- Public Offering: AI company Anthropic could begin public stock trading in November, funding expansion while its CEO urges slower development.
- Computing Costs: Its commitment to Amazon Web Services exceeds $100 billion over ten years, covering model training and running its Claude assistant.
- Business Growth: Investors argue that companies adopting Claude could sustain sales growth even if advances in its capabilities slow.
- Revenue Forecast: Anthropic is reportedly expected to exceed a $100 billion annualized sales pace by year-end.
Anthropic, the company behind the Claude AI assistant, could begin public stock trading as soon as November. The offering could finance its costly expansion while chief executive Dario Amodei calls for slower advances in AI capabilities. The New York Times reported the possible debut on September 18, citing people familiar with the plans.
Public offering documents could appear in the coming weeks, the people said. Anthropic has already submitted a confidential draft registration statement to the U.S. Securities and Exchange Commission, an action confirmed on June 1. That submission gave it the option to pursue an initial public offering, or IPO, after regulatory review. The company had not set a share count or price at that point.
The Times reported that Amodei, chief financial officer Krishna Rao and the company’s bankers had met prospective investors, including existing backers and investors willing to commit large sums. An IPO would open ownership to public-market investors and could supply fresh capital for the computing infrastructure used to build and operate Claude.
Computing Costs Extend Beyond New Models
Anthropic’s commitment to Amazon Web Services exceeds $100 billion over ten years. Amazon’s cloud infrastructure will provide capacity both to train Claude models and to run them for customers. Serving more customers consumes computing resources even when the pace of new model development slows.
Investors interviewed by the Times see room for those two activities to grow at different speeds. Most of Anthropic’s revenue comes from business customers, and companies are still early in adopting Claude, they argued. Wider use within businesses could therefore expand sales even if Anthropic takes longer to improve the models’ most advanced capabilities.
The Financial Times reported that slower development could save billions in training expenses while allowing rivals to catch up. Snother model release could counter momentum from OpenAI, Anthropic’s main rival in business AI tools. One source said Anthropic was evaluating the model’s safety as part of those discussions; the company declined to comment to Reuters.
OpenAI has chosen a different timetable for public investors. Its chief executive, Sam Altman, said on September 12 that it would remain private in 2026, citing AI safety concerns.
Revenue Growth and Profit Measure Different Things
Anthropic was expected to exceed $100 billion in annualized revenue by the end of 2026, up from a $65 billion pace in July. Annualized revenue projects a sales pace across twelve months. Investors were using the growth to justify a potential $2 trillion company valuation, the Times reported. That valuation describes what the business might be worth, rather than how much cash an offering would raise.
The Financial Times reported second-quarter revenue of $11.5 billion, fourteen times the year-earlier figure, alongside positive adjusted operating income. Anthropic had told a small group of shareholders that the adjusted measure would remain positive for a second consecutive quarter, according to the report.
That measure excludes costs including stock-based compensation, the value of shares awarded to employees.
What Amodei Wants to Slow
Amodei’s proposal to pace AI development focuses on giving safeguards time to catch up with improving capabilities. He argues that AI systems are increasingly helping build their successors, accelerating development in a way that could exceed people’s ability to understand and control them.
His proposal calls for more time to make models behave safely and for outside evaluators to examine the work, while model training and technical progress continue. He argues that gaining even an extra year or two before models reach critical capability levels could improve safety research. Wider coordination among companies and governments, in his account, could provide that time without forcing one developer to surrender its commercial advantage alone.
Public Scrutiny and Laboratory Access
Altimeter Capital investor Brad Gerstner has argued that going public would bring more scrutiny, accountability and participation to leading AI companies. A public listing would bring financial disclosures and allow a wider group of investors to own shares.
Disagree. Anthropic will IPO. The market knows how to price risk – see SpaceX. There is huge appetite to invest in the AI leaders. And its beneficial / critical that we bring even more transparency, scrutiny, accountability, & participation to these grt American companies! 🇺🇸📈 https://t.co/RBVhUBNmx2
— Brad Gerstner (@altcap) September 12, 2026
Harvard law professor Jesse Fried told the Times that Anthropic could still keep information about activity inside its lab confidential after going public.
Anthropic has recently announced an evaluation partnership with Accenture, led by Accenture’s specialist AI business Faculty. The planned evaluators would work inside Anthropic with access comparable to employees, allowing them to observe model training, examine safety practices and report incidents. Anthropic will fund Accenture’s work directly and says it will continue training and releasing models alongside the evaluators. Standards governing what information evaluators can access and how they report their findings are still being worked out.
