Anthropic's $1 Trillion IPO Faces Falling Prices and $417 Billion in Bills

Anthropic is aiming for a record-breaking stock market debut at a US$1 trillion valuation, but the Wall Street prospectus it published Monday spells out three threats that could sink the deal: rivals catching up, prices collapsing, and bills that keep climbing.

No startup has ever grown this fast. In five years, the San Francisco-based firm has become an AI giant. Now it wants public money — and the prospectus makes clear it needs it.

What does the growth slowdown mean for the valuation?

Investors have already trimmed their expectations. In mid-August, shareholders were floating a valuation of US$2 trillion. Then quarter-on-quarter growth slowed from 145% in the second quarter to 45% in the third. Expectations cooled to around US$1.5 trillion — a US$500 billion haircut in a matter of months.

The growth that remains is heavily concentrated in fresh releases. More than 80% of Anthropic's revenue last quarter came from models less than six months old. That cuts both ways: each new Claude version boosts sales, but it means the company must keep shipping winners on a strict clock.

Anthropic also flags a structural weakness. "Substantially all of our revenue" comes from its AI tools, the company says. Unlike Google or Meta, there's no advertising empire to fall back on if business takes a hit. And unlike Microsoft, there's no cloud business to cushion the fall.

How close are the rivals?

The prospectus names the first threat outright: commoditization. "The degree of differentiation among advanced AI models may narrow," the company warns.

The numbers back that up. Open AI models — ones anyone can download and modify — now trail closed models such as Claude and ChatGPT by only about four months on average, according to research institute Epoch AI. Kimi K3.1 Pro, from Chinese company Moonshot, costs half as much as Claude Opus 5.5, Anthropic's current flagship.

Among paid models from Anthropic, OpenAI and Google, none has stayed on top for long.

Anthropic even acknowledges it is holding itself back. Its safety measures "could slow our capability gains relative to competitors that do not adopt similar constraints," the company says.

Why are prices collapsing?

When products look alike, prices fall. According to Morningstar, the price of having AI process the same amount of text, at the same level of performance, has dropped 97% — from US$17 at the end of 2024 to 50 cents today. That's a 34-fold decline in under two years.

Anthropic itself cut the price of Claude Opus by 20% in September. OpenAI cut the price of one of its flagship models by a factor of five in July.

The shift is already visible at Vercel, a developer platform that routes requests across AI models. In August, more than half of the text it handled ran on open models, up from 13% in April. Even so, Anthropic still captured nearly two-thirds of spending on the platform — customers may be trying cheaper options, but they keep paying for Claude.

Malik Ahmed Khan, a senior equity analyst at Morningstar, argues the price collapse doesn't necessarily shrink the market. Even a 90% price cut would not reduce total spending on AI as long as usage keeps growing at its current pace, he said.

What about the US$417 billion in bills?

Anthropic's costs, unlike its prices, are largely locked in. Each new Claude model has to be bigger than the last, trained on more data using more computing power.

Since late 2025, the company has committed to paying at least US$417 billion to rent chips and servers. That is nearly 13 years of sales at its first-half 2026 pace — and the total could nearly double if cancellable or planned contracts are included. These commitments are due no matter the state of its business, Anthropic warns.

The bet involves the entire industry. Google and Amazon are at once Anthropic's shareholders, its suppliers and its resellers. If AI turns out to be less profitable than expected, the International Monetary Fund fears a "triple whammy" from these intertwined investments: falling stock prices, debts that are harder to repay, and promised revenue that never materializes.

The independent research firm New Constructs was blunter, calling the listing "an unprecedented test of investor gullibility."

So what changes for investors?

Anthropic is counting on money from the stock market, which is not unlimited. SpaceX, Elon Musk's rocket and AI company, raised record sums there in June. Google and Meta, which are pouring fortunes into data centres, are also seeking fresh capital.

OpenAI, whose chief Sam Altman called a listing "ill-advised" this year, has chosen to wait until 2027 — giving it time to see exactly how Wall Street treats its biggest rival. For anyone weighing the IPO, that's the real test: whether public investors will fund US$417 billion in fixed bills for a company whose prices are falling 97% and whose quarter-on-quarter growth just dropped from 145% to 45%.