Morning Edition · Friday, October 16, 2026Published at 2:26 AM EDT · New York
The company reported a second straight quarter of adjusted profitability and $1.70 of revenue for every $1 spent on compute, but slowing growth has some investors discussing a $1.5 trillion valuation rather than $2 trillion.

According to the Financial Times, as summarized by AI Post, Anthropic generated $16.8 billion in revenue in the third quarter and was profitable on an adjusted basis for the second quarter in a row. SiliconANGLE reports that the 45% growth is quarter over quarter. Executives presented the figures to investors at the company's San Francisco headquarters as it prepares for a possible initial public offering (IPO) in November.
The figures that matter most for valuation concern unit economics. Anthropic told investors it was cash-flow positive and earned $1.70 in revenue for every $1 spent on compute in the quarter. The same reporting says growth is slowing from earlier in the year and that annualized revenue is expected to land below the closely watched $100 billion mark.
The listing price itself is under debate. Earlier reporting said Anthropic could begin formal marketing as soon as the week of November 9 and list before Thanksgiving on November 26, with prospective investors putting fair value at $1.8 trillion to $2 trillion. Since then, some current and prospective investors have discussed roughly $1.5 trillion instead. The research firm New Constructs has argued the company is worth $150 billion. That same coverage lists the pressures on Anthropic: OpenAI's GPT-6 Astra regaining enterprise traction, price pressure from Chinese and Meta open-weight models, rising infrastructure costs, and safety incidents involving AI agents.
A Russian-language AI channel summed up the competitive picture with the line "Where is Qwen? Where is Kimi? Anthropic had too good a quarter". Coding and enterprise revenue have so far gone mostly to closed US labs, even as open-weight Chinese models win on benchmarks and price.
Part of a tracked trend
AI Hype Cycles and Funding Narratives
As capital floods AI, the narratives labs use to raise money and shape rules face growing public scrutiny, and the market increasingly separates verifiable capability and revenue from rhetoric on both the bullish and the cautionary side.
Start a discussion in Townsquare.
More from this edition
Anthropic, its underwriters and its private investors gain if buyers price the November initial public offering (IPO) on these margins. The investors include those in the May Series H at a $965 billion valuation. Its cloud and chip suppliers gain if the compute ratio justifies more spending.
CNBC, Investing.com and other outlets all trace the figures to one Financial Times account of an unaudited investor presentation. "Adjusted" profitability leaves out costs nobody has named, and the reporting never reconciles two different "annualized" figures: a weekly run rate reported above $100 billion in mid-September and quarterly revenue of $16.8 billion, which works out to about $67 billion a year.
An open-source-intelligence read of how likely this story is true with its real nuance, not a judgment of any outlet. It assesses the claim, weighing independent and adversarial reporting. How we label confidence.
What this means
A ratio of $1.70 in revenue per compute dollar implies gross margins far above those of a pure inference reseller. It is the first concrete evidence an investor can use to test whether frontier labs can make money on inference after a year of price cuts. The parties exposed are Anthropic's cloud and chip suppliers, whose order books depend on Anthropic's compute spending, and every private AI company whose valuation will be marked against the IPO price. The disclosure is based on company reporting, not an audit, and the S-1 filing will be the first primary document. Revenue growth is getting the most attention, but the slowing sequential rate and the gap between the $1.5 trillion and $2 trillion figures show that buyers are already pricing in the open-weight pressure documented elsewhere in this edition. The likelier outcome is a November listing priced below the $2 trillion figure that circulated in September. The price will be anchored to compute-adjusted margins rather than revenue multiples, and it will reset late-stage private AI valuations within months. That depends on the Federal Trade Commission investigation and the agent-safety incidents not producing a material disclosure in the prospectus. The read would be proven wrong if the S-1 shows a lower compute ratio than the one presented to investors, or if the listing slips into 2027.
What to watch
Observations to monitor, not financial advice.
Source: Polylog editors
Comments
0No comments yet.