On March 31, OpenAI raised at a $852 billion valuation, a month after Anthropic raised $30 billion at a $380 billion valuation. These are very large numbers. Even with the recent hesitation for $100B+ startups to IPO, at these valuations and with these capital requirements, the IPOs surely are coming, right?
Another question I have is: what will the public valuation be, compared to these private valuations? Public valuations shortly after IPO are usually higher than the most recent private valuations, but will the public actually stomach these eye-watering multiples on revenue?
tldr Anthropic is likely to list first, around Mar 2027, with an estimated first-day market cap of $630 billion. OpenAI will lag, I forecast they'll list in May 2027, with a median first-day cap of $1.0 trillion.
First, let's look at IPO timing:
Forecasts from the FutureSearch in claude.ai.
While I predict early 2027 for Anthropic and mid 2027 for OpenAI, I give a >10% chance for each of OpenAI and Anthropic that they don't IPO in the next 3 years at all! Sam Altman did say he's "0% excited" to run a public company, and both have very tricky legal issues.
Mostly, though, the ability to raise ungodly amounts of private capital removes the single greatest motivation to list. Both companies raised huge rounds recently, so I don't think it would be that surprising for them to only slowly work towards IPO and land them in 2028.
Let's look at the evidence on Anthropic. It engaged Wilson Sonsini as legal counsel in late 2025 and has been talking to JPMorgan, Goldman, Morgan Stanley. But typically time from bank engagement to IPO is 6-9 months, so still needing that step (then underwriters, then S-1, and only then listing) makes the Oct 2026 timeline that has been mentioned in the press very unlikely.
Reasons they might accelerate this: several investors are pushing the company to file ahead of OpenAI, hoping to capture being the first pure-play AI stock. Its exponential revenue growth also indicates a good time to list, since such growth can't continue forever, and the multiples might be the highest sooner.
I don't think the Pentagon's "supply chain risk" designation is that serious of a barrier, though others seem to think it could derail their IPO. The Long-Term Benefit Trust structure might be an issue. Mostly, though, I think it comes down to how much (if at all) that Anthropic wants it.
Now for OpenAI. I give a median date of May 2027. Talks are more informal than Anthropic. The company only recently finalized its transition to a Public Benefit Corporation, and CFO Sarah Friar has publicly guided toward a 2027 listing. The sooner scenario would be OpenAI accelerating to beat Anthropic to the public markets.
The later scenario involves the Musk lawsuit, disputes with Microsoft over the "AGI clause" in their partnership, and the possibility that public markets simply balk at the price given the multiples.
I still think access to vast amounts of private capital is the most significant factor. Will those investors keep coughing up the tens of billions needed to keep the party going? Or more precisely: do Dario and Sam think they will?
Now let's look at the valuation post-IPO.
The forecasted Anthropic valuation is tigher, tied more closely to its actual business, whereas OpenAI's valuation is wider given variability in hype.
Prior to OpenAI's new $850 billion valuation, I forecasted it would get the bigger IPO pop. Now that that private valuation exists, I predict Anthropic will get a bigger IPO pop, assuming that the private valuation stays what it is when they IPO.
You may think, wouldn't a later IPO mean a higher valuation? Not if you believe that the most recent private valuation takes into account the expected future, which it is supposed to. My forecast is more about the difference between public and private markets, rather than the view that the most recent private valuations are inefficiently low.
For OpenAI, I predict a median first-day market cap of $1.0 trillion, but I give significant probability that it comes in lower than their most recent private valuation (and some that it's much higher). Public demand for shares seems strongly a function of how people feel about ChatGPT, which would be very positive, or plausibly not very positive, by then.
I don't think their ads business, or social media like the discontinued Sora, or even their enterprise / Codex business will be very significant factors. As FutureSearch has shown since 2024, OpenAI revenue is all about consumer ChatGPT subscriptions, and if anything, they have stumbled towards building a great agent business and ceded that territory to Anthropic. Unless revenue is trending strongly towards their lofty $100 billion goal (a scenario FutureSearch has previously described as unlikely), I don't think the public will tolerate too crazy of a valuation, as if OpenAI were a memestock.
For Anthropic, I give a median first-day cap is $627 billion, ranging from about $350 billion to $980 billion. Anthropic's current private valuation of $380 billion was priced at roughly 20 times its $19 billion ARR, so there is tons of room to grow but exponentials are really hard to model. The lower end of the range reflects a scenario where this tapers off, largely because the AI hype cycle hits a correction.
The higher end reflects Anthropic's agentic AI and safety positioning driving massive enterprise adoption, pushing ARR well above $25 billion. Strangely, I think recent experience shows a bigger IPO pop is more plausible for an enterprise company (like Snowflake) than a consumer company (like Uber), even though the new public investors are consumers, not enterprises. Prediction markets currently give about 90% odds that Anthropic reaches a $500 billion valuation in 2026, which feels overconfident to me but does make me want to push up my lower bound substantially.
I expect this is a forecast I'll have to update frequently. Or you can get an updated forecast at any time yourself by connecting FutureSearch to Claude and asking yourself.