According to the UK's Financial Times, Anthropic, a leading global AI company, could go public as early as October this year, with some investors projecting a valuation of $2 trillion or even higher. This figure would surpass SpaceX's $1.77 trillion IPO record set in June 2026, becoming the largest initial public offering in human business history. Just three months ago, Anthropic's valuation after its Series H funding round was "only" $965 billion.

Founded in 2021 by former core members of OpenAI, the company behind Claude has reached a $2 trillion valuation in just over five years. Elon Musk recently admitted that his previous assessment of Anthropic was "clearly wrong," calling it "the obvious leader in AI today."

01. Revenue Growth Soars Vertically

Supporting this astonishing valuation is an equally remarkable set of figures. According to documents seen by Bloomberg that Anthropic disclosed to potential investors, the company's preliminary revenue exceeded $11.5 billion in Q2 2026, up nearly 14-fold from $787 million in the same period in 2025. In Q1 2026, the figure was $4.73 billion. More critically, Anthropic achieved positive adjusted operating profit for the first time in Q2. Research firm SemiAnalysis forecasts that the company's Q3 GAAP EBIT will exceed $1 billion.

The trajectory of annualized revenue (run-rate) is even steeper: approximately $9 billion at the end of 2025, reaching $14 billion in February 2026, surpassing $30 billion in April, and exceeding $47 billion in May. Investors expect this figure to reach $100-120 billion by the end of 2026. Anthropic's CFO disclosed in May that the company's net revenue retention rate has reached 500%—customers contributing $30 billion in ARR this year contributed only $2 billion a year earlier.

Anthropic's growth is not driven by burning cash for scale. Its business model fundamentally differs from OpenAI's. While OpenAI aims to become the gateway to the AI era, focusing on consumer markets, Anthropic has targeted enterprise clients from the outset. About 75-85% of its annualized revenue comes from usage-based API services, with consumer subscriptions accounting for only about 5%. In contrast, over 65% of OpenAI's revenue still comes from subscriptions.

The core logic of this model is simple: the more intelligent agent workflows enterprise clients deploy, the higher the token consumption and revenue, enabling expansion without constantly acquiring new users. The API model has no upper limit on revenue per user. As of April 2026, the number of enterprise clients spending over $1 million annually on the Claude platform exceeded 1,000, up from just over 500 in February of the same year. Eight of the Fortune 10 companies are already Claude clients. According to Ramp data, as of April this year, 34.4% of U.S. enterprises paid for Anthropic, surpassing OpenAI's 32.3%.

At the product level, the programming tool Claude Code is the biggest growth engine. Launched in February 2025, its usage has surged over 10-fold, with annualized revenue exceeding $2.5 billion by early 2026. According to SemiAnalysis, Claude Code now accounts for over 7% of all code commits on GitHub. In the enterprise-grade large model API market, Anthropic's share reached 32% in Q2 2026, surpassing OpenAI's 25% to become the industry leader.

02. The Logic and Controversy Behind the $2 Trillion Valuation

Another noteworthy figure is the rapid improvement in profitability. Thanks to hardware architecture optimization and deep utilization of custom chips, Anthropic's inference gross margin has soared from 38% a year ago to 70-85%. SemiAnalysis estimates that the company's overall gross margin has recovered from -94% in 2024 to the mid-60% range, with API business gross margins exceeding 80%. Measured by ARR per megawatt of computing power, this metric will reach $60 million for Anthropic later in 2026, up from just $16 million nine months earlier. With inference computing costs largely fixed, any increase in tokens processed per unit of computing power pushes marginal profitability toward 100%.

Of course, the $2 trillion valuation is not without controversy. Fortune magazine has raised sharp questions: Anthropic has yet to achieve any net profit; based on common valuation multiples for large companies in the Nasdaq 100 index, Anthropic would need to generate $59-79 billion in annual profit to justify a $2 trillion valuation. At a Q2 operating margin of approximately 5.1%, $120 billion in annualized revenue would generate about $6.2 billion in annualized operating profit, making the $2 trillion valuation about 325 times this figure.

However, investors may not be relying on traditional P/E models. Some analysts point out that with Anthropic, investors are not buying a software company but a call option on its AI strategic dominance. One investor bluntly stated: "If Anthropic maintains 800% annual growth, even at the lowest 30x price-to-sales ratio, it would be worth $3 trillion." Moreover, no directly comparable large model company exists in the U.S. public market.

03. Risks and Variables

Risks cannot be ignored. In June 2026, the U.S. Department of Commerce imposed export controls on Anthropic's two top models, Fable 5 and Mythos 5, forcing the company to suspend these models globally for nearly three weeks and dragging down monthly revenue growth. The Pentagon had previously labeled Anthropic a supply chain risk source. Regulatory pressure, low-cost competition from Chinese AI companies, and pricing pressure—Anthropic's flagship model costs two-and-a-half times more than OpenAI's equivalent—are all looming threats.

Meanwhile, Anthropic's deep ties with tech giants are both an advantage and a hidden danger (potential risk). Amazon has invested over $13 billion cumulatively, with a total commitment of up to $33 billion; in exchange, Anthropic has pledged to spend over $100 billion on AWS over the next decade. Google has invested about $3 billion since 2022, with the book value of this stake exceeding $124 billion by October 2025. Microsoft and NVIDIA have also joined with capital investments. This "interdependent" relationship provides Anthropic with nearly inexhaustible computing power but means it must continuously fulfill massive cloud service procurement commitments.

04. A Footnote to an Era

If Anthropic's IPO proceeds, its significance will extend far beyond a single company going public. It marks the AI industry's shift from reckless competition for computing power to a commercialization phase focused on "return on investment." For the past two years, the industry assumed large model R&D would endlessly burn cash, but Anthropic is proving with quarterly profits nearing $1 billion that top-tier model companies can generate exceptional cash flow. As SpaceX, Anthropic, and OpenAI—collectively worth trillions of dollars—converge on public markets, Wall Street is undergoing an unprecedented stress test.

Is $2 trillion a reasonable valuation for the future, or the peak of a bubble? The answer may not arrive until October.

- End -