Anthropic is reportedly heading toward an October IPO, and investor models cited in the source article suggest the company could debut at a valuation above $2 trillion, a figure that would put it in unprecedented territory for an AI listing.
According to the article, that number did not come from Anthropic itself. The Financial Times reportedly said six Anthropic investors built their own valuation models and arrived at estimates above $2 trillion for a possible October listing. At least one of those models, using an 800% growth assumption, produced a figure above $3 trillion. The article adds that Anthropic executives have not settled on a target valuation even in private discussions.
Anthropic was founded by the Amodei siblings and former core OpenAI team members. The source describes a company that was working out of a small San Francisco office five years ago and is now being pushed by capital markets into the top tier of AI companies. It contrasts Anthropic’s five-year run from zero to a discussed $2 trillion valuation with Apple’s roughly 40-year path from IPO to a $2 trillion market capitalization.
Revenue growth is driving the valuation case
The article ties much of the valuation argument to Anthropic’s revenue trajectory. It says annual recurring revenue stood at about $9 billion at the end of 2025 and reached $47 billion by May this year.
It also says Anthropic generated more than $11.5 billion in second-quarter revenue, compared with $787 million in the same period a year earlier, a 14-fold increase.
Reuters, according to the article, reported another figure that had not previously been disclosed: Anthropic internally projects revenue of $190 billion to $200 billion in 2028. That leaves investors looking at a steep progression, from $9 billion at the end of 2025 to $47 billion now and as much as $200 billion two years later.
Faced with that growth curve, Wall Street banks and investors are said to be using enterprise-value-to-revenue multiples and anchoring valuation directly to the 2028 forecast. The article notes that valuing fast-growing software companies on revenue rather than profit is common. What stands out here is the decision to skip 2026 and 2027 and price the company off a financial outcome that still sits two years ahead.
API and enterprise demand form the core of the business
The report argues that one of the clearest differences between Anthropic and OpenAI lies in revenue mix. It says roughly 70% to 80% of Anthropic’s revenue comes from APIs and enterprise contracts billed by token usage, rather than from a subscription-heavy retail base. In that framing, enterprise customers and their day-to-day model calls are the main engine of growth.
Claude Code is described as the steepest part of that curve. The article says the product line had annualized revenue of $500 million in September 2025 and reached $2.5 billion by February 2026, a fivefold increase in five months. By the company’s internal measure, it now contributes close to one-fifth of total revenue.
The piece also cites usage data from open-source repositories and enterprise users. In public GitHub repositories, the share of commits written by Claude Code reportedly rose from 4% to more than 10%.
Uber’s CTO, the article says, had stated that the company exhausted its full-year AI budget in four months, with Claude Code and Cursor accounting for much of that spend. It also cites per-engineer monthly API usage of $500 to $2,000 and says Claude Code’s internal adoption rate rose from 32% to 84%.
Menlo Ventures’ latest enterprise survey, as cited in the article, shows Anthropic accounting for about 40% of enterprise large language model spending, compared with 27% for OpenAI. In coding, Anthropic’s share was put at about 54%, versus 21% for OpenAI. Those figures, in the article’s telling, help explain why some investors are willing to base such aggressive valuations on forward revenue estimates.
Culture and governance concerns are surfacing
As IPO expectations build, scrutiny of Anthropic’s internal culture is rising as well. The article cites early Anthropic investor and Silicon Valley commentator Brian Roemmele, who wrote: 「情况比我想象的还要糟糕。投资者和员工的士气，已经降到了历史最低点。」
The source traces those tensions to what it describes as a leadership group dominated by elite PhDs. It says CEO Dario Amodei and colleagues left OpenAI with a mission-driven image centered on protecting humanity, and that this internal narrative has since become a source of friction.
Citing Wired and other foreign media outlets, the article says Anthropic has a recurring internal session called “Dario vision exploration.” At all-hands meetings held every two weeks, Dario is described as focusing less on operating metrics and more on the ultimate fate of human civilization and the catastrophic risks of AI. One former employee, as quoted in the source, said: 「这感觉不像是在开公司大会，而像是在听牧师布道。」
The article says that style has deepened internal divisions. Some younger employees were initially drawn in by the company’s broad mission, while others became more critical as personal and financial pressures grew. Newer hires, the source says, are less receptive to that internal messaging.
It also reports that frustrated rank-and-file employees created a hidden internal network to vent their dissatisfaction with what the article calls the company’s upper “priestly” class. Near the IPO, employees are presented as being caught between two realities: stock options that could be worth immense sums and a workplace culture that they say is taking a toll on mental well-being.
Growth demands, compute costs and regulation all remain in focus
The final section of the article frames Anthropic’s challenge as a structural one. Building the safest AI, it says, requires the strongest large models. Training the strongest models requires massive computing power. Buying that compute requires capital on a very large scale.
Once capital enters, the article argues, the pressure for growth and outsized returns becomes impossible to avoid. It points to SpaceX, which it says set a record in June and then fell 38% from its high over the following two months, using that move as a warning about how quickly lofty valuations can be tested.
Anthropic, in the article’s telling, is now walking a narrow line. Government regulation, rising compute costs and mounting governance tensions could all be amplified after a public listing if the company does move ahead in October.
Even so, the piece closes by arguing that Anthropic’s coding strength has already put it deep into the core race around ASI. With October approaching, markets are waiting to see whether the IPO will reset valuation benchmarks for the AI sector once again.