Latest / Tech IPO Conversations with Fexingo: Public Software Companies and Capital Markets / How Post-IPO Companies Are Using AI Compute as a Growth Strategy
Transcript
- Lucas: So you've got a software company that just raised a few hundred million dollars in an IPO. Conventional wisdom says you spend that money on sales teams, marketing, maybe a few strategic acquisitions. But a growing number of post ipo tech companies are doing something very different: they're pouring capital into AI compute. Luna: You mean buying GPUs, reserving cloud capacity — treating compute like a raw material? Lucas: Exactly. And it's not just AI startups. Look at Shopify. They went public back in 2015, but their capital allocation story has shifted dramatically. In their most recent earnings call, they explicitly said they're investing in merchant-facing AI features that require significant compute infrastructure. The market seems to like it — Shopify's up about 18 percent over the past five days. Luna: That's a big move. And Palantir — up 17.6 percent in the same period. They've been all-in on AI since last year. Lucas: Right, Palantir is a fascinating case because they were already a data analytics company. But their AIP platform — the AI integration layer — is compute-intensive. They've been spending on GPU capacity from partners like Microsoft Azure. And investors are rewarding that bet. Luna: So the question becomes: is this a smart strategic pivot or are companies just chasing the AI hype cycle? Lucas: I think it's more nuanced. For a company like Shopify, the compute spend is directly tied to product features that drive merchant revenue. If their AI tools help a store owner optimize pricing or inventory, that's a measurable return. But for other companies, it might be a more speculative bet. Luna: Speculative how? Lucas: Well, consider a hypothetical post ipo SaaS company that raises $200 million and then spends $50 million on reserved cloud compute for an AI assistant they haven't fully productized yet. That's a bet on future adoption, not current demand. And if the AI feature doesn't gain traction, you've burned capital that could have gone to sales or R&D with a clearer payoff. Luna: But isn't that the same argument people made about cloud spending a decade ago? 'Why spend on AWS when you can build your own data center?' And then cloud won. Lucas: Fair point. The difference is that the cloud migration had a proven cost-saving thesis. AI compute today is more about revenue generation — or at least, the promise of it. And the costs are enormous. NVIDIA's latest data center GPUs can cost $30,000 each, and you need thousands to train a model. Luna: So the risk is real. But the reward could be outsize if you're one of the first to scale an ai native product. Lucas: Exactly. That's why we're seeing this pattern. Companies that went public in the last few years — especially those with strong cash positions — are treating compute as a competitive moat. Luna: And if these conversations are useful for what you're building or running, listener support is what keeps this show ad-free and independent. Lucas: Yeah, it's a small way to help us keep digging into these topics. You can find us at buy me a coffee dot com slash fexingo. Luna: And we really appreciate it. So back to that competitive moat idea — are there any companies that have gone too far? Lucas: There are warning signs. A few months ago, a mid-cap enterprise software company that went public in 2023 announced they were diverting 40 percent of their IPO proceeds into a five-year compute contract. Analysts pushed back hard — they said the ROI was too uncertain. The stock dropped 12 percent in a day. Luna: Market discipline in action. Investors want to see a clear line between compute spend and customer acquisition or retention. Lucas: Right. And that's why the most successful post ipo compute strategies are tied to a specific product launch with measurable targets. Palantir, for example, has been transparent about their AIP deployments and the government contracts they're pursuing. That gives investors confidence. Luna: What about companies that haven't gone public yet but are watching this play out? Do you think it affects their IPO timing? Lucas: Absolutely. Private AI companies, especially those with high compute needs, are staying private longer to build their infrastructure before facing public market scrutiny. We've seen that with companies like Anthropic raising massive rounds instead of rushing to an IPO. Luna: So compute is becoming a strategic asset that shapes not just how companies spend, but when they go public. Lucas: Exactly. And I think we'll see more post ipo companies write compute into their growth strategy as a line item — not just an operational cost, but a capital allocation decision. The ones that do it well will separate themselves from the pack.