00:00Bessemer Venture Partners is raising $5.75 billion in fresh capital, making a bigger push
00:06into growth stage investing. The firm setting aside $4 billion for growth deals as AI startups
00:11raise increasingly larger rounds and stay private longer. Joining us now, Samir Dalakia,
00:17partner of Bessemer. This is big. Actually, I think it's the biggest in the firm's history.
00:21It is indeed.
00:24Why now? What does it reflect about the environment that venture capitalists find themselves in?
00:30Yeah. Great to see you, Ed, and thanks for the question. It is a special moment in time,
00:35in human history, genuinely, in terms of the technology stage that we are at. You have this
00:41inflection happening on so many different dimensions. You guys were talking earlier with
00:45your guest about physical AI, embodied AI, intelligence. You move into autonomy within
00:52robotics. You have autonomy and manipulation. There's so much extraordinarily exciting activity
00:58going on right now. This is the time. These companies need a lot more capital than they
01:03used to need, and they stay private longer than they ever have. We think that's a permanent
01:08structural shift. It used to take five years to get a company public, six years. Now we're
01:13looking at an average of 12, 14. We have to have the capital to support these incredible founders
01:20and their companies as they scale.
01:21I think how the firm behaves and its strategy is really important because people will see
01:26the dollar headline. That's a lot of money for a venture capital fund, even for a growth stage fund.
01:31But is it the case of trying to invest as many companies as you can in a calendar year,
01:35or are you still kind of discerning and selective about where those kind of bigger checks get written?
01:41Yeah, it is. It's a huge fund in terms of dollar amounts. The phrase we use at Bessemer is
01:46disciplined conviction. So the discipline is still extraordinarily important. As big of a fund as it
01:52is, I really do want to emphasize our early stage practice is still the foundation of the platform.
01:57Always has been. We've been around for a hundred.
01:59Well, in this case, $4 billion goes to growth. So $1.75.
02:02Yeah, $1.75 continuing to focus on early stage, inception stage, siege stage, series A investments
02:09in great companies at the very beginning. That's when we identify where the world is heading. And
02:15when we pick these great founders, the growth stage is simply to say, gosh, if for whatever reason we
02:21didn't see them early, we're going to make sure we're in the very best companies on the planet.
02:25And those companies need an awful lot of capital these days.
02:27There are some blurred lines, right? You know, I've written a lot this calendar year. It might even
02:32have been last calendar year. I lose sense of time and space. But there are seed rounds or debut
02:37rounds in the hundreds of millions of dollars at valuations that, you know, historically you say
02:42that's a growth type of investment. Yeah. But it still has early stage risk.
02:47Right. It could just be four people in a garage in Menlo Park. Like that's kind of back a bit.
02:52Yeah, indeed. We do those, I would say, fairly sparingly. Like if the team is remarkable in an
02:59incredible market that we know really, really well, on occasion, we'll do one of those. But
03:04in general, we'll look for, and it tends to be in industries or markets where the capital to get
03:11the business going simply is massive. That is what requires those. But those are high,
03:17high beta investments. I would say our disciplined conviction takes us to other places more often than
03:22not. I think it would be very useful as well to understand what you are not talking about
03:27investing in. So you as a firm are invested in Anthropic. We are. You know, $4 billion more
03:33for growth investments. Your expectation is that another Frontier Labs that is not coming along,
03:39right? We have Anthropic. We have OpenAI. Yeah. There are others thinking machines, et cetera.
03:44You're looking at a different layer or at the application or physical AI, as you mentioned.
03:50Yeah. At every layer of the stack, we'll continue to find areas of opportunity. So
03:54certainly at the application layer, we're really, I've been very excited about vertical applications,
04:00a bridge for doctors, Legora, and even up for lawyers, field guide for auditors,
04:08a lease AI for property managers. So you just kind of go industry by industry. And that last mile
04:13that each of those companies provides to the industry they serve is really, really important.
04:18So we spent a lot of time at the application layer. We're very proud, $650 million investors
04:24in Anthropic over four rounds. We invested two and a half years ago in the Series D when they were
04:30a
04:31$100 million startup. So we're very proud of that investment. But how severe, sorry to interrupt you
04:37then, how severe is the change in environment today versus when you wrote that first check into
04:42Anthropic? Oh, it's extraordinary. We're living, when you think about the capabilities of the model
04:49and of LLMs two and a half years ago to where we are today, it's a vertical line, right? It's
04:55just
04:55extraordinary. And to roll that forward, I think the last time I was on, I was talking to you about
04:59how I thought intelligence as a market would be bigger than all of, it's like oil, like oil powers
05:05all the machines in the world. Intelligence will power all the humans in every organization in the
05:10world. It's actually interesting to go back and look in time when the oil industry was founded in
05:15like go back to 1870. The use case, the primary use case for oil was lighting kerosene lamps.
05:23Cars didn't come yet. Planes hadn't come yet. Plastics hadn't come yet. Those use cases for oil
05:29didn't come for decades. And in our case, the use cases for intelligence, where the models are today
05:35relative to where they were, where they're going to be in the years ahead, we can't envision all the
05:41ways that these models are going to be used. And that's what gets me excited. That's what gets my
05:45partners excited. We're investing. When I say this is the moment in time, that's what I mean. Like for
05:50the next three to five years, we think we're going to see extraordinary innovation using the
05:55intelligence of those models. And they are in a wildly different place than they are today than they
06:00were just two and a half, three years ago. I do want to address the news. So today,
06:05Dario Amo Day will join this UN Security Council briefing alongside Sam Altman. And there will be
06:12a focus on pacing the frontier again, I'm sure. One of the things that Dario advocated for in that
06:18essay was the idea of this will only work if the frontier labs do the same thing. And therefore,
06:24antitrust waivers are appropriate because that would allow them to work together. The other side of
06:29that argument is that forms a cartel. Yeah. Yeah. Please weigh in. Well, I think that whenever
06:35you see Dario, Sam and Elon agree on anything, it's worth noting. Those three don't tend to line
06:42up on things. They are very close to the edge on the frontier of what is happening. They think that
06:47coordination of safety and guardrails matters. I'm inclined to believe them and want those folks
06:55to coordinate. I love the idea of the third-party evaluators and the independent evaluators that
07:01they've talked about. Those are just pragmatic, practical steps forward. They're in a quiet period.
07:07I'm not going to comment on anything specifically to them. But broadly speaking, does it make sense
07:12for them to share best practice around safety and guardrails? I think that is great. When we invested
07:19in Anthropic two and a half years ago, they were talking about safety and trust then. This is not
07:24new news. They are nothing if not consistent. They talked about a phrase called the race to the top.
07:30They were going to set an example of how important it was to ensure safety and trust. And they've been
07:36living in it and others are indeed following suit.
Comments