Some Anthropic investors discussed a valuation of around $1.5 trillion at the company’s 14 October investor meeting for a planned November share sale, down from roughly $2 trillion previously expected, Reuters reported, citing three people familiar with the discussions.
Key points
- The $1.5 trillion figure was discussed by some investors, rather than set as an IPO price.
- Anthropic told prospective investors that third-quarter revenue reached $16.8 billion and that it made an adjusted profit.
- Reuters put quarterly revenue growth at 45%; the Financial Times put it at 50%.
- Anthropic has committed at least $518 billion to computing and infrastructure spending over a decade.
Investors discuss $1.5 trillion in San Francisco
The valuation discussion followed Anthropic’s financial presentation at its San Francisco headquarters. Current investors and prospective buyers of shares in the planned offering were among those weighing a lower figure, four people familiar with the matter told Reuters. The people spoke anonymously because they were not authorised to discuss the meeting publicly.
Attendees met chief executive Dario Amodei and finance chief Krishna Rao. They were mostly among Anthropic’s most prominent investors. The discussion took place as investors assessed the cost of financing AI development alongside the company’s latest growth figures.
Reuters reported that some investors were reassured by Anthropic’s revenue and adjusted profit, while others were disappointed by slower growth from one quarter to the next. Investors view the proposed share sale’s reception as important to the near-term outlook for the wider AI market, ahead of a potential OpenAI listing next year.
Anthropic reports $16.8 billion in third-quarter revenue
Anthropic told prospective investors that third-quarter revenue reached $16.8 billion, up 45% from the previous quarter, as AI Affairs previously reported. Reuters put second-quarter revenue at $11.6 billion, more than double the first-quarter figure.
The Financial Times reported a 50% third-quarter increase and put second-quarter revenue at $11.5 billion. It also reported third-quarter revenue of $16.8 billion. Both accounts described a slower sequential growth rate than the more-than-doubling reported for the second quarter.
Anthropic said it earned an adjusted profit in the third quarter, its second consecutive quarter on that measure. Executives also put operating margins at 11% and said the company was cash-flow positive, according to the Financial Times. They warned investors that those figures could change as Anthropic increased investment in computing and biomedical research.
For context on the scale of that investment, Reuters reported that Anthropic’s revenue rose twelvefold to nearly $4.6 billion last year, while its operating loss exceeded $8 billion excluding writedowns of liabilities largely connected to earlier fundraising. Reuters also reported that an IPO prospectus it saw in September put Anthropic’s 2025 loss at $42 billion and projected further spending on AI infrastructure.
Compute revenue rises as clients develop models
Anthropic earned $1.70 in revenue for each $1 spent on computing power in the third quarter, according to figures reported by the Financial Times. Yahoo Finance put the comparable 2025 figure at roughly $0.63, based on $7.33 billion of compute spending and about $4.6 billion in revenue.
The Financial Times reported that Meta and Microsoft, two large Anthropic clients, now rely more heavily on models they develop themselves. Anthropic also reduced distillation activity, in which other AI developers pay to use a model’s outputs to train their own systems. Four people told the Financial Times that such activity had accounted for 5% to 7% of Anthropic’s revenue.
Anthropic and OpenAI have both lowered prices amid competition from lower-cost Chinese developers, according to the Financial Times account. Anthropic priced Claude Haiku 5.5 about 75% below its predecessor in October. Those changes sit alongside the company’s commitments to the computing capacity used to train and operate its models.
Anthropic has committed to spend at least $518 billion over a decade on cloud services, computing and infrastructure, Reuters reported.
