Anthropic Files S-1 Prospectus for Nasdaq IPO, Eyeing $2 Trillion Valuation

Anthropic has officially filed its S-1 registration statement with the U.S. Securities and Exchange Commission (SEC) for an initial public offering on the Nasdaq, where it plans to trade under the ticker symbol "ANT." Led by Goldman Sachs and Morgan Stanley, the move makes Anthropic the first frontier artificial intelligence lab of its scale to go public.

Wall Street expectations place the company's potential valuation at roughly $2 trillion, according to Yahoo Finance. That figure marks a significant jump from Anthropic’s most recent private valuation of approximately $965 billion, which had already moved it ahead of rival OpenAI’s valuation of roughly $852 billion.

Explosive Revenue and First Operating Profit

The financial statements in the prospectus highlight an unprecedented acceleration in revenue. For the first half of 2026, Anthropic generated $16.3 billion—more than triple its full-year 2025 revenue of $4.6 billion. Revenue surged from $376 million in the first quarter of 2025 to $11.6 billion in the second quarter of 2026, representing a 1,368 percent year-over-year increase.

Anthropic also achieved its first positive GAAP operating profit in Q2 2026, reaching $974 million. While the company's compute costs grew 300 percent during the same timeframe, operational income expanded fast enough to outpace that rise.

With approximately 5,300 employees on staff, Anthropic's revenue per employee sits at $6.6 million. That metric surpasses every company in Wall Street's "Magnificent 7" cohort except Nvidia. News of the filing spurred a rally across major semiconductor stocks, including Nvidia, Broadcom, and AMD.

Customer Spending and Revenue Structure

A key driver of Anthropic’s top-line growth is customer expansion. The company reported a net revenue retention rate of approximately 900 percent, meaning existing customers are spending on average nine times what they did a year prior.

More than 11,900 customers spend upwards of $100,000 annually with Anthropic, and roughly 250 enterprise clients spend more than $10 million per year. Customers who combine Claude Enterprise with Claude Code represent the heaviest spenders, paying around 20 times more than other users.

Anthropic generates 80 percent of its revenue through usage-based, per-token billing, while the remaining 20 percent stems from fixed subscriptions. The company noted that its projections for the upcoming year will remain predominantly tied to token consumption.

Government contracts represent about 1 percent of overall revenue. Meanwhile, approximately 5 to 7 percent of revenue has historically come from accounts violating terms of service—primarily through model distillation—which the company has been actively cracking down on. Customer concentration remains a factor to watch: while no single entity accounts for more than 10 percent of revenue, its five largest customers together represent 25 percent.

Long-Term Obligations and "Anthropic Direct"

Alongside massive sales, Anthropic revealed heavy, long-term capital commitments required to power its models. The filing lists long-term compute capacity contracts with partners including Broadcom, Google, Amazon, and Microsoft totaling between $780 billion and $810 billion.

Because these massive infrastructure commitments are mostly fixed while incoming revenue relies primarily on variable token usage, the company requires sustained high-velocity growth to maintain margins. To reduce reliance on third-party capacity and manage costs over the long run, Anthropic disclosed it is building its own proprietary data centers under an initiative titled "Anthropic Direct."

The S-1 filing excluded numbers for the third quarter of 2026, pointing to a potential growth deceleration during that window. However, external metrics suggest demand has remained robust into late 2026: data from platform OpenRouter indicates that model usage rebounded significantly following the October rollout of the Claude 5.5 series, led by Opus 5.5.