Anthropic plans to raise roughly $90 billion in an initial public offering that could value the six-year-old AI company at $1.5 to $1.7 trillion — and pre-market trading already implies it may go higher.

Lead bankers Goldman Sachs and Morgan Stanley intend to price the transaction on Nov. 18, with public trading to begin on Nov. 19, according to a live deal calendar. The offering combines a conventional IPO with a private share placement to several of America's most successful AI companies, including Nvidia, Google, and Microsoft.

The early signals are bullish. On the Hyperliquid trading platform, Anthropic is changing hands at a suggested value of $1.7 to $1.84 trillion — a 20% to 37% premium over the proposed IPO range of $70 to $78 per share. Trading volume on the platform exceeds $311 million. Because of that significant pre-market interest, the lead bankers will probably raise the IPO price.

How did a six-year-old startup get here?

Only in the United States, the story goes, can a company founded barely six years ago become central to national efforts to maintain AI superiority over China — and only the U.S. has the deep pools of risk capital that let a new firm reach a valuation near $1.5 trillion in a few years.

Anthropic's origin was a schism. Toward the end of 2020, several senior employees at OpenAI resigned over what they saw as its failure to address ethics and safety issues, and founded what is today Anthropic. Now the company stands on the verge of becoming publicly traded.

Its growth engine is the Claude chatbot, followed by Claude Code, an AI system built for business users. Both achieved very rapid success. Anthropic now also sells AI services to the legal, financial, and cybersecurity industries.

What do the numbers show?

The revenue curve is parabolic. Anthropic's annualized revenue run rate (the last quarter's revenue multiplied by four, an annualization shortcut) now stands at about $90 billion, versus just $9 billion at the end of 2025 — a tenfold jump in under a year.

Third-quarter 2026 revenue reached $17 billion, a 47% quarter-on-quarter increase. That top line is heavily anchored by high-value enterprise accounts rather than scattered consumer subscriptions.

The company is also extraordinarily productive on a per-head basis. Trailing 12-month revenue per employee has scaled from $1.4 million a year ago to $6.6 million earlier this calendar quarter — more per employee than every Magnificent Seven company except Nvidia.

What could go wrong?

The risks are real, and they start with capital commitments. Before the end of 2030, Anthropic has long-term purchase and lease commitments for semiconductors and compute capacity totaling about $450 billion.

The open question: can the company grow revenue, become profitable, meet those enormous capital requirements, and remain a leader in developing frontier AI models — all at the same time?

Competition compounds the challenge. Anthropic faces formidable rivals in Google, Microsoft, and Meta. In the near future it will also compete against open-source AI models (software whose code is publicly available) developed by Nvidia and numerous Chinese companies.

Is the IPO a buy?

Right now, the deal appears poised to be a major success, with shares trading well above the indicated offering price. Where the stock trades over the coming months is another matter entirely.

The cautionary tale is SpaceX, the most recent trillion-dollar IPO, whose post-listing trading history shows how quickly sentiment can shift. As always with a new company, caveat emptor remains prudent advice.

Disclosure: the original columnist owns shares in Nvidia, Google and SpaceX. James Rogan, author of the source commentary, is a former U.S. diplomat who later worked in law and finance for over 30 years.