Anthropic is preparing a record IPO, but admits three main threats

A record debut and cooled expectations

Anthropic, a San Francisco-based company that has grown over five years into an AI giant valued at $1 trillion, is preparing a record-breaking stock market debut. This is reported by AFP, citing the prospectus for Wall Street published on Monday,in which the company itself lists the key risks: competitors are catching up, prices for text processing are collapsing, and infrastructure bills keep growing.

Slowing growth and dependence on a single product line

Investors have already reacted to the dynamics: while in mid-August shareholders were discussing a valuation of $2 trillion, after quarterly growth slowed from 145% in the second quarter to 45% in the third, expectations dropped to roughly $1.5 trillion. The company acknowledges that 'virtually all revenue' comes from its AI tools: unlike Google or Meta, it has no advertising business, and unlike Microsoft, no cloud division to lean on during a downturn.

Open models are advancing, prices are falling

The first risk named in the prospectus is commoditization: according to the company, 'the degree of differentiation between advanced AI models may narrow.' According to the Epoch AI research institute, open models that can be downloaded and fine-tuned lag behind closed ones like Claude or ChatGPT by only about four months on average. Among them is Kimi K3.1 Pro from China's Moonshot, which costs half as much as the flagship Claude Opus 5.5. Anthropic adds that its own safety measures 'may slow capability growth relative to competitors that do not impose similar restrictions.'

Price war and market behavior

When products are similar, prices fall: according to Morningstar, the cost of processing the same volume of text at the same quality level has dropped by 97% — from $17 at the end of 2024 to 50 cents. In September, Anthropic cut the price of Claude Opus by 20%, while in July OpenAI made one of its flagships five times cheaper. On the Vercel platform, which distributes requests across models, the share of open models in processed text grew from 13% in April to more than half in August — although Anthropic still accounted for nearly two-thirds of spending on the platform. Morningstar analyst Malik Ahmed Khan believes that even a 90% price drop will not reduce total AI spending as long as usage keeps growing at its current pace.

$417 billion in obligations and the bet of an entire industry

Anthropic's costs, unlike prices, are largely fixed: each new Claude must be larger than its predecessor and trained on more data and compute. Since the end of 2025, the company has committed to paying at least $417 billion for the rental of chips and servers — nearly 13 years of revenue at the pace of the first half of 2026 — and with terminable and planned contracts, the amount could nearly double. These obligations, the company warns, must be paid regardless of the state of the business. The bet also involves other players: Google and Amazon are simultaneously shareholders, suppliers, and resellers of Anthropic. The IMF fears a 'triple shock' from such intertwined investments if AI turns out to be less profitable than expected: falling share prices, difficulty servicing debt, and unfulfilled promised returns. The independent research firm New Constructs called the offering 'an unprecedented test of investor gullibility.'

Who is going public, and when

To pay its bills, Anthropic is counting on money from the stock market, which is not unlimited: Elon Musk's SpaceX raised record sums there in June, while Google and Meta, pouring fortunes into data centers, are also seeking fresh capital. OpenAI, whose chief Sam Altman called a listing 'ill-considered' this year, has decided to wait until 2027 — which will give it time to see how Wall Street receives its main competitor.