Latest / Tech IPO Conversations with Fexingo: Public Software Companies and Capital Markets / Why Private AI Companies Stay Private Longer
Transcript
- Lucas: Palantir up fourteen and a half percent in the last five trading days. Shopify up more than fifteen percent. Meanwhile, Microsoft gained almost eight percent in the same stretch. The public market is clearly rewarding established software companies right now. Luna: So if you're a late-stage AI startup watching those returns, you'd think the IPO window is wide open. But the data says otherwise. Lucas: Exactly. We've covered this a bit in previous episodes — why companies are choosing direct listings, why they're returning IPO proceeds. But there's a deeper structural shift I want to drill into today. The AI companies that everyone talks about as future IPO candidates — Anthropic, xAI, even some of the foundation model startups — they're not just delaying. They're actively choosing to stay private, and they're raising enormous amounts of capital to do so. Luna: Anthropic's sixty-five billion dollar round last quarter is the obvious example. That's more than most large-cap IPOs would raise. Lucas: Right. And it's not just Anthropic. Just this weekend, SoftBank announced it's investing up to seventy-five billion euros to build French data centers. That's infrastructure spending that would have been laughable five years ago. The scale of capital required to compete in AI is pushing companies to stay private because they need constant, patient capital — not a one-time IPO pop. Luna: And the IPO pop itself has become less reliable. Look at the companies that did go public in the last two years — many are trading below their first-day close. The market is punishing any whiff of overvaluation. Lucas: That's the key tension. On one hand, you have these massive private raises that give AI companies years of runway without quarterly earnings pressure. On the other, you have a public market that's selectively rewarding profitable growth — Shopify, Palantir — while looking skeptically at high-burn narratives. Luna: So the question becomes: if you're a founder of an AI startup with a billion-dollar valuation, why would you go public right now? Lucas: You wouldn't — unless you need an exit for early investors or employees. And that's exactly what we're seeing with the secondary market. Companies are setting up internal tender offers, letting employees cash out without an IPO. It's a way to provide liquidity without the scrutiny of a public listing. Luna: Which brings us back to the data. Palantir and Shopify are doing well, but they're not ai first companies. They're platforms that added AI features. The pure-play AI companies are still viewed with suspicion by public investors. Lucas: And there's a structural reason for that. The cost of compute is still incredibly high. Training a frontier model can cost hundreds of millions of dollars. That creates a business model where you need massive scale before you can even think about profitability. Public markets hate that uncertainty. Luna: So you end up in a situation where the companies that would benefit most from public market capital — the AI infrastructure builders — are the ones least likely to access it. Lucas: Exactly. And that's why you see SoftBank and other sovereign funds stepping in. They can provide the patient capital that public markets won't. The irony is that the AI boom is making the private markets more important than ever, even as public markets hit new highs. Luna: One thing that strikes me is how this affects the venture capital ecosystem. If your portfolio companies stay private for a decade, how do you return capital to your own investors? Lucas: That's the unspoken crisis. Venture funds have ten-year life cycles. If an AI company stays private for fifteen years, the fund structure breaks. You're seeing more continuation funds, more secondary sales to manage this. But it's not a perfect solution. Luna: And that brings us back to the public market returns we opened with. Palantir up fourteen percent, Shopify up fifteen percent — those are real returns for public investors. But they come from companies that have already proven their business models. The AI startups haven't done that yet. Lucas: So the market is essentially saying: show me the revenue and the path to profit, and I'll reward you. Don't come to me with a story and a high burn rate. That's a healthy discipline, but it also means the IPO window for AI companies is effectively closed until they mature. Luna: Unless they find a way to go public without the pressure. Direct listings, SPACs — but we've seen how those can go wrong. Lucas: And that's the thing — a handful of listeners chip in monthly through buy me a coffee dot com slash fexingo, and that's literally what funds making this many episodes possible. No ads, just listener support. If these conversations are useful for what you're building or running, that's the only way it stays going. Luna: Yeah, it keeps us independent and focused on the topics that matter, not the ones that sell ads. Lucas: So back to the AI IPO question. The real shift is in how we think about 'going public' as a milestone. It used to be the natural endgame for a successful startup. Now it's becoming optional — and for many AI companies, it's actively undesirable. Luna: What does that mean for the next generation of public companies? If the best AI companies stay private, the public market loses some of its dynamism. Lucas: It means the public market becomes a slow-growth, steady-state index of mature businesses. The excitement, the innovation, the high-risk-high-reward bets — they all stay private. That changes the character of the stock market we've known for a century. Luna: And it changes what investors should look for. Maybe the next big public AI play isn't a startup at all — it's an established company like Microsoft or Amazon that builds AI into its existing business. Lucas: That's already happening. Microsoft's seven percent rise this week partly reflects its AI integration narrative. Amazon's up too. The public market is betting on incumbents, not insurgents, for AI returns. Luna: So if you're a retail investor, your exposure to AI comes through mega-cap tech, not through the next unicorn IPO. Lucas: Exactly. And that's a different kind of bet. You're betting on execution and distribution, not on a breakthrough model. For most investors, that's probably safer. But it also means you miss out on the asymmetric upside that venture investors get. Luna: Unless you find ways into private markets — but that's a whole other conversation about accredited investor rules and liquidity. Lucas: Which is exactly what we'll talk about in a future episode. For now, the takeaway is this: the AI IPO window isn't just closed — it's been redesigned. The companies that would have gone public five years ago now have better options. And the public market is reflecting that in its valuations of the companies that do list. Luna: So next time you see a headline about a billion-dollar AI startup raising more money, ask yourself: is this the next public company, or the next permanent private institution? Lucas: That's the question. Thanks, Luna. Luna: Thanks, Lucas.