Anthropic Files Confidentially for an IPO

The AI boom is about to get a public-market test. On June 1, 2026, Anthropic — the maker of Claude — confidentially filed for an initial public offering with the SEC. The company says it hasn’t settled on the number of shares or a price, and the timing will depend on market conditions and how the SEC’s review goes. But the filing puts one of the most valuable private companies in the world on a path to the public markets.

The numbers behind the filing

Anthropic enters this process reportedly valued near $965 billion following its latest raise, with roughly $47 billion in annualized revenue from Claude subscriptions sold to businesses and consumers. That revenue figure is the headline: a company growing this fast, this large, is rare — and it’s the main reason investors are paying attention.

The most recent round was a Series H that raised about $65 billion. That’s the same round Micron took a strategic stake in as part of its memory partnership, a sign of how AI labs and their suppliers are increasingly tying their balance sheets together.

Why go public now

Going public does a few things for a company burning enormous sums on compute. It opens access to deep, liquid capital markets; it gives early investors and employees a path to liquidity; and it creates a currency — public stock — for future deals. For a business whose costs scale with every new model trained on expensive GPUs and scarce high-bandwidth memory, durable access to capital is strategic, not optional.

Anthropic isn’t alone in the queue. Analysts at Wedbush framed the filing as part of a “race to public markets” alongside OpenAI and SpaceX’s xAI, predicting these debuts could open the floodgates for the broader IPO market later in 2026.

What to watch in the prospectus

A confidential filing keeps the details private for now, but when the full prospectus lands, these are the lines that matter:

• Gross margins. Revenue is impressive, but how much does it cost to serve? Inference and training expenses determine whether scale turns into profit.
• Concentration. How much revenue comes from a handful of large enterprise customers or cloud partners?
• Capital commitments. Long-term compute and infrastructure deals — like the supply agreements reshaping the industry — show up as obligations.
• Growth durability. A $47 billion run-rate is only as good as its trajectory and retention.

The risks

The bull case is obvious; the risks are real too. Model capability is converging as open-weight models close the gap, which pressures pricing. Compute costs are volatile and tied to the same memory supercycle squeezing everyone. And a frontier lab’s value rests heavily on staying at the frontier — an expensive place to live.

The takeaway

Anthropic’s filing is a milestone for the entire AI sector: the first real chance for public investors to price a pure-play frontier lab. The revenue growth is extraordinary and the strategic relationships — from Micron to its cloud partners — are deep. Whether the public markets reward that growth or focus on the cost of producing it is the question 2026 will answer.