Anthropic heads to Nasdaq: up to $90.4 billion raised at a $1.02 trillion valuation

The company whose models now underpin thousands of products is going public. Anthropic, the developer of Claude, has begun its road show: meetings with investors where it pitches its story ahead of a Nasdaq listing. If you build a product on Claude, this isn't just financial news. It describes the conditions under which your supplier will operate.

What exactly is being sold

The public offering covers 825 million Class A shares at a price of $70 to $78. Of these, 801,351,354 shares are sold by the company itself, and 23,648,646 belong to existing shareholders. Anthropic will receive nothing from the sale of those. If all shares go at the top of the range, the public portion will bring in about $64.4 billion. A $26 billion private placement is added to that, bringing the total to as much as $90.4 billion. The underwriters, Goldman Sachs and Morgan Stanley, also received a 30-day option on 123.75 million additional shares. The ticker will be ANT.

The anchor investors are Google, Nvidia, AMD, Broadcom, Micron, Microsoft and TSMC. Note that nearly all of them also supply Anthropic with chips, memory or cloud capacity. This is precisely the kind of "circular" financing that skeptics criticize.

The company's valuation reaches $1.02 trillion. In May, a private round valued it at $965 billion, while the $1.5 trillion discussed earlier was deemed too expensive by investors, who demanded a discount. For scale: the largest IPO of the year so far remains SpaceX, which raised $75 billion in June at a valuation of almost $1.75 trillion.

The money is there, but the obligations are huge

In the third quarter, Anthropic's revenue grew 47% from the previous quarter to $16.96 billion, and its net loss narrowed. The company attributes the growth to demand for Claude, especially in industry solutions for healthcare, finance and the legal sector. But the same metric grew 144% in the second quarter, so the pace has slowed noticeably. For an investor paying for growth, this is an important signal.

The main risk is buried in the prospectus: compute capacity commitments to partners such as Amazon and Google total nearly $900 billion. Part of this future infrastructure depends on financing that Anthropic does not control. If revenue doesn't grow fast enough, such contracts will become a heavy burden.

The market backdrop is far from ideal. Interest rates are high, investors increasingly doubt the sustainability of the AI investment cycle and fear data-center overcapacity. On Thursday, the Nasdaq posted its sharpest one-day drop in 19 months. The triggers were Democratic gains in the US midterm elections and concerns that data-center construction will face tougher regulation. The IPO itself has already been postponed more than once, and on Monday evening the deal details were briefly sent to banks and immediately withdrawn.

What developers and product managers should do about it

There's no need to panic: a company with roughly $17 billion in quarterly revenue is not a startup that will shut down tomorrow. But public status changes the rules. Quarterly reporting and margin pressure will arrive. How this will affect API prices, limits and terms is not yet known. This is our assumption, not a forecast, but it is still sensible to build such a scenario into your plans.

A few practical steps:

• Keep an abstraction layer over your model provider so that switching from Claude to another model takes days, not months.
• Build your own test suite for quality evaluation. Without it, you won't be able to compare alternatives fairly when the time comes.
• If Claude is business-critical, lock in prices and SLAs in an enterprise contract rather than relying on the public price list.
• Read the prospectus section on risk concentration and dependence on partners. It is the best description of what your supplier stands on.

Limitations

The share price is given as a range, final terms may change, and some information comes from Reuters and the company's official announcement. The market's reaction to the listing will show whether it is ready to pay such sums for AI companies at current interest rates. For you, the main takeaway is simple: your model provider has received public money, and with it public obligations, and it is worth checking in advance how ready you are for that.