# The labs are going public, and the public isn't sold Anthropic just filed to go public at a valuation near a trillion dollars, with OpenAI close behind. At the same time, 57% of Americans say AI's risks outweigh its benefits — while using it more every month. Two things worth thinking about: what happens to a company's 'safety-first' principles once a stock price depends on relentless growth, and what it means that the foundation you build on is now run for shareholders who priced in a future that hasn't arrived. This isn't market commentary. It's about the ground shifting under everyone building on these models. - **Author:** Fedor Chishchin - **Published:** 2026-06-08 - **Tags:** business - **Source:** https://fedorthinks.com/blog/the-labs-go-public --- Two facts from this spring, side by side. First: Anthropic [filed to go public on June 1](https://www.cnbc.com/2026/06/05/tech-download-anthropic-ipo-ai-valuations.html), after a round that valued it near **$965 billion** — bigger than OpenAI, which is right behind it. Second: in a national poll, **57% of Americans said AI's risks outweigh its benefits**, with only [26% feeling positive about it at all](https://www.fool.com/investing/2026/04/22/57-of-voters-believe-ai-risks-outweigh-the-benefit/). The companies building AI are about to become some of the most valuable public companies on earth, and the public that uses their products mostly doesn't trust them. That gap is worth more than a shrug, because the IPOs change something real for everyone who builds on these models — not just for traders. When the foundation under your product becomes a public company, the incentives that shape it change, and they change in directions you should see coming. ## What you're actually buying at a trillion dollars Start with the valuation, because it tells you what kind of company you're now depending on. Anthropic's revenue grew astonishingly fast — to a reported $47 billion run-rate — but no normal accounting justifies a near-trillion-dollar price. What's being priced isn't the business as it exists. As one analysis put it, these companies are [asking you to price what they believe humanity becomes tomorrow](https://fortune.com/2026/06/02/anthropic-ipo-openai-valuation-ai-bubble/) — a "narrative premium," with open comparisons to the dot-com era. A valuation built on narrative is a valuation that has to keep _feeding_ the narrative. Growth can't just be good; it has to be spectacular, every quarter, forever, or the story cracks and the premium evaporates. That pressure doesn't stay in the boardroom. It flows downhill into pricing, into which features get shipped, into how much compute they're willing to subsidize — into all the things you quietly depend on staying stable. ## The mission meets the earnings call Here's the tension I'd actually think hard about. These labs sold themselves, in part, on restraint. Anthropic literally [walked away from a Pentagon deal](/blog/your-model-has-values-baked-in) on principle. That posture — "we'll refuse the harmful thing even when it pays" — is easy to hold as a private company answerable to mission-aligned investors. It is much harder to hold when you're a public company and a missed revenue number wipes out tens of billions in market value on a Tuesday. I'm not predicting these companies will abandon their values. I'm pointing out that the structure changes the pull on them. A public company has a legal and practical duty to its shareholders, and shareholders who paid a narrative premium expect the narrative delivered. Every "no" — every refused use case, every safety gate that slows growth — now has a price tag attached that didn't exist before. The values from [my last piece](/blog/your-model-has-values-baked-in) aren't just the maker's choice anymore; they're a choice the market gets a vote on. ## The strangest part: distrusted and unavoidable Now layer in the public mood, because it's genuinely odd. People are wary — 57% think the risks win — and yet usage keeps climbing: ChatGPT's user share rose from 48% to 56% in a few months. The same study found comfort with a local data center [collapsed from 69% to 35%](https://www.cnbc.com/2026/04/15/public-opinion-ai-data-centers-anthropic-openai-ipo.html) even as the build-out accelerated. The public is using more of a thing it trusts less, and resenting the infrastructure that makes it possible. That's an unstable place to build a trillion-dollar business on. It means the political and social licence for AI is thinner than the valuations imply — and when sentiment is that sour, regulation, backlash, and cost pressure can arrive faster than the growth story assumes. You don't have to predict a crash to take the lesson: the ground these companies stand on is shakier than their price tags suggest. ## What it means for you You're not investing in the IPO; you're building on the company. So the practical reads: - **Don't assume today's prices are permanent.** Some of what you pay is subsidized by growth-stage economics. Public-market pressure to show margins can push prices toward the real cost of compute. Build with that [meter in mind](/blog/the-meter-was-always-running), and stay swappable. - **Don't assume today's posture is permanent either.** A model's refusals and safety stance can shift as the company optimizes for growth. Keep the guarantees that matter to you [on your own layer](/blog/your-model-has-values-baked-in), not borrowed from the lab's current mood. - **Watch the narrative, because it's load-bearing.** These valuations depend on a story staying intact. If it wobbles, the effects — funding, pricing, consolidation — reach all the way down to your stack. Don't bet your product on any single lab's invincibility. ## The bottom line The labs going public is the moment AI stops being a research adventure and becomes a set of trillion-dollar companies answerable to shareholders who paid for a future on credit. That's not inherently bad — public scrutiny can be healthy — but it changes the forces acting on the models you build on, and it does so while the public that uses those models is, by the numbers, deeply ambivalent about them. So watch this less like an investor and more like someone whose house sits on this land. The valuations say these companies are the most certain bets in the economy. The polls, the data-center fights, and the narrative-premium warnings say the ground is less solid than that. Build on it — you don't have much choice — but build the way you would on shaky ground: keep your footing portable, keep your own guarantees close, and don't mistake a confident price tag for a stable foundation.