Hello everybody, thank you for being here this morning in Davos. Bill, come on up. I want to introduce — my name is Dan Primack, Business Editor at Axios, and this is Bill Ford, Chairman and CEO of General Atlantic.
Great, hey Dan, good to see you. So thanks for having me.
So Bill, let's start here. There's been so much talk just over the last couple days in Davos about the AI cycle, and you know, we're kind of in year three of this AI cycle. Is it an AI hype cycle or just an AI tech cycle?
Well, let's just step back for a moment. And this is the fifth big tech cycle of my career. I think about the PC cycle of the '80s, the internet cycle of the '90s, the mobile cycle of the 2000s, and then the cloud cycle of the 2010s. And now we have AI, which might be the biggest of all of them. The first thing to think about is that these cycles play out over decades and not years. And so the opportunity that's going to be created from AI for investors and users is going to be over the long, long term. But like all of those things, there's a lot of enthusiasm early on, and there's, you know, which I think leads to people might call it a hype cycle or whatever, but valuations get pushed up to reflect the enthusiasm of what's happening. And I think that's happening now. A lot of capital has flowed in again because the long-term potential is very, very real. So in some ways it's a treacherous time for investors because...
Is it more treacherous when you talk — you're right, there's been past cycles, but the quantum of capital that's being invested in some of these companies dwarfs anything we've ever seen in the private market.
Absolutely. And there's three or four concerns about it. You know, one is the amount of capital being invested — it's very hard to pick winners and losers right now. And the thing that's different about this cycle versus the internet cycle is that the incumbents have a huge advantage because of the capital intensity of what has to be built, number one. The ability to harness engineering talent — you can't do this with three engineers, you need a heck of a lot more than that. And the availability of data. So you start with the big guys having a tremendous edge. So if you're backing an attacker, like a venture capital firm would do, or a growth equity firm like General Atlantic, you've got to assess what is the impact of the big incumbents. I mean, whether it's Google or Meta, Amazon, you know, Microsoft — again, I believe they have a big advantage. So I think we have to wait a bit. I think we're going to see some application companies come that will build on the infrastructure that's getting laid down now and will be good investments. But you sort of need to be active but be patient at the same time.
I was curious to pick up on something you said at the beginning. You said this may be — of those five shifts from your career, five big tech shifts in your career, this might be the most significant. Do you think this is going to be the most significant?
I do, actually. I think it has the ability to more profoundly impact more industries. You think about healthcare, education — two industries that have been hard to change through technology. I think AI has the ability to do that, and so I think it could be the most profound.
One thing that General Atlantic, a growth equity firm, often invests in later-stage private companies. Obviously you have a lot of capital under management. You have not yet done one of these foundational LLM — you know, OpenAI, Anthropic sorts of deals. Why not?
We have done one, actually — Runway, which is basically doing LLM for the creative industries, where they're trying to help advertising firms and creative departments internally build presentations, ad campaigns through a creatively oriented or vertical LLM. So we are doing a bit, but we haven't gone into the big ones. I think again it goes back to what I was saying before: capital intensity, valuation, and wanting to really have a better understanding of how do they fit vis-à-vis the big guys who are all coming at this in such a large way.
I am told you now have an AI on your investment committee. Is that correct?
We have a non-voting member of our five-person investment committee. She's the sixth member. We call her Ada. And what she's done is look at 45 years of investment history, investment memos, deals we did, how they turned out, financial models, incorporating all that information. And think of it as a private LLM. I mean, we're using Llama as our third-party open-source platform, but trying to use all that data to actually vote on our investments. We all vote on all of our investments at our investment committee, and we've asked Ada to vote as well, even though it's not an official vote.
Has Ada ever disagreed with the human voters?
That's interesting. When has that disagreed with you?
I don't know if she's disagreed with me, actually. We're learning — a very good question, a very good question, actually. But there's a human factor in investing that Ada can't take into account.
If Ada disagreed with you, what would that look like? But also, would you sit there and go, 'Wait a minute, maybe I'm wrong,' as opposed to if one of your partners [disagreed], they'd be like, 'No, I know he's wrong.'
No, I think all five of us take it seriously with Ada. And here's the thing: what Ada might do — Ada would say, 'When I look at your history, when I look at external data, I don't see margin expansion that's in the investment memo. I don't see margins going from 5% to 25% on that time frame in that industry in that structure, so I don't really believe it.' To the extent you guys are now, we might be saying, 'Yeah, but this entrepreneur is special, and I think they can do something, and the business may pivot to a different direction.' Ada can't take that into account yet — doesn't have the interaction with the individual. We can still do that. So when we differ, it's her saying, 'I don't like the market size relative to valuation, I don't like the margin expansion story, and so I'm flashing caution to you.' And a lot of times we've hit those points, and I'd say 80, 90% of the time we're voting the same way. But there's a couple of times you say that human factor and what makes an entrepreneur special — we just heard it, you know, in the conversation before about what they can do — you've got to factor that into your decision.
Could you see in 10 years Ada having a vote, an actual vote?
And you think LPs would be good — how would LPs react to finding out that they're in part investing in an algorithm?
Well, they want good decisions. I mean, at the end of the day they want performance. And I'd be the first to say, if Ada can do better, cut their fees, because now...
No, no way, no, no, no, come on, not that, Dan. Dan, don't go there, that's bad.
No, no, but if we can make better decisions, right, because of Ada, they're happy.
But they can't cut their fees.
No, we don't want to do that.
Let me actually ask about — let me go into something else a little. Actually, you know, let me stick on one AI thing, and then we've got Chuck Robbins after this, I'm going to ask him too. A big announcement last night coming out of Washington about this half-trillion-dollar investment into US data centers. I'm wondering, when you think of that, A, what do you think that means for the startup environment, and B, do you think it's real or do you think it's announcement? In other words, it's real work to build these things, it's not just flipping on the map, brick and mortar here.
I think the way to look at all of these announcements, including this very important one from last night, is that if you look at the large tech companies — put Oracle in the mix now, and SoftBank obviously has terrific insight and expertise — I think they're convinced this is real, that this AI transformation we're talking about is a fundamental change. It will rewrite technology, rewrite business and society to some level. And even though we might not have all the evidence yet, we've got to invest for the future now, and we've got to make that bet. And so you see people at the large incumbents saying we're going to put down large investments on the anticipation that this is going to be there. And yesterday was a great example of that.
And you guys, General Atlantic, has investments on the data center side, on the infrastructure side. Do you feel you're now competing with the federal government on this?
I don't know if we're competing, but we have a digital infrastructure business. We're doing it more outside the US where we have real strength, and we're building data centers, digital infrastructure, and we're in conversations with all the large hyperscalers who want the capacity, and they're saying to us, we're tripling the size of a data center business we have in Malaysia, for example, because ByteDance is saying to us, we want more capacity for what they're doing in both the traditional TikTok business but also AI. And so we're going to participate in our own way, selectively, but going head-to-head with the big guys is probably not a formula for success for us.
I said I want to talk about LPs. Let me ask — the exit environment for private equity and for venture capital, growth equity over the last couple years has been sluggish to say the least in terms of both M&A and IPOs, and we talked about this a bit downstairs. So on the M&A side, there's been this argument that DOJ and FTC scared all the big guys off from making acquisitions. I get that. Why was the IPO market so slow when the public equity market was so strong?
I think you asked that question and it's a great one and it's a conundrum. But I really do think that this sense of overregulation from the SEC, even if it wasn't specifically targeted at the IPO market, caused entrepreneurs to hesitate about going public. And if you think about it — and we were talking about this — there's a lot of private capital available for companies to stay private longer. A lot of CEOs of private companies got comfortable with private ownership and saying, why do I want to subject myself to the rigors of the public market if I can wait a little bit longer? And we were talking about — while not directed towards crypto, towards market structure — coming out of the SEC was quite negative towards the private sector and quite negative towards markets, and it made it easy for companies, I think, to say, let's just wait and see. And so we have three of the worst IPO years on record for the last three years. Now we have 26 companies in our pipeline who are actively getting ready to go public.
And again, as we were talking about briefly, do you believe they are all really, or do you believe some of the CEOs are saying, well, GA says we should be, let's tell them we're getting ready to go public?
Well, I think there's a gentle hand on their back from a lot of private equity firms saying, you know, we need to return capital to our investors, you know, this is a partnership, and work with us on this. But I really think — I've thought a lot about this — and I think it really is a triple win, the IPO market, because number one, it allows companies to raise capital. They do need to raise capital. Yes, they can do it privately, but they can also do it publicly. Second, it gives a chance for us to return capital to our investors. And what I love about the public market is people can make different decisions — some can hold their investment, some can sell their investment. And the last thing is, I think there's a real point about the public participating, public investors in the best growth companies. I mean, think of the public being invested in Apple over the years, Amazon over the years, Google over the years, right, all of which of course went public much younger than most of the companies you're talking about.
