OpenAI Files for IPO, Days After Anthropic

Two S-1s in One Week

OpenAI submitted a confidential S-1 registration statement to the SEC on June 8, 2026 — just days after Anthropic filed its own confidential IPO paperwork. Two of the three most valuable AI companies on Earth are now formally on the path to public markets, and the timing is almost certainly not a coincidence.

A confidential S-1 filing (technically a "Draft Registration Statement" under the JOBS Act) lets a company begin the SEC review process without immediately making its financials public. The company's numbers stay private until at least 15 days before a roadshow begins. It's the standard playbook for high-profile tech IPOs — Stripe, Instacart, and Reddit all took this route.

What's unusual here is the pacing. Anthropic and OpenAI filing within the same week suggests both companies — and their respective banks — see the same window opening and neither wants the other to define the narrative first.

Why Now?

Several forces converged to make mid-2026 the moment:

• Revenue acceleration. OpenAI reportedly crossed $10 billion in annualized revenue earlier this year, per multiple press reports (Reuters, The Information). Anthropic hasn't disclosed comparable figures, but its enterprise traction — the $1.8 billion Akamai deal, the $200 million Gates Foundation partnership, and the SpaceX Colossus 1 compute lease — signals a business that's scaling fast enough to withstand public-market scrutiny.
• Valuation momentum. OpenAI's last private round reportedly valued it north of $300 billion (per TechCrunch's June 8 reporting on the filing). At some point, the secondary market becomes less efficient than public equity for both liquidity and future capital raises. Both companies appear to have crossed that threshold.
• The SpaceX signal. SpaceX has been preparing its own IPO path, and the broader market appetite for "generational tech" listings is clearly strong. If SpaceX — a company that famously resisted going public — is moving, it resets the calculus for every other mega-cap private company.
• Compute capital requirements. Training frontier models requires billions in GPU infrastructure. Both OpenAI and Anthropic have been raising massive private rounds, but public equity markets offer a deeper, more permanent capital base. Anthropic's recent infrastructure deals alone (Akamai, SpaceX) represent multi-billion-dollar commitments that need long-term financing.

What Confidential Filing Actually Means

It's worth being precise about what has and hasn't happened. A confidential S-1 is the starting gun, not the finish line. Here's the typical timeline:

Stage	What Happens	Typical Duration
Confidential filing	S-1 submitted to SEC; financials stay private	We are here
SEC review	Multiple rounds of comments and revisions	2-4 months
Public S-1	Full financials disclosed (15+ days before roadshow)	1-2 weeks before roadshow
Roadshow	Management pitches institutional investors	1-2 weeks
Pricing and trading	Stock begins trading	Day after pricing

So we're likely looking at Q4 2026 or Q1 2027 for actual trading — assuming no delays, market disruptions, or strategic pauses. Both companies could also withdraw their filings at any point. Filing doesn't obligate you to go public.

The Strategic Chess Match

My read: the near-simultaneous filings aren't coincidental, and they create an interesting dynamic.

Whoever prices first gets to set the valuation benchmark for "frontier AI company." If OpenAI lists at a $300B+ market cap, that anchors how investors value Anthropic — and vice versa. Both companies have strong incentives to go first if they believe their story is stronger, and to go second if they think the other's IPO will validate the category and lift all boats.

There's also the talent dimension. Public equity means liquid stock compensation. The AI talent market is brutal, and both companies are competing for the same researchers and engineers. Whoever goes public first can offer something private-company stock can't: liquidity without waiting for a secondary sale or acquisition.

The real race isn't who files first — it's who prices first. That company gets to define what a frontier AI business is worth on public markets.

What We Don't Know Yet

Confidential filings are, by definition, opaque. Here are the biggest open questions:

• OpenAI's corporate structure. OpenAI has been converting from its unusual capped-profit structure to a more conventional for-profit corporation (per OpenAI's own announcements in 2025). The S-1 will need to explain this clearly to public investors. How that transition is presented — and whether any legacy nonprofit obligations remain — could materially affect valuation.
• Anthropic's unit economics. Anthropic has signed enormous compute deals (the SpaceX Colossus 1 lease alone involves 220K+ GPUs, per Anthropic's announcement). Investors will want to know the cost structure: what does it cost Anthropic to serve Claude at scale, and what are the margins?
• Revenue mix. How much of each company's revenue comes from API access vs. consumer subscriptions vs. enterprise contracts? The answer shapes which public comps investors will use — and therefore the multiple.
• Profitability timeline. Neither company is believed to be profitable. The S-1 disclosures will reveal burn rates, gross margins, and — crucially — whether either company has a credible path to profitability or whether this is a "grow now, profit later" story.
• Microsoft's role. Microsoft owns a significant stake in OpenAI and has committed billions in infrastructure. The S-1 will need to detail this relationship, including any revenue-sharing arrangements, and how Microsoft's cloud partnership affects OpenAI's independence as a public company.

How This Compares to Previous Tech IPOs

If OpenAI goes public at anything near the $300 billion valuation range reported by TechCrunch, it would be one of the largest tech IPOs in history. For context:

• Meta (Facebook) went public in 2012 at roughly $104 billion.
• Alibaba listed in 2014 at about $168 billion.
• Arm Holdings returned to public markets in 2023 at approximately $54 billion.

A $300B listing would dwarf all of these at the time of their IPOs. The question is whether public-market investors will accept frontier-AI valuations that are priced on potential rather than current earnings. The enterprise AI market is growing fast, but "growing fast" and "worth $300 billion today" are different conversations.

Anthropic's valuation is less publicly discussed, but its most recent private round reportedly valued it at roughly $61 billion (per early 2025 reporting). That number has almost certainly moved significantly upward given the infrastructure deals and revenue growth since then.

What This Means for the AI Industry

Public listings force a level of transparency that private companies can avoid. Once these S-1s go public, we'll see — for the first time — audited financials for the two most important AI companies outside of Google DeepMind (which is buried inside Alphabet's reporting).

That transparency cuts both ways. It could validate the AI investment thesis if the numbers are strong, attracting even more capital to the sector. Or it could trigger a correction if the revenue doesn't match the hype — similar to how WeWork's S-1 disclosure in 2019 exposed a business that couldn't justify its private valuation.

The honest take: I think both companies are in fundamentally stronger positions than WeWork ever was. They have real enterprise customers paying real money for products that are clearly useful. The question isn't whether the businesses are real — it's whether they're worth the specific multiples the private market has assigned.

For developers and users

Public companies face quarterly earnings pressure. That can change how products evolve — more focus on revenue-generating features, potentially less tolerance for expensive research that doesn't have a clear commercial path. It also means more data about pricing, usage, and strategy will become public, which is actually good for anyone trying to make informed decisions about which AI platform to build on.

For competitors

If both IPOs succeed and both companies raise significant public capital, the competitive moat around frontier AI gets deeper. The cost of training and serving state-of-the-art models is already prohibitive for most companies. Public-market access to billions in fresh capital widens that gap further.

This is particularly relevant for open-source players like Meta (with Llama) and Mistral, as well as newer entrants like Recursive AI and xAI. The question isn't whether open-source can compete on model quality — it's whether it can compete on the infrastructure spending required to stay at the frontier.

The Bottom Line

Two confidential S-1 filings in one week is the clearest signal yet that AI's startup phase is ending and its public-company phase is beginning. The actual IPOs are still months away, and a lot can change between now and then — market conditions, regulatory developments, or even a strategic decision by one company to let the other go first.

But the direction is set. OpenAI and Anthropic are both headed for public markets. When the full S-1 documents become public, they'll contain the most detailed look we've ever had at the economics of frontier AI. That data will reshape how everyone — investors, developers, competitors, and regulators — thinks about this industry.

I'll be watching the SEC's EDGAR database. You probably should too.