According to people familiar with the matter, Anthropic is in talks to have Nvidia (NVDA) participate as a cornerstone investor in what could become the largest initial public offering in history. The people said Anthropic is seeking to raise as much as $100 billion, with a valuation that could reach about $2 trillion, and Nvidia is considering investing as much as $10 billion. The plans are still under discussion and may change. Bringing in a cornerstone investor like Nvidia could boost market confidence in the listing.
Founded only five years ago, Anthropic is rapidly moving from a private unicorn toward the public market at an astonishing pace. As of the end of July 2026, its annualized revenue run rate had surpassed $65 billion, more than a sevenfold increase from about $9 billion at the end of 2025. Preliminary revenue for the second quarter of 2026 exceeded $11.5 billion, compared with only $787 million in the same period of 2025. Even more encouraging to the market, the second quarter achieved a positive adjusted operating profit of about $559 million.
However, high growth comes with enormous costs. Anthropic’s net loss in 2025 was close to $42 billion, five times the roughly $8.3 billion loss in the previous year. Unlike OpenAI, where more than 65% of revenue comes from consumer subscriptions, about 75% to 85% of Anthropic’s revenue comes from enterprise API calls, with the flagship product Claude Code serving as the core variable for growth. A significant improvement in inference efficiency has further strengthened the virtuous cycle of its business model, with the blended gross margin rising to the mid-60s percentage range, and the inference gross margin climbing from 38% a year earlier to 70% to 85%.
What supports the $2 trillion valuation is not merely the current $65 billion ARR. According to reports, Anthropic internally expects revenue of about $190 billion to $200 billion in 2028. If calculated based on $100 billion in revenue by the end of 2026, the $2 trillion valuation corresponds to a forward price-to-sales ratio of about 20 times, lower than the price-to-sales ratios of more than 50 times often seen among popular AI stocks such as Palantir and Nebius.
According to reports, Anthropic has delayed its IPO timetable. The prospectus originally expected to be made public in early September has now been moved to late September, the roadshow may begin as early as mid-October, and the listing could be completed just a few days before the U.S. midterm elections in November. The fundamental reason for the delay is that the standards of scrutiny in the public market differ from those in the private market. After listing, investors will have to face audited gross margins, computing expenses, and customer concentration.
At this critical moment in its IPO push, Anthropic plans to grant shares with extra voting rights to Chief Executive Officer Dario Amodei and other co-founders to withstand pressure from outside shareholders. The company also plans to retain its existing non-shareholder trustee structure, using special shares to elect a majority of board members. This governance structure will still face strict scrutiny, as institutional investors are increasingly skeptical of dual-class share structures.
Anthropic’s listing will become a historic moment in the commercialization of the AI industry. From an “escapee” from OpenAI to the world’s most highly valued AI unicorn, its five-year journey is itself a microcosm of the commercialization of the AI industry. Whether the $2 trillion valuation is a bubble or a value reassessment is an answer that may only truly be revealed after the company’s public financial reports are tested by the market. When Anthropic officially rings the listing bell on Wall Street, what it brings is not only the capitalization of one company, but also a comprehensive test of the entire AI industry’s valuation paradigm, commercialization path, and governance model.
