we think, right? Certainly that dumb article about DoorDash being disrupted by Base44 next week is dumb, but God, this acceleration—it’s just, it’s hard to— it’s so fast.

In practical terms, I’ll give you the practical ramification. Almost all the B2B software we use today is terrible now. It’s terrible. I can’t talk to it. I can barely bring myself to use WordPress; I can’t change anything in WordPress. All these products are terrible.

A lot of the ones that the founders we love—the founders Harry talks to—the products are terrible now because AI software is so good. Blow your brains out to input data for 2 hours into your system, right? It’s terrible. And if nothing else, that is accelerating so quickly that the leaders cannot keep up with the fact that their products are so dated. They’re so dated.

Rory O’Driscoll

I think that’s broadly true. And I think when you narrow it—that’s why I said when you narrow it down to the micro—the impact of AI on the 1 million to 1.5 million workers in the US, in the software and tech industry broadly defined, could be way more disruptive, right? And I think that’s actually a useful conversation.

But it’s worth pointing out that, plus or minus, 5% to 7% of all jobs in the US are disrupted every year, right? And this is 1%—less than 1%, maybe 1% to 1.5%—of all jobs if the entire software industry got nuked, right? It still wouldn’t be the same as losing the car industry 10 or 15 years ago.

So my point is, again, differentiate the micro discussion of: Are B2B software companies in trouble, and how much are they in trouble? That’s a really good discussion, and if you make your living investing in B2B software companies, it’s the only thing that matters, right?

But jumping from there to saying civilization as we know it ended is, just as you say, Jason, clickbait, and we should just ignore it, right? The truth—I mean, I’m going to be even more direct—

Jason Lemkin

We didn't bleed in Silicon Valley when the car industry went down the toilet. Don't hold your breath thinking they're going to come for us and say, if the only thing that's impacted here is the B2B software industry, my suspicion is the rest of the world will go, “Yeah, I'm willing to lose those guys.”

So I think we discard the macro, but then I think you're right. The question is who wins and who loses in the 2% of GDP software business with AI. I think that is a valid—that is the question. You're right. So do you think all that software is crap and looks outdated today?

That's what you're saying. Also, one other point—and we talked about Tobi Lütke from Shopify before—is about the best of the best, right? I think Shopify is on top of these things. It has the same number of employees it did 3 years ago. It has not added a single net headcount in 3 years and has grown 50% to $12 billion in revenue.

At a meta level, that is a decline, too. It has grown its revenue 50% to $12 billion and not added a single net headcount. That already is an economic loss to the tech lifestyle we lived just a couple of years ago, right?

I was literally talking with a group of B2B CEOs at scale the other day, and I made a statement that everyone thought was a joke at first, until they thought about it. I said, “One of the leaders in the next 12 months is going to do an Elon Musk and just cut half their team in 1 day. They're going to lay off half the entire company.”

Elon—we thought Elon was crazy at X, but it stayed up. One of these folks that is not making a big transition to the new world is going to realize they can just go from 4,000 to 2,000 employees and be fine, right? This trend could accelerate, even if Shopify is the same for 3 years.

Anyhow, listen, just for fun, then we can break. I asked Claude—who isn't that crazy guy on Twitter trying to get views—to make my point: What would happen if, with AI, we were able to reduce tech headcount by 50%?

Claude said, the short version: $600 billion to $900 billion in GDP impact; 4 to 5 million total jobs lost, including all multiplier effects; and local economic devastation in 5 to 6 cities where tech is concentrated. It would be one of the largest economic shocks in U.S. history outside of a world war or pandemic.

I'm not saying it's true, but I just asked Claude to parse the data if we lost half our headcount. And even in some ways—now, here's where I'm not smart enough—even not growing headcount for a decade is losing headcount, too, right? Because our revenue will grow so high, right?

But go back to your thing. I think you're right about one thing. I think the number of companies that will do that kind of dramatic headcount reduction, and the most obvious place, will be all the levered, PE-backed, highly levered SaaS companies.

Because if you're not levered, if you're public, if you have time, you can do the Tobi playbook and just hold headcount flat and rely on growth. You don't have to do traumatic surgery; you'll just be able to become steadily more efficient.

If you owe 6 times EBITDA on debt and you bought the thing for 8 or 9 times and now it's trading at 4 times, you have to start paying down that debt, and there's no growth. So I think a lot of those guys will look at very draconian expense-management plans.

7. Is Tech Private Equity and Thoma Bravo F***** in this Market

I'm not picking on—I didn't mean to pick on that, but I'm using them as an example. So, please, your answer is not dependent on Coupa and Anaplan, but that ilk. You're growing at 20% with no founder CEO. So, a hired CEO, growing 20%—what do you do?

I think Rory's making a more brutal and cogent point, which is there are folks like that that are growing 6% and have massive debt leverage. They have no choice but to shrink, right? The math doesn't solve any other way.

At 20%, you at least have options. You can probably do the Tobi playbook. If you're north of 20%, probably the math barely pencils out at that level, right? That compounds to enough growth over 4 to 5 years to pay off your debt. But in the single digits, all these Blue Owl and friends, they're just upside down on this debt. They can't pay it off.

What happens?

Guest

I think if the debt is upside down, hence the Blue Owl thing, then the equity is gone. A lot of these folks—they're not just going to roll over and die. They're going to do exactly what Jason said. They're going to try and cut expenses and knuckle down.

The debtors will extend the debt because they won't want to crystallize the loss. The equity will run the business and try hard to pay down debt. They will reduce headcount dramatically. They'll be a much less attractive place to work. They'll be much less attractive vendors to supply to because they won't be investing in R&D.

But, and this is the sad comment, I think the inertia in a lot of software contracts is such that it'll take a long time to die. So it'll just be a steady, nasty grind, right? I don't think it'll be the cataclysmic, all-going-wrong-on-Friday kind of thing. I think it'll just be a long 5-year grind.

Jason Lemkin

But, yeah, my guess—I don't mean to go off topic—is we will end up seeing more and more of 5 to 6 to 8 startups at $50 million to $100 million to $200 million in revenue mashed together at nominal prices of 2 times revenue or less. They'll just all be mashed up.

No one knows it's the best idea. Everyone knows that the 11 companies Clari has bought do not have perfect synergies, right? But it's the best idea we have for these companies at 2 times—1.5, I know it's probably not even an ARR multiple, but 1.5 to 2 times revenue.

People will just say, “You know, is enough already. It's time.” At 1 to 2 times revenue, they will capitulate, and you'll see these sort of Frankenstein B2B companies that have their many constellations or whatever. But it's the best play left, right?

You'll bring in professional management that will have 20 products, right? We'll see 20 unicorns merge into one thing that will IPO in 2027.

Guest

Agreed. That—except the last sentence. They'll try an IPO in 2027. But you're right, Jason. It is. I think that's a great idea—it's the best idea they've got.

You've got this $50 million revenue thing with no acquirer. Maybe if you had 5 $50 million revenue things in roughly the same market, you could build a $250 million thing that would be profitable enough and/or get some growth at scale.

It'll be a miserable way to consolidate all that debt. But, yeah, that's the—I mean, those are going to be the most impacted by this kind of decline in the perceived value of recurring-revenue companies.

We can either talk about Figma's earnings or OpenAI's spending increases next. Which one do you want to do?

Jason Lemkin

I don't know. I want to know from Rory when he thinks the OpenAI gravy train ends. I mean, what is it—another $110 billion? We've all had this board meeting, guys. You walk in: There's good news and bad news. The good news is we're making up more hardware and other revenue 3 years out. The bad news is I need another $110 billion to get there.

8. OpenAI Massively Increases Spending Plans: Analysis

Okay, let's ground it in numbers then. OpenAI is doubling spend to $66.5 billion by 2030, but they're upping their revenue forecast by 27% to $280 billion based on products that mostly don't exist today: hardware, ads, everything else.

Guest

At one level, it's been interesting just watching the conversation that Claude now gets the benefit of the doubt and people believe it can kill everything. OpenAI, which was the darling, now gets no benefit of the doubt and almost looks like it's a clown show.

I think the truth is probably no one is ever as good or as bad as they seem, right? OpenAI is still the winner. I think it's pretty straightforward. OpenAI is still the clear winner in the consumer space in terms of mindshare.

I think what they're grappling with is how ambitious they can be in these capital markets, and they clearly still want to be ambitious and they're still getting funded to be ambitious. But you're right: When you run the math out, it's the consumer product that does exist. Then you add on to it the enterprise product that exists but is not doing as well as Anthropic.

Then you add on to that agentic products that don't yet exist. Then you add on that, quote-unquote, consumer monetization beyond subscription, of $30 billion to $77 billion, which is basically ads and other stuff.

What you recognize is that you're spending a large amount of current real dollars that NVIDIA and all these other people think they're going to collect in anticipation of a whole bunch of future anticipated dollars, which are entirely credible to believe. But there's a lot of leaning into the future here.

So far, as long as people believe some version of—going back to the Citrini memo—as long as these companies are perceived as so powerful that they can destroy everything, they will be able to get money. Because the truth is, if you believe you're the thing that can kill everyone else, then the only rational response as an investor is, “Oh my God, if the models are going to take over the world, I better get me some models.”

All this fear-mongering is good for both of them. Do you think Sam Altman is right to be as aggressive as he is being on all fronts—hardware, ads, discovery, Codex, health—or would it be a better play to be much more focused, in the Dario Amodei mindset, on enterprise coding? He's expanding more and more, as we're seeing, but it's still enterprise-focused and more specific.

Jason had said this, but just pragmatically, you've got to say that if you were 10 times ahead of a competitor and now you're only 3 times ahead of a competitor, at some level you did allow them to gain market share, and the objective of the game is to beat the other guy.

So, you’ve allowed the other guy some room. You’d have to say it wasn’t the right play this year or last year, right, to allow Anthropic so much space in the enterprise, and maybe doing all the things you were trying to do took your eye off the ball on winning the 2 or 3 things that you must win.

Now, you’ve also got to say that this team and this man have created the single most valuable, exciting AI company on planet Earth. So, I’m not going to sit here as a little minnow criticizing him. But what you can say pragmatically is that 2023 and 2024 were good years for OpenAI, and 2025 was a good year for Anthropic. If it’s a 10-year race—2 years up, 1 year down—bring on the next year.

Rory, don’t put yourself down. You’re a GP at Scale. Okay, you can take Sam. Fuck that.

Jason Lemkin

I do think—I’m confident, going to Rory’s point—I’m confident that Sam’s got one of the best benches in the world, right? So, I’m confident the bets they’re making are the right bets to make today, right? Given the plan, given the fact that they’ve committed to an insane growth number, that they need to raise an additional $110 billion of capital, I’m confident it is the right plan.

Also, honestly, if it doesn’t work out, he can dump the hardware business in a heartbeat. He can dump all this stuff. He can dump anything if it doesn’t work, and they will ruthlessly dump it if it doesn’t work. They’re not committing everything upfront today.

So, you’ve got to pick your 3 to 5 best bets at the start of the year. Maybe, in the age of AI, you’ve got to change them every week, but these are the best bets. The funny thing is—and granted, the OpenAI slides were leaked to The Information, so they weren’t presented publicly—it just feels so much like a startup board meeting you see where you walk in at the start of the year and someone’s got these things. There’s a stacked chart that looks beautiful, but 3 of the colors have never been done yet. They’re aspirational.

And so the good news is, “Rory and Harry, we’re raising our forecast 30%, and we just need another $80 billion to do it.” It just felt like that on steroids. But I think those slides were for the believers, right? So the believers, the SoftBank and others, will give him more money. I think that’s the point, because you’re not going to believe that the revenue from hardware and ads and research are real unless you want to believe. If you want to believe, you’ll believe it, right? If you’re a skeptic, you’re going to take your marker out and just delete those bars. But, man, it felt like a couple of board meetings I was in in December.

Yes, except, as you point out, with 3 zeros attached to every number.

Jason Lemkin

We just need another $111 billion.

9. Figma Fights Back: Earnings Through the Roof

You’ve got to say these are the 2 fastest-growing companies in history. That’s just a statement of fact. Jason, you said last week that, essentially, Lovable and Replit’s rise was a fault line for Figma, and their ability to do, you know, $300 million to $400 million was a take from Figma.

Jason Lemkin

Maybe just a theoretically missed opportunity in their sweet spot. Just theoretically missed. Yeah.

Okay, a theoretical miss. Figma came out with its Q4 2025 earnings. They were very good: accelerating growth at $1.22 billion ARR, growing 40% year over year versus 38% year over year in Q3. They also have amazing retention of $10,000-plus customers: 97% GRR and 136% NDR. This was great. This was a great quarter for Figma all around, and the stock was up 15% after earnings.

How do we see these numbers? Talk me through this. This is a company whose leader is clearly saying, “I know what I’ve got to do. I’m not in any doubt. I’ve got to add AI capabilities. I’ve got to go from design all the way to coding, and I’ve got to add that and make it happen.” And they seem to be getting decent adoption of it, right?

This is what winning looks like in SaaS in a sector where it’s very credible—very credible, perhaps more credible than most—that an AI-native product could disrupt you, right? And I would argue something like accounting, you’re 5 years away from AI-native being disruptive. In something like design and codegen, creative and codegen, it’s here right now.

10. Momentum Versus Value: Four Public Stocks to Buy

So, this is what fighting back looks like with a generational-talent entrepreneur, a company still at its peak, i.e., not stale, and clearly trying to punch back. I said, well, if this was a boxing lineup in Vegas, this would be one of the marquee headline events: in the right corner, Dylan Field, heavyweight champion of the world; in the left corner, Lovable and Replit, with ringman Jason Lemkin. So, what do you think, Jason? How’s the fight going?

Jason Lemkin

Look, I think it’s tough because it was a great quarter. There’s nothing not to love in the quarter. It’s an epic company, an epic quarter. We’ve just given up on the present. We’re all panicked about the future, and you get no credit for a great quarter, right? You get no credit.

I read one analysis of monday.com, who we all love, and the criticism of monday.com was, “You’re constantly beating and then lowering.” We want you to constantly be beating everything. We’re looking for 10 quarters of AI dominance, and so it’s just tough.

Even on Figma, I guess, at a meta level, it’s a race. Listen, I’m not as big an expert in Figma as I am, say, in DocuSign, but there’s a similarity between DocuSign and Adobe Sign and Figma, which is that their products are much more workflow-oriented, they’re much more systems of collaboration, and they’re much less just about getting pixel-perfect designs created.

But I would just be shocked if, at some point in the next 8 to 18 months, Claude Code can’t automatically make designs that are as elegant and beautiful as a designer can, because you already can with some custom LLMs for images. You can build epic images. The delta is that you could literally mock a website that Claude Code would build from a design perspective. I find it hard to believe that, at the end of the year, we’re going to mock it. I find it hard to believe that every GPT on planet Earth can’t create genuinely custom, artistic, artisanal, beautiful designs. I would be shocked if that’s not possible by the end of the year.

And what do you think that means for Figma?

Jason Lemkin

Exactly. I mean, that’s why it’s hard to be bullish on anything from the past right now. It’s hard to be bullish. I know it’s the job, but, man, it’s stressful. I would love to know how stressed Dylan is. Maybe it’s a 2, but I’d be stressed. Most of the CEOs I talk to are pretty stressed, right? Some of the executive teams aren’t. They’re checked out, but most of the CEOs are stressed.

I sit in the investment committee this week and I said, “Well, there are 2 areas where I’m finding value in security. One is where there are deep integrations and partnerships, which are difficult and the AI companies will never do. We’re an investor in Airwallex. We’ve discussed it before. They have hundreds of banking relationships and partnerships with Southeast Asian and Indonesian banks that no one—not Anthropic or OpenAI—will ever have.”

And then the second is where there are deeply technical, complex coordination challenges, like Fuse Energy, another one of our investments, which owns everything end to end, from energy supply and generation to delivering it to a consumer’s home. So, that’s where I put my bets. There are still areas of secure value.

Rory O’Driscoll

I think you’re talking past each other because, of course, you’re right, Harry. The 2 examples you cited are a financial company and a physical power company, and even the most deluded AI believer doesn’t think that AI is going to take over fintech or energy. So, I agree: you can invest comfortably in those 2 areas and not have to spend a ton of time thinking about AI.

When Jason says it’s hard to be comfortable with anything from the past, perhaps the more precise version of that is the core question: it’s hard to be comfortable with anything from the past in core B2B software, or maybe in core software.

I mean, it’s funny. When Adobe tried to buy it, at the time it seemed like they were overpaying because it was a big deal at the time, right? But at least they were buying a 20-year business. Scott Belsky was paying twice what anyone else would pay to get shit done. Kudos to him, actually. Kudos to him for putting his job on the line. Overpaying—but so what? You were buying a 20-year position like Creative Cloud and Photoshop.

You fast-forward today: Figma says one of its biggest growth drivers is integration with Claude Code, and Figma made it natively integrated. And if you go into Replit—and I’m sure it’s true in Lovable—you can natively integrate Figma too. But it’s like Fortnite. What if the native integration just overlaps more and more throughout the rest of the year?

And then sometimes I’ll integrate Claude Code and Figma, but some other times I’ll just have Claude Code do the whole design. It’s getting pretty good. It’s pretty good. It can design it. It has ingested every single website and mobile app on planet Earth. It can reproduce an iterative version of that that’s just as good. Why can’t it? It’s only because they haven’t focused on—

Jason Lemkin

I think I'm only going to bet on the winners right now. I'm going to do the Andreessen version, the Thrive version, and Figma may make it, because I'm looking at what stocks are actually up.

Okay. The ones you might not want to buy are basically Palantir, Figma, AppLovin, Cloudflare, and Shopify. Those are the only ones that are up over a year. Over a year, those are the only 5 that are up over a year, right? You could find other esoteric ones, but these are the core ones. Deal. Those are probably the ones I'm going to buy. So, I don't know, maybe 18 months ago I decided I was going to go bargain shopping on public stocks.

But that seems paradoxical. You said that you could see Claude being able to make and design in the same way that many are using Figma today, but then you would—

Jason Lemkin

Yeah, because there's so much uncertainty. I'm going to be a momentum investor here. There's just so much uncertainty. I'm going to bet on the ones that are winning because I believe success will beget success, that the best people will continue to grow. In an age of uncertainty, I don't want to bet on the ones that are down. I mean, GitLab's a great company; it's down 59.62% over the last year. I don't want to make that bet. That's for somebody else. I've already made those bets in the past; I lost money on all of them.

I just don't want non-founder-led companies.

Jason Lemkin

I know, but I'm going momentum.

Rory O’Driscoll

I actually think, Harry, Jason's answer was utterly coherent and, in a sense, utterly logical. You're exactly right: I see what you're doing, Jason. It's a momentum play, right? All the data says—and I always struggle with this—in the short term, over 6 to 18 months in a public market, momentum plays work; value plays don't. Over a 5-year period, value plays work and the advantage of momentum goes away.

A whole bunch of people have published on this, right? Of course, the trick is to figure out when you're transitioning from one to the other. You're effectively taking that risk. You're basically calling the market right now a momentum market. That's what you're doing when you do this, and so far, you've been right, to be clear.

Jason Lemkin

Well, at least I'm saying that when everything is in a state of flux, I'm going to bet on whoever has the gravitas. Momentum is gravitas because, at least for a little while, momentum builds on itself. This isn't just fake hype of marking up a round. This is real momentum.

So, what about this? Because I wrestle with it—we're getting to genuine, actual investing, because it's actually been the right play, right? It's been the right play in the public markets; it's been the right play in the private markets. The companies that had good rounds have good follow-on rounds, right?

You have to ask yourself: logically, that's not extensible forever, right? Because trees don't go to the sky, as they say on Wall Street. Let's take Palantir: a massive momentum play, and yet over the last 6 months—or the last 3 months, anyway—it's been hit, not quite as hard as the boring-ass SaaS stocks, but it's been stomped pretty effectively over the last 3 months, correct?

Jason Lemkin

Yeah, it's down 27%.

How do you respond to that information? If you're a momentum player, how do you respond to it? Something that's trading at 40 or 50 times revenue—I mean, it's down from 70 to 46 times revenue—has probably got risk of blowup in it. How do you think about the risk of blowup versus wanting to still be in the momentum play?

Jason Lemkin

To make it really practical, I decided that, for me, a year is the right measurement point. There are all different time frames: year to date, 3 months, 1 week, 1 hour. I feel like a year shows the durability of this, at least looking backward. I'm going to pick from those 5. It would've been different last week, when I would've picked favorites and tried to find dislocations.

If I look at the public stocks, Klaviyo versus Shopify, Klaviyo is essentially a derivative of Shopify. It's essentially just a derivative, with almost 100% of the revenue attached. Yet in the last year, Shopify is still up at least 63%, but Klaviyo is down 58%. So, if I'm a bargain hunter, I'm going to go to Klaviyo. I'm not going to buy Shopify; I'm going to buy Klaviyo. It's got to be a bargain.

Interesting.

Rory O’Driscoll

No, I think, actually, funny enough, because I often tend toward value. But in this case, I would argue your instinct is correct, because, bringing it back to the CEO of Shopify, if Shopify is going to thrive—and that guy looks like he wants to thrive—he has to build agents on top of his stuff. Fundamentally, he has to take the market cap that has currently accreted to Klaviyo onto Shopify. He has to take that business.

And so I actually think you're right in that case. Again, I'm genuinely trying to seek knowledge here, right? I think in that case, the bargain hunt would be a mistake, because the adjacent competitor—namely Shopify, which is less dislocated than Klaviyo—probably has to kill you in order to survive. Provided they execute well, they're going to take some of your revenue.

Okay, what about this one? Palantir last year, even with its ups and downs, was up 41% in the last 12 months.

Jason Lemkin

Yeah.

The worst performer—one of our personal favorites—Atlassian, worst performer over the last year, down 74.85%. So, with just those 2 data points, which one? Atlassian doesn't have the direct disruption risk that Klaviyo has with Shopify, right? I can still argue it's oversold—Klaviyo. I can argue it's oversold despite that disruption risk—but minus 74% versus plus 41%, this is the greatest bargain. The delta is over 100%.

This is actually very illustrative, and it's something I've been thinking about a lot. What you're doing here is momentum investing. You are just looking at the price action independently of the valuation and saying, “Do I go with the momentum?” You're not saying, “Is there a price at which you own Atlassian?” By the way, this is not a criticism. I think in the short term, momentum has been the only play that worked, both in the publics and the privates. I believe we've been in a momentum market on the private side, also. It's too hard to find new shit; just find the shit that's working and pile on, because the next round is going to be in 6 months and it's going to be 2× the last round, independent of value.

To be very clear, I think that it's not a criticism; it's a positive comment. But if you said to me, “Should you buy the stock that's up 44% or the stock that's down 71%?” If I'm a momentum player, the answer is buy the 44%. If I'm a value player, the correct answer is, “I don't know until I see the numbers.” Then I've got to look at the value relative to growth.

Jason Lemkin

Well, Atlassian's accelerating. A year ago, it was growing 20%; now it's growing 23% at $6.3 billion in revenue. It's accelerating, and yet it's the biggest decliner of the entire group. This is not down to single-digit growth.

That's my point. Therefore, I might argue, perhaps wrongly—and, again, I'm perhaps wrong—that you'd start to nibble at that one versus being afraid to catch the one at 43× revenue, even though it's had the momentum. But I'm also humble enough to say the momentum play is what's worked.

Jason Lemkin

Yeah, that's why this is hard. But I think what you see in where we started the conversation is the problem with perfectly priced momentum plays: it doesn't take a lot to knock you off-kilter.

God, I need to get thinking about these picks. Fuck me.

Jason Lemkin

Yeah.

Jesus. Okay. I love Mike, by the way. I think he's awesome.

Rory O’Driscoll

Yeah. I mean, as armchair value investors, if we're abstracting away from everything else, this is the best pick on the list. Of the ones that have been beaten down the most, it's the one that's accelerating. None of the rest are accelerating.

So, if you had to be simplistic, what's the greatest dislocation in the market? The stock price is down the most, but with the most revenue acceleration that's still above the fold, per our conversation. As armchair value investors, you couldn't do better than Atlassian. You literally could not find anything better that's accelerating and is beaten up.

Also, if your concern is revenue durability, the increasing portion of multi-year contracts with large enterprises relatively answers a lot of those concerns.

Yeah. If there's any humans left to buy the product, of course, but you're right.

Rory O’Driscoll

Of course. Of course.

11. Jack Altman Joins Benchmark Capital

Listen, I want to finish on one, which I think is very interesting from a venture perspective. We saw Jack Altman, who raised $275 million last year for Alt Capital. He's a wonderful dude. I'm sure everyone loves Jack. No one in the Valley dislikes Jack.

He made the move to Benchmark—a very big move—leaving his firm, or kind of shutting shop on his firm, to join the great GPs that are at Benchmark today. How do we analyze this move? Is this symbolic of the further consolidation of venture? What did we think about this move? It was a big surprise to the ecosystem.

Jason Lemkin

I think it was a clever move by a very shrewd firm. I think it's been their MO for 15 to 20 years: we have a very compelling offering to make to any GP. We'll make you broadly equal in a very successful partnership with a lot of autonomy, and therefore you can have your pick of proven talent.

Rory O’Driscoll

You're not in the growing-talent business. You're in the picking-talent business, and in general, you can make people a compelling offer that most people are inclined to take, right? This is just an extreme version of that.

Jason Lemkin

Yeah. To me, the more interesting thing at a meta level is, if he really took his last fund—he raised like $400 million in 2 years—and if he took his last $250 million fund as, in essence, a solo GP, forget how it's structured. It's him. It's all him. And he gave it all back to the LPs. That is not a minor give.

Even if you're made whole, because I was offered to be made whole a couple of times in the old days, it's made whole with asterisks and daggers. You have to stay. You have to deliver. And maybe it's made whole no matter what the world brings.

So if you're confident you can triple $250 million and retain massive economics in it in terms of carry and fees, it's a— The interesting thing is, there are so many GPs who would love to have what Jack had. There are so many folks stuck at venture firms, stuck working for people, and they're like, "My God, if I could have $400 million in 2 years to invest in whatever the fuck I want." And Jack gave up the dream of 95% of folks stuck in venture.

Rory O’Driscoll

I think, again, Jason, you're on a roll at the moment. That is the other interesting point: most people are swimming the other way, which, by the way, is an implicit and embedded compliment to Benchmark.

Jason Lemkin

Yeah, I think it might be more than that, though, because you have to ask yourself why he would do that. But I remember in the old days, and I'm sure Harry got similar offers in those days—maybe more—but in the old days, I got an offer from a mega-firm that I didn't even know, because I had just raised my first fund. They said, "Well, we'll just—just come here, do SaaS. We'll make you whole. We'll make you everything."

And they're like, "Honestly, you've only raised $70 million for your first one." That's like nothing. That's like, "We'll just—we can guarantee you you're going to make $10 million off every exit we have, and we'll pay you $2 to $3 million a year, and you don't have to do any—why? We'll make you whole." But I'm like, "Who do I have to work for? I didn't sell my last company to go work for somebody." It wasn't even a 10-minute conversation. I couldn't even get it to work in my head.

Rory O’Driscoll

Yeah. Jason has enough self-knowledge to know he is destined never to work for another human being again, and humanity is graceful.

Jason Lemkin

For someone that I'm in love with—I mean, not personally. I mean, if I so respect the CEO and they would let me work in my box, it's not true. I would do it, but it has to be in that special situation, right, for someone that you just respect so much.

Rory O’Driscoll

But, so, he gave up a lot. That's just the interesting thing: he gave up a lot, and it was worth it to him to be part of this entity, this brand. I just think—I don't know what it says, but it says a lot about 2026, that you would give up the dream of 95% of venture, that you would give it up, right?

Worst case—you know, I mean, not to get tacky on this pod—but even as a solo GP, even if you just manage the fees on $400 million, it's not a terrible lifestyle. You can still afford to eat at a pretty good restaurant from time to time. You might even be able to rent an apartment in Harry's building.

I think people that successful aren't necessarily into minimizing the downside. I mean, I'm sure the attraction is to work with a great group of people and build a great fund. And yeah, you're right: I still wonder. Many people would say, "I'd prefer to be on my own," especially if you've already gotten there on your own. But it's a compelling offer.

Jason, would you leave SaaS to do a $500 million fund with me and Rory?

Jason Lemkin

I wouldn't. I wouldn't do it because I don't think I would be successful.

Why?

Jason Lemkin

Look, everybody—this is a very niche industry, right? This is as niche as it gets, right?

I could paint a picture to you, Jason, that this could actually be the next greatest 5 years of your investing career. The insights that you have as an investor today, because of your proximity to it, make you better than ever as an investor. I would argue—and I think Rory would probably agree with it—

Jason Lemkin

It should be. I think you're right. It should be. It should be. Whether it will be remains to be seen, right? It should be, right?

And so you should be more aggressive than ever, not less, would be my argument to you.

Jason Lemkin

I just don't know if I could sit in Monday partner meetings again for 4 hours. I don't think I could. I see you out there, Harry, on the road with all your portfolio companies on LinkedIn and celebrating their Series A. I'm the most loyal person to the founders I invest in, but that's not me anymore. I've done that. I'm just not—I can't do it.

I told my LPs and co-investors I'm not doing any more AGMs. I'm not standing up there with, "Here's the numbers." I'll go talk to you. I'll drop by your office, but I'm done with this performative, all-day circus of an AGM. Not that I think there's anything wrong with it—I actually think AGMs are very important—I'm just not going. I'm just not doing one again. If you want to work with me, as long as I don't have to go to the AGM, maybe it's okay.

But if you're going to do something like this, you have to be sure, especially if the person has a perspective, or you have to be able to leverage their strengths and backfill their weaknesses, right? So my particular strengths are deeply knowing everything around this agentic go-to-market and having a large group of founders that I've helped who believe in me. People do trust me, as Harry knows; he just had a conversation about it, right?

But I'm not good at some other stuff, and most venture—kind of going to Harry's point about Benchmark—should be 5 flat partners. There's this certain genericism of most of venture that, if you don't fit into those things, you might not thrive at different entities, right? There's still a little bit of solo hunting and meeting on Monday, and weird, consensus-driven outcomes where everyone's not that happy about Rory's or Jason's deal, but I have to do it because Harry wants to do his deal.

There's just a niche thing. And I think if you want certain people who are talented, you have to let them do their thing and nothing else. That's the key. And I think in venture it's harder to do that than at Anthropic, to tie it all together. In Anthropic, they're going to find your niche if you're off the charts, right? Although even there, I wonder.