Anthropic Just Filed Confidential IPO Paperwork — and the AI Mega-IPO Race Is On

A $965B company knocked on the stock market's door — Anthropic's June 1 move

On June 1, Anthropic submitted a confidential draft S-1 to the U.S. Securities and Exchange Commission for an IPO. The company even put out its own official announcement. The key is the timing. Just last week it closed a $65B Series H that valued it at $965B post-money — the world's most valuable AI startup — and before the ink was even dry, it kicked off the path to going public.

Why is this big news? Because the vague expectation that "AI startups will go public someday" has become a concrete event: actual paperwork landed at the SEC. And Anthropic planted the flag before OpenAI. The market sees both companies potentially listing as soon as the fall, with OpenAI expected to take similar steps within days or weeks. AI companies that once looked like money-losing research labs are now walking onto Wall Street's main stage.

And one number explains all of it. Anthropic's revenue run-rate has surpassed $47B — up more than fivefold from $9B at the end of 2025 in barely six months. To the suspicion that "the valuation is just a bubble," it answered with revenue.

Meet the players — Anthropic, and the confidential-S-1 device

Anthropic is an AI safety–focused research and product company founded in 2021 by OpenAI alumni. Its flagship model, Claude, earned a reputation for strength in coding, long-document handling, and safety, and rapidly penetrated the enterprise market. Over the past few months it raised $65B through Series H, expanded overseas hubs (Seoul, Milan), and exploded its enterprise customer base — graduating from "AI lab" to "large AI company." The $47B revenue run-rate is the proof.

The confidential draft S-1 device is worth explaining. To list in the U.S., you typically file an S-1 registration with the SEC. File it "confidentially" instead of publicly from the start, and you can work through the process with the SEC privately without disclosing financials, risks, or business details to the public. The key advantage: you can flexibly adjust the listing timeline and price while watching market conditions. So Anthropic has declared "we're going public," but hasn't set the share count or offering price yet. It's holding its cards to play them when the market window is best.

SpaceX is an important supporting character here. Elon Musk's SpaceX filed a public S-1 on May 20, targeting a June 12 Nasdaq listing under "SPCX," aiming for a $1.75–1.8T valuation and a raise of up to $75B. If it lands, it's the largest IPO in history. Many analysts say it's no coincidence Anthropic's confidential filing came right after.

The details — the AI mega-IPO trio by the numbers

Here are the key figures in a table.

Item	Value	Note
Anthropic valuation	$965B (post-money)	From last week's $65B Series H
Anthropic revenue run-rate	$47B	Up from $9B at end of 2025 (5x+)
OpenAI valuation	$852B	As of March
SpaceX target valuation	$1.75–1.8T	Targeting June 12 Nasdaq SPCX
SpaceX target raise	Up to $75B	Largest IPO ever if it lands
Expected listing	Fall 2026 (Anthropic & OpenAI)	Confidential review stage

Anthropic's $965B surpasses OpenAI's $852B from March — the "most valuable AI startup" crown changed hands. But the number to watch more closely is revenue. A $47B run-rate means "that many enterprises are actually paying for Claude," and it's the strongest rebuttal in the bubble debate around AI valuations.

The timing matters too. Analysts see it this way: "Once SpaceX confirms investor demand with its June 12 mega-IPO, Anthropic wants to ride that hot window in the fall." It's a strategic move to lay the listing groundwork before the investment fervor for AI and growth stocks cools. The confidential filing maximizes that flexibility.

Who gains what — Anthropic, investors, and employees

For Anthropic, going public delivers both "ammunition" and a "stamp of credibility." Training AI models costs astronomical compute, and there's a limit to sourcing it endlessly from private markets. Entering the public market grants access to a far larger capital pool, and being a listed company itself signals "stability" to large enterprise customers and partners. Cleverly, the confidential filing leaves room to tune timing and price to market conditions.

For early investors and VCs, an "exit" is finally in sight. Private shares are hard to buy and sell and valuations are largely estimates — but a listing lets you cash out stakes at real market prices. For investors in the giant rounds through Series H, an IPO is the channel to realize returns. It's also the first sign that capital betting on the AI boom is actually starting to come back.

For employees, stock options and RSUs gain a path from "numbers on paper" to "real money." That said, there's usually a lockup period right after listing, so you can't sell immediately. Still, "your stake gets a public-market price tag" is a powerful motivator and a strong tool for attracting and retaining talent. In an AI industry with a fierce war for genius researchers, "an IPO in view" is a weapon in the hiring market too.

Historical parallels — have giant-valuation IPOs always worked?

Tech history has plenty of cases of "listing with a flashy, astronomical valuation." And the outcomes were extremes.

Success — cloud SaaS sticking the landing. Companies like Snowflake and Databricks drew attention with high valuations, but earned the market's trust through solid revenue growth and clear monetization. Lesson: for a "pricey tag" to be justified, revenue growth and unit economics ultimately have to back it up. Anthropic's $47B run-rate is precisely the card meant to satisfy this condition.

Cautionary — the dot-com and WeWork bubbles. Conversely, many listed on flashy stories and giant valuations alone, then collapsed when results didn't follow. WeWork's pre-IPO valuation cratered and the IPO itself was scrapped; during the dot-com bubble, revenue-less companies vanished one after another post-listing. Lesson: the market is generous to "growth stories" for a while, but eventually asks "do you make money?" Whether AI companies' giant valuations can meet public-market expectations is still an open question.

Comparison — growth stocks that listed while losing money. Companies like Uber and DoorDash listed while in the red and endured long profitability debates. They eventually turned profitable through economies of scale and cost control — or, failing that, saw their shares suppressed for years. Lesson: AI companies carry a structural burden where "revenue can explode while profit stays distant," given enormous training/inference costs. That cost structure is exactly what investors will scrutinize most relentlessly in the Anthropic and OpenAI IPOs.

Competitor counter-plays

OpenAI looks like it got beaten to the punch, but it'll follow soon. Armed with ChatGPT's dominant consumer brand and an $852B valuation, it'll pitch the market the narrative that "we're the real leader." If both list around the same time, investors will line up revenue growth rates, cost structures, and enterprise-customer concentration side by side. We're about to witness an unprecedented scene: a direct comparison of the AI industry's #1 and #2 via public financials.

Big Tech — Google, Meta, Microsoft counter with "infinite capital, no IPO needed." They already have enormous cash flows and in-house models (Gemini, Llama, MAI), so they don't need to raise money by listing. In fact, when Anthropic and OpenAI face quarterly-earnings pressure as public companies, Big Tech can emphasize "the stamina to wage a long war" and pull in talent and customers.

Elon Musk's SpaceX/xAI are competitors from another angle. SpaceX's June 12 mega-IPO proves "investment demand for AI/space growth stocks is alive," opening a path for Anthropic — while also competing for the same investment dollars. If SpaceX absorbs limited IPO investment demand at scale first, it could affect the pricing of the fall's AI IPOs too.

So what actually changes

For AI industry workers and founders, the biggest signal is that "an exit is real." Once AI startups' giant valuations move beyond private-market estimates to public-market validation, follow-on startups get a benchmark for valuation and fundraising. At the same time, becoming a public company means baring quarterly results and cost structures — the first time "an AI company's true economics" gets revealed transparently.

For investors and general readers, you'll see "where the AI boom's money flows" verified in public books for the first time. Once a number like a $47B revenue run-rate is validated in formal financial statements, the long debate over whether the AI industry is "story" or "results" gets data attached. Still, how much the enormous training/inference expense line delays profitability remains a core variable. Whether the giant valuation can meet public-market expectations won't be clear until the fall.

For enterprises and developers using Claude, there's no big short-term change. The listing process won't immediately alter the product roadmap. But long term, being a public company brings "quarterly results" pressure that can affect pricing policy and feature priorities. Securing stable capital to invest more in model development is a positive — but it's worth keeping in mind that profitability pressure could, one way or another, show up in free/low-cost tier policies.

References

• Anthropic — Confidential draft S-1 submitted to SEC (official)
• Fortune — Anthropic files its S-1 first, but the race with OpenAI is just beginning
• TechCrunch — Anthropic files to go public
• CNN Business — Anthropic files for a potentially trillion-dollar debut
• CNBC — Anthropic confidentially files IPO prospectus with SEC