Sequoia Capital’s Roelof Botha & Axios’ Dan Primack

Thank you, Don. Thank you for having me.

So, there are a few things we know how to do well at Sequoia. One of them is to make great investments and the other one is transitioning leadership, which we've done many times before. I joined Sequoia 23 years ago and I wanted to learn from the best from Mike Moritz, Doug Leone, and Jim Goetz. I ended up helping us lead investments in YouTube, Instagram, Square, MongoDB, Unity, and a couple of other companies. And then in 2017, so just more than eight years ago, the partnership elected me to become the steward or the managing partner of the US and Europe business. And at that point, my focus shifted to overall team and investment performance. In the time that I was steward, we distributed over $50 billion to our investors and we saw the IPOs of Airbnb, DoorDash, Snowflake, Zoom, Robinhood, Unity, NuBank, and the pending acquisition of Wiz, to name a few. But the thing that I'm most proud of in that time honestly is the team that we've been able to assemble. When I look at the quality of the bench that we have at Sequoia, we have Andrew and Constantine and Sonia and Stephanie and Bogomil and Sha and Luciana. We just have a firepower team. And when I became steward, the most important thing was to think long term, to think about how to leave Sequoia in a better place than I found it. And with that in mind, I always had a view on who the next generation of leaders would be. And Alfred and Pat are both exceptional investors. And people outside may not know this, but they've been co-leading the early and the growth teams for Sequoia for many, many years. And this is their time.

That's part of the secret we have at Sequoia. We've done this successfully over time. We want to make these changes very quickly. I think there's a concept which I love is interregnum, which is when you have a period in between administrations and that's when organizations suffer stasis. No decisions get made. And so if you go back in our history, we've typically done this very quickly and to many people it may seem as a surprise, but internally it's always been deliberate.

No.

That's the challenge of management is to balance trade-offs. Every decision involves a cost-benefit analysis.

I think in any of these situations, there is a point at which free speech becomes hate speech. And I think that is clearly a threshold around decorum and decency that is an important boundary that one needs to set with how you respect free speech.

Not sure about that.

Oh, we just invested in a company called Flow Engineering that's building a requirements management system for hardware development.

I'm not saying it's real. I'm saying we have... we have a company called Apora that provides international remittances using stablecoin as an infrastructure for non-resident Indians.

So I think the bubble is a pejorative term in what I think is one of the most tectonic changes for society in terms of how we can harness this technology. And there's an interesting book that I read recently that talks about how these sorts of periods end up concentrating enormous amounts of capital and minds on opportunities that can be transformative. And I think AI is right up there. When I look at the sort of companies we're able to see walking through the doors at Sequoia, they touch industries that to me were hard to conceive as being impacted by technology and being upended by technology in the way that you're seeing right now.

Absolutely. And I was in Silicon Valley in '98, '99. I saw what happened with the dot-com era. I saw what happened with mobile. Saw what happened with cloud computing. This surpasses all of them and more.

I think there are so many industries in America that have ossified. There's this index, it's called HHI, the Herfindahl-Hirschman index, which measures the relative concentration of industries. And research by the Fed Reserve actually shows that more than 75% of industries in America have become more concentrated over the last 20 years. In most cases they've increased concentration by 90%. So you have so many industries that are relatively ossified and ripe for disruption. And AI isn't just a new distribution mechanism or a new interface. It is fundamentally a set of capabilities that can upend industries. And we're seeing that in cybersecurity, in customer support, in the legal profession, in finance, just industry after industry where an entrepreneur walks in and they've encountered a problem and they describe what they're going to do. I just go, 'Wow, I hadn't thought of that.'

Well, the two things when I think about many of the other bubbles that we've encountered and I don't think it applies here. I think this is far more constructive than it is a bubble. They're often debt-fueled and most of these companies are spending money that's on their balance sheet. This is existing cash that's being used. So you don't have the same leverage impact and then deleveraging impact that you may have had in other situations. The other thing I'd say from all the evidence that we have is that the majority of cycles being used right now are going towards inference rather than training. And so I'd say two years ago when I heard these arguments I was much more sympathetic towards it because when it was training it wasn't obvious that it was actually solving customer problems. That's changed. And unlike what happened in the dot-com era where people were building infrastructure that lay fallow often until 2005, 2006 when Web 2.0 took off, in this case every spare cycle is being used. There's no spare capacity.

That's always the risk. And I remember when I joined Sequoia Capital, Michael Moritz had this great phrase that yes, there are elephants in the industry, the big incumbents, but they're more likely to trample on you by accident than on purpose. And I think we're always looking for companies that have a keen insight and an ability to focus in the way that the big companies cannot. I'd say the other difference by the way this time is AI has become a board-level issue for Fortune 500 companies, right? And so I think bigger companies are able to take advantage of this technology in a way that I don't think they were geared up in mobile, cloud, or in the internet. It doesn't mean they'll always get it right and it doesn't mean that there aren't enormous opportunities for us as venture investors.

What do you get though if you have a key?

Yeah.

But you can't force them if they're not motivated to do so.

It's always a challenge when you have big technology shifts. I think we certainly saw this when we were investing in web companies in the mid-2000s and then mobile came around. And the challenge for those companies was could you become mobile first or maybe even mobile only and sort of let go of what you had before. And Meta, Facebook at the time, is a great example. I remember how people were doubtful that they were able to make the mobile transition and their stock plummeted and people thought, 'Oh, you'll never be able to monetize mobile traffic.' How laughable is that in retrospect? So I think a similar challenge exists in every wave. And there are companies where the founders see the opportunity and are moving very aggressively. And our job as board members is typically to help shine a light on what's around the corner.

Not that I can think of.

Maybe.

I think SpaceX has a bigger chance of being the most valuable company. I think SpaceX by itself is responsible for 80% of all mass moved into orbit last year. It dominates that particular industry. I'd say in the case of OpenAI, an incredible business. They have several competitors as large language models, as foundation model providers. And I think Google is starting to really shine with Gemini as a consumer service. They now have over half a billion monthly active users. So it seems as though there's real competition now for the consumer aspect of what OpenAI is, and so it doesn't have runaway market dominance characteristics.

Well, OpenAI also doesn't have a board seat.

But we've... I mean I've known Elon since he recruited me to PayPal back in 1999. And so a lot of that comes down to your trust and faith in the individual. And Elon doesn't want to be encumbered with some of the bureaucracy that goes along with more traditional boards. It doesn't mean that he's not open to advice. We text frequently. He has many people that tell him things that are uncomfortable to hear because he wants to build really successful businesses.

Well, it depends if you include what's happening in robotics as part of AI.

You'd include that because of the computer vision capabilities.

And I think it clearly is benefiting from the work in foundation models. We have an investment in a company called Skild that actually is providing a foundation model for robotics. So you can take a $10,000 robot and have it be as performant as a $100,000 robot because of the intelligence that you imbue. Cobalt, Robco, several other companies. So I think there's this era of us being able to mechanize a lot of labor. That's interesting. And then I think stablecoins are absolutely fascinating. We're an early investor in a company called Bridge that Stripe acquired that's helping with commercial international payments and totally circumventing the need to use the legacy Swift banking system, which is expensive, cumbersome, and lacks transparency.

No, industry is okay. Asset class is bad.

Well, I think partly the threshold to become a public company has changed pretty dramatically. I mean, I was thinking about MongoDB. I was at the board meeting last week actually, and when we went public in 2016, the company's market cap was about a billion dollars. Today it's worth 25 or something, and that was at the bottom end of the threshold. I think you can't take a company public today at a billion dollars. I mean, it probably costs you $50 million a year just to comply with Sarbanes-Oxley and all the internal control requirements.

I think the era of being able to own 15, 20% of a company by the time it goes public and have it be a billion-dollar company is gone. That return math doesn't work. So unless you see a path to this company being worth four billion at a minimum, you can't bank on it being an independent company and then you're dependent on M&A, which candidly hasn't changed that much under the new administration. It's still pretty darn unfavorable. Hence the acqui-hire dynamic that you alluded to earlier. So M&A is not as easy given the regulator.

I did. I was wrong.

I think the vibe in general was that it was a more pro-business administration than we'd seen under the previous administration. But M&A is still an area... maybe back to the comment I made about the Herfindahl-Hirschman index, that people recognize that there are so many concentrated industries and that's bad for America's competitiveness longer term.

It is coming. I agree with you on that. I think the thing which I worry about a little bit is for retail investors to invest in the fund structure of venture capital, to me is a little bit more risky because it's...

I think what would be more interesting is if there's an ability to trade secondary shares in more mature companies or standalone companies. So a company like SpaceX or Stripe. These are companies that are of massive scale but remain private, and retail investors don't have an ability to participate in them. Maybe that's an interesting idea. We have a photo. Can someone throw this up on the screen for me?

No, no, no. That's Drew. I'm right there.

They've all... everybody here has accomplished some interesting things and none of which had to do with PayPal. I think the work that David has done as the AI... David Sacks, he's incredible. He's a little bit more salt and pepper than he was back then. But I think he's actually done an incredible service to this country in helping create the legal framework for AI and for crypto to thrive.

I serve on the boards of four public companies that all started as seed or early stage investments for Sequoia: Nvidia, Unity, MongoDB, and Block. I serve on 12 private company boards. I'll continue to be an adviser to the firm. And right now, I'm doing the thing I love most, which is meeting with entrepreneurs to hear about their ideas.

Yes, I actually have a company that's presenting this Thursday.

I love this company XBOW. It's X-B-O-W. It's the person who started Copilot inside Microsoft. He's a Dutchman who was a computer science professor and he started this company less than two years ago to solve penetration testing and cybersecurity. And today if you go to HackerOne, they are number one in the world. Their software is literally the best hacker in the world less than two years after the company got started. And if you look at the rate at which AI is progressing, this thing is growing and improving at an exponential rate. And I sort of think about the possibilities of this type of technology. I love that business.

Thank you.