“Cursor 已死”纯属胡扯：原因在这里｜Miles Clements

Miles, we are in person. I love it when you're in town. It's so lovely to see you, man, and it makes it so much more special doing it in person. Thank you for joining me.

1. Where is True Alpha and Value in a World of AI

Yeah, thanks for having me. It's always fun being here.

I want to start with the core question that I think every investor is thinking about, which is: How do we ascertain true value in an AI world where technology seems so transient and revenue seems so uncertain?

I think in terms of evaluating these AI categories and companies, there's a pretty useful framework, which is basically trying to understand a company's time to value and then the durability of that value. A number of these companies shine on different dimensions.

If I were to look at legal AI, accounting AI, or a company like Basis that we just invested in, I actually think these companies don't have an immediately quick time to value. When you look at the deployment cycle and getting lawyers and accountants sold on the technology, that can take a little while. But once it is hooked, the durability of that value is transformational to these firms.

On the other end of the spectrum, I would take some of the very early vibe-coding companies. Very quick time to value: You start vibe coding, and all of a sudden you have a weekend-warrior pickleball app ready to go overnight. You can start using something very quickly, but the bottom just fell out for a lot of these apps because there was no durability of value.

The reason I think coding has become the vertical in AI is because it shines on both dimensions. You can start using Cursor in an afternoon, and by that evening you're 10 times more productive. The time to value is very short, and then the durability of that value compounds as the team starts using it.

Claude Code, Cursor, and all of the great products out there—I think this is why coding has become the vertical that is the battleground in AI today.

2. Why it is Total BS that Cursor is Dead

Jerry Murdoch from Insight said on the show the other day—not me, but something I overheard from my portfolio—“No one's using Cursor anymore. Everyone's using Claude Code.” We just saw Chamath tweet last night, “We're going to have to move off Cursor because it's simply too expensive.” The Twitter sphere seems to have turned against Cursor with the “Cursor is dead” meme. But then they hit $2 billion in ARR. I'm trying to understand what is going on here.

I think there are a couple of things at play, and I saw the Chamath tweet and listened to the Jerry Murdock show. With all due respect to those guys, I think there are a few things at play.

First of all, this market is growing enormously, and I don't think a lot of these companies are actually experiencing success at the expense of the others. Take Claude Code as an example. First of all, what an amazing product, and how lucky are we to live in this technology cycle where we have all of these tools at our disposal?

Claude Code has absolutely captured the imagination, in part driven by Opus 4.5 and Opus 4.6. I think the success of Claude Code is also very much tethered to the success of the underlying model, so it has captured the zeitgeist. That's unmistakable. With that said, I think these things are so market-expansionary that it's not necessarily coming at Cursor's expense.

I think they're market-expansionary on 2 dimensions. First of all, they're bringing so many new cohorts of users online. People who would not have been software developers a year ago today can be software developers with these tools. They're also expanding the market in terms of consumption.

When you look at the ARR growth leaked for both companies, a lot of that is not net-new companies paying per-seat pricing. A lot of that ARR is consumption, which is off the charts for both tools. I think that's one thing that's going on.

I think another is this misunderstanding about Cursor being tied to the IDE. In some ways, Cursor is a victim of its own success. They were so disruptive and so innovative around the IDE a year ago that people can't help but overmake that assumption.

What is happening, though, very clearly, is that the world is moving to agents. No one has been more vocal and thoughtful about that than Michael [surname unclear] from Cursor. I just look at the numbers. According to Michael's post, which was public on Twitter a few weeks ago, there are 2 times more people using agents in Cursor than using the tab feature. Ninety percent of Cursor users are daily active users of the agent product.

The agent product grew 15 times last year. The cloud-agent product, which was new as of October 30, is now responsible for 35% of merged PRs in Cursor. Those are happening by cloud agents.

With all due respect to Jerry Murdock, I think he said, “I thought about these metrics, and this company needs to pivot.” All due respect, I thought about playing in the NFL, but instead I walked onto a college football team and was the fifth-string inside linebacker. You're not looking at any real metrics. Who are these people to make these judgments?

I get a little spun up about it. The thing that's so cool about the team is that they are focused, they are unfazed, and they're just building.

Do you think they are fundamentally challenged because of their reliance on Anthropic in their models and what that does in terms of cost inflation for end users of Cursor?

I don't think so. In a number of dimensions, the beauty of Cursor is their ability to be multimodel. I think it's valuable for a couple of reasons.

First of all, we put this survey into the market. You'll have to have me back on the show to give you the full readout because it's only 90% of the way complete. We just wanted ground truth on what's going on with the mindset of developers today.

One of the things that we're learning is that 50% of developers switch model families on a daily basis, and 95% of developers switch models on a daily basis. I think the world wants to be multimodel, and that experience is fundamentally enabled by Cursor.

3. Why Cursor Were Not Wrong to Build Their Own Models

The other thing that comes from being multimodel is that you basically become an index of AI innovation. You get this compounding product benefit where every new feature and every new enhancement that the Cursor team makes obviously improves the product experience, but every improvement with the underlying models also improves the capabilities of Cursor. You get this compounding product flywheel that's very unique.

Was Cursor wrong to focus on building its own models?

I don't think so. I think what they're going to be able to achieve is incredible. I also think we need to frame in the right context what their aspirations are with these models.

Generally speaking, there are generalists and there are specialists. Cursor is going to build specialized coding models that are going to serve specialized coding tasks, especially for a lot of enterprise users. They don't need their models to be good at poetry or to teach you how to make an apple pie. Their models are there for professional coders to do professional work, and I think that's very powerful and will continue to make the product experience really differentiated.

4. What is the Upside When Investing in Cursor at $27BN?

Can I ask you, when you're investing, what was the first-round price?

$9.5 billion.

Okay, $9.5 billion. When we're doing something like $9.5 billion and $27 billion, what are we underwriting it to? If I was your partner, I'd be like, “Totally get it, and this is super exciting, but what's the upside here?” How did you think about that?

I think there are a couple of ways to frame the upside. One is that you think about platform companies that are publicly traded and own their domains. There are very few of them out there.

Salesforce historically has been the go-to-market platform company. CrowdStrike and maybe Palo Alto are the platform cybersecurity companies. There has never been a platform company for engineering as a vertical. Engineers are the fastest-growing, most dynamic vertical there is, and no one has ever owned that.

You've had companies that have built tremendous value by biting off pieces of the stack. Atlassian is a hugely valuable company that we love. It began around issue tracking. Datadog began around monitoring. These have been $50 billion to $100 billion companies built over time, addressing one portion of the engineering product stack.

No one has built the platform company to own it all, and we think they have that aspiration. So that's one thing.

The other is that we were also joking before the show that I think sometimes getting overly fixated on the financial metrics in this environment can leave you with an unsatisfying taste in your mouth. This company is growing so quickly that, on a multiples basis, our first investment was at roughly 4 to 5 times year-end ARR.

That wasn't anything that we talked about or part of the underwrite, because now it's at $2 billion and you did it at $9 billion, essentially.

The company said, I think a week ago—or it was leaked—that they'd passed $2 billion. So, yeah, that's a fair assumption.

Yeah, I totally get it. What was it when you did it, just because you need to have that mental plasticity, give or take?

We had a conversation with Michael where we said, “Where is the business today, commercially?” He told us, and I'm not evading the answer—I don't specifically remember—but maybe it was like $100 million of ARR, give or take.

Yeah.

We said, “What do you think is realistic for the end of the year?” He said, “I think maybe our aspiration is, if these assumptions go right and these are the products we're going to launch, I think we can get to $500 million.” Andrew Braccia, who I was working with, and I sort of looked at each other and were like, “I think we should haircut it and call it $300 million.” Getting from $100 million to $300 million would be extraordinary for this kind of a company.

They ended last year somewhere in the billions, I think has been reported. But it really never was about—and still is not about—financial metrics. The financials of this company, to me, are purely a reflection of the product-market fit, and it's unlike anything I've ever seen.

When you're so off in your ability to predict revenue at year-end, how does that change your go-forward investor mindset? Do you just place no value on revenue predictions? How do you think about that?

I think revenue predictions are important in that they encode a lot of business assumptions. If we get this product right, if our pricing here is correct, if our penetration of this customer segment works out, we should be at this rough revenue scale. But the idea of having a budget so that you can hold the founders' feet to the fire quarter after quarter is just not really relevant.

To me, the less important thing is if a company finishes 10% below plan or 10% above plan. We're not public-market investors. We're not managing to earnings calls. We care a lot about the inputs that go into the assumptions, but the output is a little bit less important.

When we think about that and the revenue numbers that you see there, it makes other things seem quite boring. It does. This is the sad case. Are we in a world where triple-triple-double-double is dead when you can have a company like Cursor going from $100 million to $1 billion?

Absolutely not. Send me all of your triple-triple-double-double companies that you're not interested in investing in.

I was thinking this last night. Everyone says this on the show. I guarantee you'll be like, “Fuck no.”

No, no, no. Here's why. I think you can actually be successful in this market investing in consensus. And I think you can actually do really well investing in non-consensus. I think you get hammered sitting in the middle. So, a company that's not growing 15x year over year, that's fine. There are all these other really important inputs that go into it that I think can make for a really interesting investment outcome.

5. You Can Pick Any VC to Join Accel, Who Does Miles Choose…

I'm sorry, I still don't quite understand. If you have a pot of money and you can put it in companies that are growing 15x, why then put it in companies that are growing 3x, 3x, 2x, 2x? The opportunity cost of your cash is real. As your partner, I'd be saying, “Why are we doing that?”

Yeah, but this is where we're ignoring all of the other important inputs, right? Quality of the founder, what market are they in, what ownership are you getting in the investment—all of these other things factor in, too.

But I think one thing that's happened in our market is investors have tended to just flock to the extremes. Either we're AI maximalists—we're going to buy the basket, ownership and valuation be damned; we want everything—or we hate the valuations, they make no sense, and we're going to sit on our hands and wait until things cool off a little bit.

The reality is, the best funds in the world, the best investors in the world, embrace the nuance. The right answer is always somewhere in the middle, and constructing a basket of companies where maybe some were undisputed breakout leaders and you didn't get the ownership that you wanted, but you wanted to be a part of that company and you wanted to be partnered with that founder—there's room for that in a portfolio.

But there are also bootstrapped companies in Little Rock, Arkansas, where you can have a different ownership threshold and work with a really special founder and build the company in a different way, and you can do very well that way, too. We don't really run from the nuance; we embrace the nuance, and there's a lot of benefit to being a multi-stage, multi-strategy firm.

That's wonderful. But your funds are too big to embrace nuance, dude.

How so?

6. Do Sub $10BN Outcomes Even Matter to a Fund the Size of Accel?

I'm sorry. You need to have $50 billion-plus companies to return your fund sizes?

I think we will. Think about this: a decade ago, how many trillion-dollar companies were there in the world?

No, and you're right. I use this stat, too, and the expansion of outcome sizes, but, dude, they're very, very rare and they take 17 to 20 years when you look at the majority of them.

7. Why Would Any Founder Go Public Today

But this is the cycle repeating itself. To answer my own rhetorical question, which nobody asked me to do: a decade ago, there were 0 companies worth a trillion dollars. 5 years later, there were 6 public companies worth a trillion dollars. Today, there's a dozen companies worth a trillion dollars in the public market. Plus, you have the labs, you have SpaceX, and companies in the private market.

So the sizes of the outcomes are enormously bigger, and I absolutely think that firms can make substantial returns in the late-stage business given those outcomes. Now, I will say it's really hard if that's the only thing you do. If all you're doing is buying late-stage momentum companies, I do think that's hard. There are people that do it well, but it's hard.

I think being a multi-stage, multi-strategy fund where you also have a really focused early-stage effort and a growth effort, I think you can absolutely continue to support companies at every phase of growth and make a lot of money.

Okay, but can you do vertical SaaS growing triple-triple-double-double?

Yeah, I think, again, with the right valuation metrics, I wouldn't exclude—I wouldn't write off—a company purely because that's the growth profile. Now, I see the point. You have to focus on large outcomes, and I agree with you there.

8. Losing ServiceTitan: Investing Lesson Learned…

I'm telling the team, we need to do 2 things. One, we need to replace seats. We're replacing labor. And 2, I need to see $1 billion in revenue. Before, it was like $100 million, and we can sell it for $1 billion or IPO. A $1 billion exit doesn't do [__] for us now.

Yeah, I agree with you. As much as I enjoy sparring with you, I agree with you on this point. If you can't articulate the big outcome, and if the founder can't articulate the big outcome, that is probably a sign that you don't want to be involved with the company.

But I think what you're describing is basically the mistake that we made on a company like ServiceTitan. We had fallen in love with [unclear]. We were chasing this round; it was going to happen in the $250 million or $300 million range. We had these rigid rules about it: you definitely can't pay more than 6 to 8 times forward for vertical SaaS, and you definitely can't pay more than 10 times forward for vertical SaaS. We lost it because we got cute on price, and then that went on to be a $9 billion company.

If you really understood the depth of the market and if you really understood what they were disrupting in that era, you would have done it, even though it was a vertical SaaS where you might have otherwise historically thought it was constrained.

When we said about Cursor, I liked your description of the platform company for engineers, and I see that grand play. But then it kind of goes against something that we noted down before you said: “‘Who will win?’ is a narrow-minded framing of the market.” Are they not paradoxical?

I think Cursor will win. I think there's huge value to being the winner in these markets. But the reason I think the framing is overly simplified is that people forget we don't operate in monopoly markets in this country. The forces of capitalism don't permit it, and if they did, then the federal government wouldn't permit it.

I think the best software company in the world is AWS. AWS has around 35% market share. Everyone aspires to win, and you get into business with these founders because you believe that they can win, but I also think the way that a number of these verticals are going to play out, in a number of the AI categories, there's going to be a couple of really big companies in several of them.

Do you not think we do legitimately operate in monopoly markets? Let's look at NVIDIA. Let's look at Apple for consumer hardware. Some, like Salesforce for CRM—and Salesforce is a $250 billion business.

Yeah. But I think it's different when you get into the megacap companies. There are monopoly conversations, and that is what the federal government is there for—some would argue. I would not argue, but that's what the federal government tends to do these days.

I think in the private markets, at the scale of companies that we're talking about, I just don't think so. I'll give you one framing for the winning conversation. We've talked about—and you talk on the show a lot about—Deel. People say Deel has won the market. Alex is phenomenal. Deel has won. We're not investors in the company, but I think it was published that they passed $1 billion of ARR. It's incredible.

It's like, welcome to the big leagues. ADP has $20 billion of ARR. You're 120th the size of ADP. And, by the way, in this market, you've got Paychex, which is a $60 billion company, and Paycom and Paylocity.

So, I think the venture framing of “this company won” is not always—I think it can be a little bit oversimplified. Do you reflect on those two? You're not in Deel or Rippling?

Separate conversations. We're not in Deel because we're in Remote, and I'm thrilled that we're in Remote. I think Job and Marcelo are very special. I think their product vision is very different and unique.

I'm in both. Welcome to 20VC.

Exactly. I'm sure we'll talk about conflicts somewhere in the conversation.

9. Missing Rippling: What We Learned

Yeah, it's fine. We don't embrace conflicts.

The Deel conversation is separate because I'm very thrilled that we're investors in Remote. The Rippling one—yeah, I think about this one a lot. This one stings.

I think a lot about the physics of these businesses and the product mechanics behind a lot of these companies. What I mean by that is, I think a lot of investors tend to look at, “What's the product? What's the growth rate?” Et cetera, et cetera. No one really has an appropriate appreciation for what I think of as the marginal ease of ARR accumulation.

What are the downstream levers that you're putting into place that you can pull on in the future that will allow you to grow at these crazy growth rates in year 4, 5, 6, and 7? How do you build this growth mechanism that is better than, “I put in a marketing dollar and I get out $1.20 of revenue”?

I think nobody in the world does that better than Parker Conrad. The first time he outlined the vision, I was like, “This is really compelling.” I think that's what he does. He has this innate sense for pockets of margin that other people wouldn't go build companies around, like laptop provisioning and physical IT leasing. That would be a tough standalone business, but as a revenue line item for a company like Rippling, I think it's really interesting.

I just think that Parker is a generational founder. We don't get it right all the time, but he's certainly someone I wish we were in business with.

Why are you not? Was it because of the Remote situation or price?

I think it was a couple of things. Parker previously had a reputation. I'm not going to opine on whether it was deserved or not, but he had a reputation that I like to think he's now totally overcome. That just came up in the conversation, and in a market where Moon was moving very quickly and other people were moving quickly, it probably made us a step slow.

I think this is also one where we stuck to our knitting on the investment framework and the ownership thresholds. The opportunity to get involved was going to be at a high valuation, and maybe there was—I don't remember the specifics—but there was a mechanism where you could invest more over time.

It would have required us to break a lot of rules. I don't regret not breaking the rules in general, but this would have been a time when it could have been worthwhile. I'm slightly confused right now as to whether we should break the rules on Series A's. The prices have gone from 20 on 100 to 20 on 200 to 40 on 400.

I'm forced every day to question whether we should break the rules on ownership for these incredibly fast-growing, hot AI companies, and we go back and forth on it. We're friends sitting in a coffee shop. What would you say to me if I was debating that?

Yeah, I'm chuckling because there's this funny quote that comes to mind. I've been very lucky at Accel to learn from a lot of really great people. One of them was Jim Brier. Jim used to say this thing, which I think he was paraphrasing from somebody else, but it was basically that investing is an art and a science.

The science is understanding how to properly value a company, and the art is understanding when to break the rules. I just think in this market you've got to do that constantly. Generally speaking, sticking to your rules is a good place to be.

Now, I do think the vocabulary around what a Series A is in this market is just very different. I actually think there are multiple subcategories of investing that go on in Series A land. You just have to decide what you want to participate in and what you don't.

It's okay to say no. You don't have to be in every single round. Breaking the rules is something you should do very, very rarely. You said that brilliantly wanky phrase, “the marginal ease of ARR accumulation.” I'm going to give you 5 tequilas and then ask you to say that again. Sounds wonderful. Where did you think there was marginal ease of ARR accumulation where there maybe wasn't, and what did you not see?

I think as the market has gotten more competitive, the pressure to be right, to pick correctly, has never been greater. It causes you to extrapolate. You have to extrapolate from early data points.

There have been investments where a company went from $1 million of ARR and then, in the period before they fundraised, they had a $4 million quarter. It's like, they've got it—the product-market fit snapped; this is it; it's time to forward-invest. You can extrapolate these trends, and then it turns out sometimes they just had an anomaly quarter. I have fallen into that trap before.

Do you have that more and more now, when we see companies being maimed by others so significantly?

Yeah, definitely. I agree with this. I think this is why the benchmarks that used to give us all comfort are largely obsolete now. You have to be really clued into the usage intensity of your product and really understand how people are using it, because growth can obscure and blind you to a lot of underlying ills in the business.

I do think that being clued into how people are engaging with the product, whether you're an enterprise company or a consumer app, is more important than ever.

Do you find it hard, the binary nature of this world? Honestly, we come into work sometimes and we're like, “What the fuck are we doing?” I'm being serious. I was talking to my dear friend Jason Lemkin the other day, and he's like, “Fuck this. I've had enough of this. I just want to do an Anthropic SPV and go home. I don't want to pick the winner in a SaaS company that—oh my God.” We feel so unimportant.

Yeah.

I have to be honest with you.

No, I fucking love it, to be really honest. I'm so lucky to be in this industry, and the competitive thrill of chasing down these founders and chasing these deals—it's awesome. How lucky are we to get to do this?

No, I understand where you're coming from, but I love it.

When we look at the big exits this year, you've got Databricks, you've got Anthropic, you've got OpenAI, and you've got SpaceX. As a partnership, do you guys lament that you're not in them?

Of course.

What's the internal management?

We are in some of those companies, but, yeah, absolutely. Nobody is harder on us than we are. Nobody has a perfect track record, and we want to understand where we went wrong. Every breakout company that we're not a part of, we want to know where we went wrong.

We also do that in the interest of getting it right going forward. When we look to the future, there are a lot of things that we're really excited about—a lot of companies where we are sort of the investor of record. We intersected with them very early, continued to buy up all the way through the growth stages, and we're excited about those.

But absolutely, we hold ourselves accountable when we miss companies.

I spoke to one of your LPs before, and they said, “Help me understand why we're not in any of the foundation model companies. Why are we not in Anthropic and OpenAI as an Accel LP?” Was that just a miss, or was that a belief that they wouldn't be good companies?

A lot of firms missed the model companies early, and we're guilty of it. Nobody has looked in the mirror harder than we did and course-corrected.

Can I ask, when you did, is it like a partnership meeting—“Have we fucked up?”—or is it an unspoken rule, like the British people when it rains and we just pretend it doesn't rain and we walk anyway?

No, it's the most important conversation there is. It's a global offsite where every partner at Accel sits in a room together and we say, “How did we not get this right, and how do we fix it going forward?”

What are the 50 best private companies in the world right now? For how many of those companies are we not just a passive shareholder, but the investor of record? What is our score? And then, what do we think is the next set of 50 companies, and how many of those are we going to win?

If we're not getting better, no one will beat us up more than we will ourselves.

So that's what the conversation is. It's the most important thing for the entire partnership globally.

Are you playing a coverage game? When we had Anish Acharya from Andreessen Horowitz on the show, he was 100%: “We are playing a coverage game. We get split up by themes, stages and categories, and you are expected to see 100% of yours. If you miss, it is not acceptable.” For me, for example, we play a different game. I don't have to see 100%, but I need to hit one of the big ones.

Yeah, we're organized a little bit differently, but of course we hold ourselves to the same standards of coverage. The aspiration is 100% coverage and a 100% win rate, right? No one in the industry does 100% of both, but if we're failing on one dimension or the other, we're going to talk about it and understand where we need to be better.

10. What is Accel's Win Rate

What was your win rate today?

How would you measure it?

When you go for a deal with a term sheet put down?

Mine individually or as a firm?

As a firm, but with a term sheet put down.

I'm not evading the question. I don't know the answer, so I'll speculate. I think a healthy win rate would be around 80%.

And the reason it's not 100% is because—

No one's going to have 100%.

Some people have come on this show and said that they have a 100% win rate.

Andreessen Horowitz?

I wasn't trying to call him out specifically, but I've just heard it said before.

Well, fair. If I said I never lost a deal, I'd be happy if someone said it to me. I don't mean to be combative about it. My polite and professional response would be: I think if you're not putting yourself out there and losing sometimes, you're not chasing competitive enough things.

I really like sticking my nose into a competitive fight like that. I have no right to win, and I really like doing that. But I also find a lot of joy in finding these founders who are just doing things a little bit differently. Maybe they've bootstrapped the company. Maybe they're located in some geography that's far away from Silicon Valley. Having these nonconsensus ideas that other people might think are silly, or that they might not really have their heads wrapped around, I think that's great too.

But part of it is that kind of growth equity, like technology venture, isn't inherently the most attractive in this AI world? I think that business has gotten harder.

Yeah, and frankly, that was the core of our early growth strategy.

Unbelievable. The bootstrap, 1Password-quality. Totally from nowhere. Amazing. I love this.

Absolutely. That is still out there, and we still do a lot of it. We aspire to be the best in the world at it.

Is it still out there in the world?

It is. This is the funniest thing. Every time we have an off-site strategy conversation, we keep saying, “There's no more bootstrap. There's no more bootstrap,” and then you find a Laravel. They're still there. They're really hard to find, and I think we're the best in the world at finding them.

Interesting. It makes sense from an opportunity cost of capital to put your money there versus just putting another $100 million into Cursor.

Well, again, nuance in portfolio theory: a part of our business will always be doing that. It's a very distinct fund. At the moment, we have a $1.4 billion growth fund, and we have a larger, later-stage pool of capital.

Yeah, no, 100%. It's $1.4 billion, and then you've got later stage, which is like $3 billion or $4 billion. I'm sure it's $4 billion. Is the growth fund subscale, then? David George has got $6 billion or $7 billion to play with, and Josh has got $9 billion to play with. Is that subscale, or should we think of later stage as the growth fund in the same way?

In many ways, the market today is like what the venture market was in 2000, but inverted. The idea was: I'll do my Series A, I'll get 30% ownership, I'll take a bunch of dilution, and when the company goes public, I'll own 20% of it. That was roughly the math.

11. How VCs Approach Ownership Has Changed

Today, you have to back into 20% the other way. You do what the market will allow in the earliest possible investment. You sponsor a tender, you do a growth round, you do an IPO round, and you can ladder your way up to 20% ownership. You have to be a multistage fund to do that.

Or you hope and pray that the expansion, the multiple or the size of the exit is so much bigger than it was. It's not $1 billion to $5 billion; it's $50 billion to $100 billion. Having 5% is actually just as meaningful as having 20% of the prior company.

Sure. I think you won't be surprised to find that I don't think hoping and praying is a great strategy.

Dude, we are all fucking hoping and praying right now. I'm sorry. Figma is an $11 billion company. The unbelievably, unwaveringly brilliant founder, Dylan Field, and this being the swan song of software, is $11 billion, which is incredible—an incredible, incredible outcome—but it doesn't return your growth fund.

We're arguing two separate points, and I agree with you on the Figma case study and all these fundamentally incredible businesses out there that have gotten beat up. That's a separate point that we should circle back to.

The other point is: do you have to swing for the fences? I go back to Arthur Patterson, co-founder of Accel, who always says this thing: “Focus on hitting singles and doubles, and let the home runs take care of themselves.” What he means by that is, if you're just constantly stepping to the plate and trying to say, “I can see at the Series A that this is going to be a $100 billion exit,” you will just overswing and you will fail.

No, but isn't that what I'm deliberately—I'm not actually. I'm fundamentally disagreeing with that. That is not what venture is about, especially at the Series A. You want to have a diversified portfolio enough that you have 1 or 2 of them hit, but you want 30 swings. Swing the fuck out of this, and it could be $100 billion. We're not here to do the singles and doubles.

I think there are different ways to practice the craft. I do think the market has evolved a little bit, and a single and double today might look different than it did in 2000.

But I think what he means is: know what you're good at, focus on founder relationships, stick to whatever your particular strategy is, and just try to do that really well. Don't just go slinging into momentum-chasing opportunities where you're not going to be any better than the next investor.

I think that advice is fair.

Do you not think we're all momentum-chasing? If we look at the AI entry for you guys and then the defense entry with Anduril, we're all slightly momentum-chasing.

Again, I think I would go back to nuance, subtlety and portfolio theory. There are absolutely companies where it is justified to chase momentum. We don't like to use that vocabulary, but we see a company like Anthropic and how valuable they are as a technology partner to every other company in our portfolio.

The momentum is very obvious, but the business logic and the business fundamentals are also very obvious. Does it make sense to have a relationship with Anthropic? Absolutely. So, guilty of that.

You did the $180 billion round?

We've invested in a few rounds of Anthropic.

Can I ask, what was the first round you did?

The 180 round.

When you were doing that as a team, how did you think about outcome planning there and sizing that?

I think that company and a small handful of companies in the private market today are operating on a different plane. I think it is not bombastic to say that some of those businesses could be trillion-dollar companies, and I think people who are underwriting these rounds believe that.

That is a different category. The danger in this business is ascribing the characteristics of an Anthropic, an Anduril, an OpenAI or a Stripe to things that don't really fit the paradigm at the Series A.

I get that, and I agree. But when you're doing an Anthropic round at $180 billion, are you saying we fundamentally think this can be a $2 trillion company and a 10x?

Implicitly, yeah. We're not—

What do you need to see to write the check? What, 3x to 5x is enough on growth?

Yeah. Again, there's never a partnership conversation where we say, “Hey, we've built a model and squinted our way to a 3x outcome.” That's just not exciting.

But that's not a deal you should do.

No, no. The reality is that I do think a lot of these funds revert to the mean. If you can generate 3x net funds, that's a pretty good business to be in. But if all you do is aim for 3x investments, of course that's not really the math that gets you there.

If we can have a conversation about how this company is special, its reach is unprecedented, its founders are very, very different, and the comps for this business would be platform companies like Google, Microsoft and Amazon, then of course you want to participate in those companies in the private markets.

You said you'd be spicy. Do you feel better or worse to be an Anthropic shareholder after Anthropic versus the Pentagon? I could argue both sides.

Yeah, you were definitely going to give me some spicy ones. I think—

I can feel your compliance team just shitting themselves.

They're just crying.

Look, how can I answer this question? How can you not admire the founders for sticking to their knitting, sticking to their conviction, and sticking to their principles? Now, I have no idea how this is going to shake out, right? But how can you really blame a founder for saying, “I’m sticking to the mission”? I get it, and I respect it.

Did you not write Dario’s memo for Claude?

I’m definitely not intelligent enough to ghostwrite anything for Dario. I think this is an opportunity for a lot of these companies. They signal virtue, and they believe in a world where AI is going to be a force for good. Then there are commercial opportunities where that gets put to the test.

Totally get that. And we’re seeing it bluntly play out for him in terms of loyalty, in terms of talent, and totally in consumer adoption. Mike [surname unclear] put it: they’re doing a million a day in net-new consumer signups.

Yeah. They passed GPT in the App Store.

Isn’t it ironic, though? This is what was needed for them to surpass.

No, I don’t actually believe that they were doing it for that reason. I don’t think they did it as a calculated business move. I think this comes down to—

No, I think it was an accidental bit of luck. It worked out well. I think it comes down to ethics and principles. Call me old-fashioned, but if you behave the right way, you will be rewarded.

Yeah.

Yeah, I totally agree with you. Are you in OpenAI as well?

12. What Happens to Companies Like Miro and Snyk with High Prices to Live Up To?

There are a lot of businesses today that we’re in historically where I’m just like, I don’t know what’s going to happen. I love Snyk, I love Miro, I love 1Password, but they were done at such high prices, and the new reality is very real. What happens? How do you opine and think about businesses like that when you sit in the partnership meeting?

The market has gotten so humbling. I mean, the greatest companies of 3, 4, 5 years ago, many have gotten totally beat up in the public markets. I believe many are oversold, but that’s a separate conversation.

I think this is where it comes back to this being a human business. Who is the founder that you’ve gone into business with? What is that founder going to do when their back is against the wall? If you look at Snyk, I’m in Guy’s new company, but he ain’t there.

What do you do? It’s $300 million ARR, growing 15%. Its last price was $7 billion.

I think this is, in some regards, where, as the founder of the company, we lose sight of this. That’s not a great setup for people who might have invested at $17 billion, but it’s a great business with a great product and a great customer base. There will be an outcome for that company. It is humbling relative to the valuations of the 2021 era.

13. Why it is a Great Time to Be Thoma Bravo and Vista

But again, who is the team that you’re in business with, and how are they behaving? How are you behaving, more importantly, as an investor, when the team’s back is collectively against the wall? I think that is the best reflection of what happens. Do these businesses go public? Do they get taken out by M&A? What do you think is the route for them?

I think it’s probably a good time to be in the LBO business. I think it’s probably a good time to be in the Thoma Bravo, Vista, Blackstone, KKR business. There will be homes for a lot of those companies that get themselves to a sustainable place, and they will find homes. These homes for a lot of companies will be different from what the aspiration was when the founder started the company. That’s just a reality of this market.

I totally agree. Are you with me in the camp of, “When the founder goes, my conviction goes”? Now, when Andrey is at Miro, I’m like, Andrey is still batting. If Andrey is still batting, I’m still there.

14. Why Founder-Led Companies Are Always Better

There is unmistakably something special about a founder company.

Mike being at Atlassian—when I interview him, I’m like, I still feel it, and his passion is still there.

Never bet against Mike Cannon-Brookes. Absolutely.

When the CEO’s there, I’m like—

It’s not that it can’t work. There are incredible professional CEOs. If I could have Frank Slootman come be the CEO of a number of companies I work with, I bet the founders would say, “Yeah, that’s a great trade.” I mean, there are incredible professional CEOs.

What have you changed your mind on most in the last 12 months as an investor?

I believed this thing a year ago that, in hindsight, I feel very stupid for having said. I believed that all of the generational investments in AI had been made. I looked at my partner Dan Levine incubating Scale AI, building a relationship with Alexandr Wang in 2016, and making that investment. The early investments in the labs—I sort of thought, listen, the bets were made 8 years ago, and it’s too late, and now we’re all sort of fighting for what’s left over.

That was a really stupid thing to say, and I no longer believe it. That’s probably the thing that I’ve fundamentally changed my mind on, both because those companies will be bigger than the outcomes that I probably envisioned a year ago, and there is still time to be a part of some of them, and because the innovation flywheel is just getting started. We are barely scratching the surface.

When you had the Scale exit—for context, $14.9 billion, amazing exit—Dan was unwaveringly the first investor there from the dorm-room-style moment. Epic. When you had that, and the company got an offer for $14.9 billion, is there high-fives and “This is awesome” around the table?

No. I mean, there is an appropriate congratulations and acknowledgment to Dan. There is a huge, loud, full-throated thank you to Alexandr, and then everybody gets the fuck back to work. It’s a humbling industry, and you are only as good as the next thing that you do.

How do you analyze that market today? There’s one that I really struggle to get my head around in a way that’s not cynical, not paranoid. I just don’t know. There are so many different providers that are all at very meaningful revenue scales.

Yeah, the Scale-Mercor market.

Yeah, as we said, with your Mercors, with your Turings—I mean, there are 10 or 12 of them, micro ones.

I probably struggle with services businesses in general getting valued on extreme, extreme ARR multiples, but there’s no disputing the strategic value of it.

15. Does Miles Feel Happier or Sadder to be an Anthropic Investor Post Pentagon Debacle

Totally get that. You said something about the value of different revenue multiples, and we’ve spoken a lot about you and Mike at Atlassian before. There are clearly things that Mike is not able to do because he’s public that private-company founders like the Collisons are able to do. How do you think about the benefits of public versus private today, given the liquidity so inherent within secondary markets, like we’re seeing even as early as with your Linears, where you’re doing tenders for them? Clay has tenders, and then Stripe, on bigger scales, has obviously much, much more liquid markets.

Why would anyone go public?

Well, the reverse is true, too. There are things that Mike can do as a public CEO and that public companies can do that private companies cannot. But I think you’re asking the right question.

I mean, I think there’s a reason a lot of these founders are staying private longer. What are the things that you typically needed to access the public markets in order to do? Liquidity for employees—you can certainly do that now as a private company. M&A currency and just increasing your valuation benchmarks, or your valuation mark, you can totally do that as a private company. So I think that is all true.

With that said, I think that applies to the 10 best private companies in the world. Databricks can do those things. Stripe can do those things. There are a lot of companies that just do need to get public.

Totally get that. The trouble is those companies need to get public, but they’re in the $2 billion to $10 billion range. Does anyone care about the $2 billion to $10 billion range anymore?

I think you’ve seen this phenomenon where I would actually peg the range a little bit lower. These companies that have gotten public in the $2 billion to $5 billion range, and then they never really break out—I think that has been a difficult threshold for a lot of these companies to break through.

I do think this is why you see a lot of good companies waiting. People say, “Oh, it’s because the investors will be underwater.” I don’t think that’s actually the reason. I think it’s because, generally speaking, you want to go public and you want to be able to have a fairly clear line of sight to hitting the $5 billion threshold and trading beyond that, because it’s murky below that.

Is the SaaS apocalypse an overreaction, or is it actually the fact that we were just bluntly valuing them far too highly on relatively mediocre 18% to 20% growth rates, and this is a realization of that?

Fundamentally, people are valuing the future cash flows and the future terminal value of these companies differently, and I don’t think that’s wrong. But I do think this has been an over-rotation.

What is the most oversold stock?

16. When is the Right Time to Take Chips Off The Table?

We’re not a part of Figma, but we have a lot of respect for that company. I know that Dylan is a generational founder, and it’s a very important company with an incredible financial profile. It just feels, in a lot of ways, for a lot of reasons, oversold.

We mentioned the liquidity inherent within companies now as it goes later and later stage.

How do you think about when the right time is to take chips off the table?

I think you have to operate from the first principle of what is best for the company. Now, if the company is saying, “Hey, we’re going to do a big tender and a secondary round, and it’s okay if investors want to sell,” then I think in those circumstances, it’s generally wise to diversify.

But it’s got to be the right thing for the company and for the founders, first and foremost.

Can I ask? I’m sorry to be so annoying. I used to be so nice. You should have done the show 5 years ago, when I was a sweet little boy. You know, that’s just not true.

When you look at, say, WeWork, Benchmark was fantastically smart to get out of it. When you look at Lightspeed and Jeremy Liew selling in Snap, they were very wise to get out of it. We’re seeing prices so far ahead of company traction now. It’s not in the investor’s interest to sell, but Jesus, we’re paying 4 years ahead of time. It’s in our interest.

Yeah, but it’s so situational. As a principle, do I think it’s good to get liquidity back when it’s available? I do. But it’s so situational.

You use the WeWork example. We were not a part of WeWork, but had I been a shareholder in that company when it was worth $50 billion, I probably would have been seeking liquidity. I don’t know a whole lot about the commercial real estate market and the office space market, but that just feels rational.

But why does it not seem rational to seek liquidity at Miro at $17 billion?

I think that was—you know, we didn’t take liquidity out of Miro at $17 billion. But again, what was Andrey doing? What was the founder doing? What was the course that he wanted for the company? That’s the only thing that matters.

The example I would point to is CrowdStrike. Sameer Gandhi and John Locke intersected with CrowdStrike when there was about $1 million of software revenue and a $9 million consulting business. That was the company. I think they invested in 2011 at $160 million post-money.

There have been nonstop opportunities to diversify and sell CrowdStrike stock. It’s a public company; you could do that today. But Sameer and John led the next round. They led the next round. They bought in the IPO, and it’s a $100 billion company today. We’re sure glad we didn’t take chips off the table.

The question there is, you mentioned, obviously, buying into the IPO and the decision to hold thereafter. Obviously, Sequoia has the evergreen vehicle, which I think there’s been a lot of talk about. Do you think venture firms should have the responsibility of managing the book into the public markets, or do you think it should be distributed to LPs and dispersed from there?

17. Should VC Firms Have Evergreen Funds and Be Responsible for Public Positions

Yeah, I think fundamentally we’re in the business of identifying outlier founders. If you’re a multistage fund, that gives you the flexibility to stick with some of those founders for the long run. You should definitely do it.

In the George Kurtz case, it was absolutely worth doing. In the Mike Cannon-Brookes and Scott Farquhar case, it was absolutely worth doing. But not every company has the mechanics to compound for a long time. You can’t just do it as a blanket rule.

I don’t like public markets right now, and I think it’s just a bad place to be because you’re seeing the casinoization of public markets, where a Catrini report can wipe billions of dollars off Anthropic, and Anthropic doing a security release impacts CrowdStrike. Are you high?

I think the public markets are no longer rational.

Yeah. To me, it’s not good or bad; it’s just a different asset class, and you should stick to what you’re good at. I don’t think we would be excellent stock pickers, but I think we’re pretty good at what we do in terms of early-stage technology investing.

I just think it’s an asset class that I’m never going to be best in the world at understanding—public stocks—and I think that’s okay.

Who’s the best sourcer inside Accel? Sourcing is finding great companies.

Christine Esserman, really, really good eye for companies. She is relentless in getting in front of founders. She’s great.

Who’s the best picker? Like, picking noses.

Andrew Braccia, by far. Andrew is wise. Andrew has seen success at incredible scale. He’s our best picker.

When it comes to winning—a core part that we don’t talk about enough, I don’t think—who’s the one where you’re like, “Okay, we need to win the deal. We need to bring them in”?

I think Sameer Gandhi is incredibly compelling and hits it off with founders in a very special way.

Do you think the best founders need your help? I was going through the pillars of venture there in terms of sourcing, selecting, securing, and servicing, and I was like, do you think the best founders actually need your help?

I think “need our help” is an overstatement. I think of the role of a good investor as being that there are basically these bumper decisions that come up a couple of times a year.

If you’re a founder, your life is a bunch of little decisions and then a couple of really big decisions. The little decisions are design decisions about the product, pricing, whether we should dial up CAC, and whether we should make this hire. You don’t need an investor micromanaging you through all the little decisions.

I do think every year there are probably a couple of big decisions where having a good sounding board can be really useful. Should we do this partnership? Should we make this acquisition? Do we need to pivot? I think having a good investor, or just a good partner for the business, can be really useful. But it’s all about striking the right balance.

You like being a board member?

Yeah, I love it.

Who’s the best board member you sit on a board with?

The best board member I’ve ever seen in action is my friend Ravi at Sequoia. He’s a very good board member. He was at Sequoia, and now he’s doing his own company.

Why?

Ravi has done a lot of different things. He’s been an operator, a buyout guy, and a growth equity investor. But I think it’s more about his demeanor and the humility with which he delivers feedback.

He has this way of saying, “Let me politely make an observation,” and you can choose to accept it or reject it. There’s just wisdom and humility in how he communicates, which I appreciate.

If you’re a founder listening to this, what advice would you give them on how to observe potentially unhelpful behavior from a board?

I think there’s generally an inverse correlation between how vocal somebody is and how helpful they actually are. The person who just has to get the first and last word in, shows up at the board meeting, and has to teach you something that you didn’t already know—I don’t actually think that is the model for wisdom as a board member.

There’s a brilliant one: “The coin box that rattles loudest has the least in it.” Another one that a fan told me the other day, which I thought was helpful, is, “My lesson from boards is VCs are great at identifying when to hire someone, and they’re awful at identifying who to hire.”

Yeah, your buddy is probably the CRO of Atlassian. That’s not great for my $10 million business. That’s not a good thing.

That’s exactly right, dude. I’d love to do a quick-fire with you.

Okay, I love it.

Seed firm, Series A firm, and growth firm that you have to invest in—and it can’t obviously be Accel. You put all your money in Accel.

Yeah, yeah, of course. Seed fund: I really like the guys at Liquid 2, Nate Redmond and Matt Mulvey. Those guys are prolific. They have an incredible network, great taste in companies, and they are kind, enjoyable people to work with. When they send me something, I take it very seriously.

Good. I like them too. Series A.

The vocabulary around what a Series A is these days has evolved, so I’m not sure how you would bucket these guys. I really like the team at Meritech.

I think they have very good taste in companies because they do some Series B and later-stage things also, but—

Great-tasting companies, and they are gritty and they hustle. I would say it’s not coincidental that Max and Alex were also trained at Summit Partners. I really respect that part of their pedigree. I really like those guys.

Great. How can you not acknowledge how successful Josh Kushner has been at Thrive? I really admire the way that they have scaled a business that not only can initiate investments and invest across funds, but really reflect their conviction at the late stages.

We compete with them fiercely. We also work with them. I’ve gotten to work with Miles Grimshaw through the Cursor board, which has been a great experience.

Miles is amazing.

He’s great. He is my much more articulate, intelligent VC alter ego—the other Miles. But we have a lot of competitive respect for those guys.

He’s also a marathon runner in something like 2 hours and 10 minutes. This guy is like a—

Yeah, no, I mean, I could finish a marathon in 2 hours and 10 minutes on a motorcycle, but it’s different.

Honestly, I see Miles and I’m like, “I need to be better as a human being.”

Different strength.

I totally agree. I could win in an arm-wrestling match.

If you could add 1 person to your team, who would you add? This can be completely hypothetical. It can be Pat Grady. It can be Josh Kushner. It can be Elad Gil. You’re like, “They’re going to move the needle most in our ability to win.”

I’d probably try my very hardest to talk Mike Cannon-Brookes into retiring from operating and becoming an investor. He would never do it. But if I could go to war side by side every day with somebody, it would be Mike, and I’d never, never bet against Mike.

What about other VCs? I agree with you. I think Mike would be amazing. I’d probably take Neil Mehta. I don’t know anyone who has the investor breadth that Neil has, from doing Windsurf’s first round and sticking with them through many pivots to doing Carvana in the public markets and having that breadth of aperture.

I might—this is actually a really good question. Or Micky Malka. Micky Malka’s ability to see trends so early is just exceptional.

Do you ask this one a lot? This is a really good one.

Thank you. I saved it for you.

Oh, thank you for that. This is great. I’m jet-lagged, discombobulated, and I didn’t prepare for this.

I think somebody who I not only have a lot of professional respect for but somebody who I personally just like a lot is actually Matt Bornstein at Andreessen. Do you know Matt?

He works with Martin. He’s deeply technical and very thoughtful, and was instrumental in their finding the Cursor investment. Matt doesn’t like admitting that he also has an MBA from Harvard, but we sat next to each other for a semester at school, and I really like him. I enjoy spending time around him, and I think he’s really, really smart.

What advice would you give to someone starting their career in venture today?

I would give the same advice that Arthur Patterson gave me and says all the time to us as a firm, which is—it’s just about professionalism. Arthur says that any firm can be professional over short periods of time, but his aspiration in starting Accel with Jim was to maintain a standard of professionalism over long, extended periods of time.

That means respecting the process, respecting the partner meeting, respecting the portfolio review, respecting the rituals of the firm, and going about the job in a professional way. I would give that same advice.

Tell me, what deal did you not do that you wish you had done in the last 12 months?

I think 11 Labs is a clear company that we wish we had been a part of. As you know, we haven’t spent enough time with the founder, which is our loss. I think we really regret that one.

Try and do the 11 billion round?

We didn’t. There are—as I said, nobody has a perfect success rate. This is one that at the next offsite we will beat ourselves up over. But, as I understand it, a very special founder, very clearly an important part of the modern AI stack, so that one stinks.

What’s worse, losing or not seeing it? Because losing—everyone says not seeing it, not seeing it—but losing really sucks.

They’re equally bad, but losing stings more. Having had the opportunity and failed stings.

Can I ask which loss hurt us the most?

I don’t know that I would characterize it as us losing. But the company where I really, really loved the founder and we didn’t get there was Shiv [company unclear]. We actually hosted this AI dinner a couple of weeks ago, and I was like, “I’m going to manipulate the seating chart and get to sit next to Shiv Rao,” because I just think he’s generational and very good. I regret that we didn’t get to work with him.

Mhm. What win feels the best? That moment of jubilation.

There was one where I was going through some personal things and happened to be able to compete for and ultimately win the opportunity to work with Linear. That one, on a personal level, was maybe the best week. It had been the worst couple of months that I had experienced in a long time.

There was this very surreal week where it felt like Kari might decide to raise capital. I basically decided that I was going to go park myself in Southern California. He lives in Del Mar, outside of San Diego, so I would stay there until he basically decided whether or not he was going to raise money.

When you say “parked yourself there”—I don’t mean this badly—but you literally parked outside of his apartment?

No, I got a hotel room somewhere. I would get up and go for runs and see if he wanted to hang out, trying not to bother him. But in the event that he said, “Yeah, I’d love to get lunch,” I just wanted to be nearby. It’s a little bit creepy as I say it out loud, but at the time it seemed right.

I had a lot of stuff going on at home. It was my best friend Craig’s birthday. Craig, by the way, is the only reason I got into this industry to begin with. He got me my first job and then my second job. I was flying back and forth: I would go home, see my kids, go to San Diego, sit there, try to hang out with Karri, go home, attend Craig’s birthday—which he wouldn’t have cared about, but he’s my best friend and I needed to do it—and go back down to San Diego.

There was just a lot coming to a boil in my personal life. When Kari called and said that he wanted to work together, it was pretty euphoric. I will always remember that week. It’s been a special company to work with, but on a personal level, that one felt pretty good.

Actually, to be fair, the 2 companies I’ve never had more requests for introductions to were Linear and Lovable. Every growth investor wanted to meet Kari before that round. It was annoying, to be honest. And then Lovable was really annoying. That was really annoying before the round that did, because we were in the round before.

Who else are you getting bothered about right now? I should go see them while I’m here.

Oh gosh. Yeah, there are 2 or 3. It’s so funny. You see the investor wins where it’s just like—

And you don’t bother sending it to the founders, because it’s like, “I’ll send you a list of names of people who want to—”

Yeah.

And that’s it.

Yeah.

And it’s even worse for me because I often have them on the show, so people assume that you’re great friends.

No, I just met whoever it is when we did the show.

I just assumed you were best friends with everyone who comes on the show.

Yeah, best friends.

Final one for you, dude. What are you most excited about when you look forward? I think it’s really important to be optimistic. Optimists make money. Pessimists are right. What are you most excited about?

Honestly, the thing that I’m the most excited about is watching the younger team at Accel flourish. I’m not smart enough to predict where the world is going to be a decade from now, but I can tell you that Christine Esman, Ben Quazzo, Josh, Rohan, and a bunch of folks on the team—Gonzo and everyone who’s going to be mad that I’m leaving them out—we have such a talented team.

They are the unsung heroes of the firm who don’t necessarily get the attention that they deserve. I’m so excited to see what they’re doing a decade from now, and I’m proud to know them.

Dude, it’s such a pleasure to have you on. It’s so nice to see you in person. Thank you for joining me, Miles.

This was a blast. Thanks, Harry.