- Anthropic is rolling out cheaper AI models and steep government discounts to drive adoption and justify its massive IPO valuation.
- The company priced its IPO at $83 per share, above the proposed range, valuing it at roughly $1.5 trillion on a fully diluted basis.
- Third-quarter revenue jumped 47% sequentially, but price cuts and discounts trimmed about $600 million from sales, and operating losses returned.
Anthropic, the San Francisco-based AI company behind the Claude assistant, is slashing prices and introducing new products as it seeks to convince investors that its lofty valuation is justified, according to The Information. The moves come as the company prepares for its shares to begin trading on Nasdaq (NASDAQ:NDAQ) under the ticker ANT later today.
The IPO was priced at $83 per share, above the proposed $70–$78 range, raising $81.3 billion before fees. The offering included about $79.4 billion in shares sold by Anthropic itself. At the offer price, the company commands a valuation of roughly $1.5 trillion on a fully diluted basis, or about $1.1 trillion based on basic shares.
To support that valuation, Anthropic is making Claude both more capable and cheaper to use. It says its Sonnet 5.5 model costs up to 30% less for most work, while Opus 5.5 costs 40% less to run on typical workloads than its predecessor, Opus 5. Separately, the company has struck a deal offering eligible U.S. government buyers a 40% discount.
Those price cuts, however, are already weighing on revenue. Model price reductions and contract discounts reduced third-quarter revenue by about $600 million, according to a summary of the prospectus. The company reported third-quarter revenue of $16.96 billion, up 47% from the second quarter, but returned to an operating loss.
Revenue for the first nine months of 2026 reached $33.24 billion, a more than tenfold increase from $2.55 billion a year earlier. As of October 31, Anthropic’s annualized revenue run rate was approximately $81 billion—a snapshot of the current sales pace, not booked annual revenue. Roughly 79% of revenue in the first nine months came from usage-based charges rather than subscriptions.
The company has also been investing heavily in infrastructure. It spent nearly $9 billion on computing capacity in the first half of 2026, nearly four times the prior-year amount. Demand is broadening: as of September 30, Anthropic had 11,900 customers generating at least $100,000 each in trailing 12-month revenue, up from just over 1,500 at the end of 2025.
The IPO surpasses SpaceX (NASDAQ:SPCX)’s offering in size and marks a critical test for the AI industry: whether a company that builds and sells advanced models can sustain strong financial results as a public company. The offering follows a private round that valued Anthropic at $965 billion.
Competition remains intense. In a September snapshot from expense platform Ramp, OpenAI’s GPT-6 Astra accounted for about 13% of tracked enterprise AI spending, versus about 8% for Claude Fable. The GSA deal with the U.S. government extends discounted access to federal, state, local, tribal, and territorial buyers with no minimum spending commitment. It does not resolve Anthropic’s separate dispute with the Pentagon over its supply-chain-risk designation, which has prompted litigation.
For now, the immediate test is where ANT trades relative to its $83 offering price. The more meaningful test will come in subsequent earnings: whether customers acquired at lower prices keep using Claude, and whether the cost of serving them falls fast enough to improve margins.