Anthropic Expects Second Profitable Quarter Ahead of IPO
From April through June of 2026, Anthropic posted about $559 million in operating profit on $10.9 billion in revenue. People familiar with the matter told the Financial Times that the company had informed a small group of shareholders it expects positive adjusted operating income for a second straight quarter, the three months ending September 30, 2026. Dario Amodei is the chief executive. Adjusted operating income, as framed for those investors, excludes expenses such as stock-based compensation.
Founded in San Francisco in 2021 as a public benefit corporation by former OpenAI members including Amodei, and known for its Claude models, the firm had spent its early years carrying the cost of building and training large AI systems. On the measure shared with that small group, it had crossed into profit. A second consecutive quarter would mean the result held through the end of September.
What sat underneath the claim was a revenue curve steep enough to change the operating arithmetic. Second-quarter revenue rose roughly fourteen-fold year over year, to more than $11.5 billion from $787 million in the same three months of 2025. The quarter before had already shown the break: first-quarter revenue reached $4.73 billion. From one quarter to the next the top line more than doubled.
By May the annualized run rate—the pace of sales if sustained for a full year—stood at $47 billion. By the end of July it had climbed to about $65 billion, up from $9 billion at the close of 2025. The increase was more than sevenfold.
Gross margins exceeded 80 percent before revenue-sharing with distribution partners such as Amazon and before the cost of training the models. The figure measured what remained after serving systems already built. The largest bills for the next generation of models sat outside it.
Amodei set the tension in a short, flat line. “We must slow the pace at which we improve the capabilities of AI models. Progress will still seem fast, and we must make wise use of the time we gain.” The sentence anchored an essay published on September 12 and 13, 2026, under the title “We Must Pace the Frontier,” a public call on the artificial intelligence industry to ease the rate at which new model capabilities arrived. “Pacing does not mean halting model training or technical progress,” he wrote. Training could continue. Technical work could continue. What he asked the industry to reduce was the speed of the capability climb itself, and to treat any time recovered as something to spend carefully. Sam Altman and Elon Musk publicly echoed support for the slowdown call. Jim Cramer questioned Amodei’s slowdown commentary on CNBC.
The essay sat beside the competitive pressure already building among the labs that trained the largest systems. Slowing the frontier was one argument. Keeping pace with the firms that kept building was another. Joy Brookhart, a researcher at the semiconductor and AI firm SemiAnalysis, put the problem in one sentence: “If this level of profitability and growth continues, competing with Anthropic will be very difficult due to their substantial computing resources.”
The warning turned on compute. In May 2026 a Series H funding round had valued Anthropic at $965 billion. Capital on that scale could keep buying the servers and chips a frontier lab needed; profit that held could lock the advantage in place.
Sam Altman did not answer with a matching listing plan. In a Fortune interview published September 12, 2026, he said the timing was not right for an OpenAI IPO and confirmed there would be no public offering until the next year. OpenAI would stay private past the end of 2026. Anthropic’s own path toward a public market was already being assembled.
Business Insider reported that Anthropic had chosen Nasdaq for its potential initial public offering, the first sale of shares on a public exchange. The report arrived hours before the Financial Times profitability story. On June 1, 2026, the company had filed confidentially with the Securities and Exchange Commission, the step that starts a private firm toward that sale. A potential IPO could value Anthropic at $2 trillion or more. The company is seeking up to $100 billion in the offering. The $2 trillion valuation is roughly 43 times its annualized second-quarter revenue.
By mid-September the Nasdaq choice, the capital target, and that multiple were all in the open reporting around the deal. The confidential filing was more than three months old. The exchange was set. The process was reported to begin in mid-October 2026, taking the offering out of confidential preparation and into the public sequence that ends with shares listed and trading. The aim attached to the start was fixed: complete the listing before November 3, 2026, the date of the U.S. midterm elections. Finishing ahead of that day meant the full run—from the mid-October opening through pricing and the first trades—had to close inside the short stretch of weeks that remained.
