August 21, 2026 — In less than five years, Anthropic has gone from a splinter group of OpenAI researchers to a company that may soon surpass SpaceX to become the largest IPO in history. With annualized revenue exceeding $650 billion and a potential $2 trillion market valuation, the Claude developer is reshaping the AI landscape even before its stock begins trading [citation:1][citation:2][citation:3].

The company is expected to file its public IPO documents as early as the end of August, with trading potentially beginning in October [citation:4][citation:8]. The stakes could not be higher: Anthropic's initial public offering may match or exceed SpaceX's $75 billion primary offering — or $86.2 billion including overallotments — making it the largest IPO in U.S. history [citation:4][citation:12].

But behind the staggering numbers lies a fundamental question: is a company that only recently turned an operating profit really worth $2 trillion, or is the market pricing in a future that may never materialize?

The Numbers: 650 Billion in Annualized Revenue and 14x Growth

Anthropic's growth trajectory has been among the steepest in technology industry history. According to disclosures made to investors, the company's annualized revenue run rate reached $650 billion by the end of July 2026, up from $470 billion in May and just $90 billion at the end of 2025 — a more than sixfold increase in approximately seven months [citation:1][citation:3][citation:7].

For the second quarter of 2026, Anthropic reported approximately $115 billion in preliminary revenue, up from $7.87 billion in the same quarter of 2025 — growth of about 14x year-over-year [citation:2][citation:4][citation:14].

Perhaps more significantly, the company achieved its first quarterly adjusted operating profit in Q2 2026 [citation:2][citation:3][citation:14]. SemiAnalysis projects that Anthropic's GAAP earnings before interest and taxes (EBIT) will exceed $10 billion in the third quarter, representing a margin of approximately 6% on a run-rate basis [citation:2][citation:7].

The company's gross margins have also seen dramatic improvement. Inference gross margins rose from 38% a year ago to between 70% and 85%, driven by hardware architecture optimizations and custom silicon utilization [citation:2][citation:7]. Overall gross margin is now in the mid-60% range, while the API business alone achieves gross margins above 80% [citation:2][citation:7].

However, these impressive numbers come with a significant caveat: Anthropic reported a net loss of approximately $420 billion in 2025, up nearly fivefold from the previous year, reflecting the enormous cost of training frontier AI models [citation:4][citation:8].

Why Anthropic Outran OpenAI: The API-First Strategy

Anthropic's business model differs fundamentally from its chief rival, OpenAI. While OpenAI derives more than 65% of its revenue from consumer subscriptions, approximately 75% to 85% of Anthropic's annualized revenue comes from API-based usage fees paid by enterprise customers [citation:2][citation:7].

This distinction has proven critical. As enterprise customers deploy more agentic workflows and automate increasingly complex tasks, their token consumption — and their spending — grows without requiring Anthropic to continuously acquire new users [citation:2][citation:7].

The results of this strategy are visible in several key metrics:

• Enterprise adoption: As of June 2026, 34.4% of U.S. companies were paying for Anthropic's services, surpassing OpenAI's 32.3% share [citation:2][citation:7].
• Enterprise LLM market share: Anthropic held 32% of the enterprise large language model market in Q2 2026, compared with OpenAI's 25% [citation:2][citation:7].
• High-value customers: More than 1,000 companies now spend over $1 million annually on Claude's API, up from just 500 in February 2026 [citation:2].
• Fortune 10 penetration: Eight of the Fortune 10 companies are paying Anthropic customers [citation:2].
• Net revenue retention: 500% — meaning the same cohort of customers is spending five times more than they did a year ago [citation:2].

The Claude Code programming tool has been a major growth driver. Launched in February 2025, the tool's usage grew more than 10x in its first year and was generating over $25 billion in annualized revenue by early 2026. It now accounts for more than 7% of all GitHub code commits [citation:2].

IPO Timing: A Race Against SpaceX and OpenAI

Anthropic's IPO plans are accelerating. The company is working with Morgan Stanley, Goldman Sachs, and JPMorgan on the offering, and is expected to file publicly by the end of August [citation:4][citation:8][citation:12]. Trading could begin as early as October 2026, according to multiple reports [citation:1][citation:6][citation:14].

The timing is significant for several reasons. SpaceX's June 2026 IPO set a high bar, raising $75 billion initially ($86.2 billion with overallotments) at a valuation of approximately $1.77 trillion [citation:4][citation:12][citation:8]. Some investors are betting that Anthropic could surpass that record [citation:1][citation:6].

Prediction market Polymarket shows Anthropic's probability of becoming the largest IPO of 2026 at approximately 44%, only slightly behind SpaceX's 55% — a gap that has narrowed sharply in recent weeks [citation:1][citation:6].

OpenAI, which filed confidential IPO documents in June, appears to be moving more slowly. The company is reportedly considering a 2027 debut, leaving Anthropic a potential opening to establish itself as the first major AI pure-play to go public [citation:1][citation:6][citation:8].

The Broadcom Factor: AI Infrastructure as a Service

Behind Anthropic's rapid growth is a parallel story about how the company secures the computing power it needs. Broadcom, the chip designer, is currently negotiating a $60 billion to $100 billion AI chip financing facility that would directly benefit Anthropic and other AI companies [citation:5][citation:13].

The structure is notable: the financing would include $60 billion to $70 billion in senior secured debt, partially guaranteed by Broadcom, plus approximately $300 billion in subordinated debt, bringing the total to as much as $100 billion [citation:5][citation:13]. Blackstone and Apollo Global Management are in discussions to participate [citation:5][citation:13].

Under the proposed structure, investors like Blackstone and Apollo would purchase custom AI chips that would then be leased to Anthropic for infrastructure use [citation:13]. By providing credit support to the senior debt portion, Broadcom helps secure investment-grade ratings and lower borrowing costs [citation:13].

This financing approach reflects a broader trend in the AI industry: companies are building "equity + credit + long-term compute contracts" stacks to fund the enormous infrastructure required to train and run frontier models [citation:12]. Anthropic is also finalizing a revolving credit facility of more than $10 billion ahead of its IPO [citation:4][citation:8].

Valuation Debate: Is $2 Trillion Justifiable?

Even by the standards of the AI industry, a $2 trillion valuation for a company with no net profit and a $420 billion loss the previous year is extraordinary.

Investors defending the valuation point to growth. If Anthropic reaches the $1 trillion to $1.2 trillion in annualized revenue that investors project for the end of 2026, a $2 trillion market cap would represent a price-to-sales ratio of roughly 1.7 to 2, far lower than many high-growth SaaS companies [citation:1][citation:2][citation:7].

Critics counter that the projected revenue figures may be too optimistic and that Anthropic's losses could persist longer than expected. The company is still in the early stages of building the infrastructure it needs to remain competitive with OpenAI, Google, and Meta, and those capital requirements are unlikely to decline in the near term [citation:10].

There is also the question of timing. 2026 has been a strong year for IPOs, with new listings raising $160.6 billion as of mid-August [citation:12]. But market conditions can change quickly, and a valuation of $2 trillion would require enormous investor demand [citation:10][citation:12].

Key Takeaways