Anthropic Files S-1 for Nasdaq Listing Under Ticker ANT

On Monday, October 26, 2026, Anthropic publicly filed its Form S-1 registration statement with the U.S. Securities and Exchange Commission, setting the stage for one of the most closely watched technology listings in a generation. The artificial intelligence company applied to list Class A common stock on the Nasdaq Global Select Market under the ticker symbol ANT, with Goldman Sachs and Morgan Stanley leading the underwriting syndicate. The filing arrived after weeks of speculation about timing, valuation, and how a frontier lab would explain hundred-billion-dollar compute commitments to public investors.

Unlike software IPOs where gross margins and seat counts tell a familiar story, Anthropic's prospectus is dominated by infrastructure leases, revenue retention at enterprise scale, and explicit discussion of catastrophic risk from advanced AI systems. That combination makes the document less a quarterly earnings preview and more a referendum on whether capital markets will fund safety-conscious scaling the way they funded cloud and mobile a decade ago. Investors who opened the PDF on Sunday night found a company that had already outgrown the narrative of a research boutique.

What the numbers say

Anthropic reported $4.585 billion in revenue for calendar year 2025, up from $386 million in 2024, according to the registration statement. In the first half of 2026 alone, revenue reached $16.286 billion versus $1.163 billion in the year-ago period, a pace that would have been unimaginable for an enterprise software vendor at the same age. Subscription revenue in the first half of 2026 was $3.2 billion compared with $180 million a year earlier, while consumption-based revenue—often tied to API usage—was $13.06 billion versus $983 million.

The company also disclosed a first-half 2026 net loss of $80.748 billion, a figure that requires careful reading: roughly $72.388 billion came from fair-value changes on convertible notes, derivatives, and simple agreements for future equity rather than from operating cash burn alone. Still, the operating picture is not that of a mature profit machine. Anthropic cautioned that revenue growth may slow, costs will rise, and it may not achieve sustained profitability; it does not expect to pay dividends in the foreseeable future.

As of September 30, 2026, Anthropic counted more than 11,900 customers contributing over $100,000 in trailing-twelve-month revenue, more than 1,900 above $1 million, and about 250 above $10 million. Net revenue retention was about 900%, a metric more common in elite vertical SaaS than in research labs. Revenue per employee over the trailing twelve months exceeded $6.6 million, compared with about $1.4 million a year earlier, signaling that the company's go-to-market engine had caught up with its model quality.

Governance built for a controlled company

Anthropic's corporate structure is unusual even by Silicon Valley standards. After listing, a Founder LLC will hold one Class F share carrying 50.1% of the vote on general shareholder matters but no economic interest, allowing founders to retain control while selling economic exposure to the public. A Long-Term Benefit Trust comprising five trustees with no financial interest will select a majority of board seats through Class T stock.

The company injected roughly 847 million Class N shares—about 5% of fully diluted equity—into an Anthropic public benefit fund. All seven co-founders pledged 80% of personal equity to charity, and donation-matching expenses tied to that commitment were $541 million for the year ended December 31, 2025. These provisions signal that Anthropic wants investors to treat safety and public benefit as first-class governance constraints, not marketing footnotes.

Securities lawyers noted that the dual-class and trust arrangements may limit activist pressure during periods when safety investments depress near-term margins. Retail investors considering ANT will need to decide whether they are buying exposure to API growth or underwriting a governance experiment that explicitly subordinates short-term returns to long-term risk reduction.

The compute balance sheet

The S-1's risk factors read like a treaty between a cloud provider and a nation-state. Under arrangements with SpaceX AI, Anthropic agreed to spend up to $120.5 billion through 2029. A Broadcom lease implies total rent payable over the lease term, excluding interest, of about $161.2 billion. Expected spending on compute and infrastructure agreements could reach $125 billion to $150 billion.

Separately, Anthropic will spend up to roughly $114.5 billion with Google through 2033. Compute and infrastructure expenses are expected to remain the largest operating line item for the foreseeable future. For investors, the question is whether API pricing power and enterprise contracts can outrun depreciation schedules and lease escalators—a bet the market is now being asked to price in real time.

Analysts at several banks circulated models on Tuesday morning showing that even aggressive API price declines would leave Anthropic with positive gross margins if utilization of reserved clusters stays above 70%. The filing itself does not guarantee that utilization, which is why the roadshow will focus as much on customer pipelines as on benchmark scores.

IPO timing and market context

People familiar with the process told the Wall Street Journal that Anthropic is targeting the week of November 16 for the listing, though dates can shift as the SEC review continues. The filing did not yet disclose share count or price range; those details typically arrive in an amended S-1 after investor feedback. Reports suggest the offering could raise $100 billion or more at a valuation potentially in the $1.5 trillion to $2 trillion range, figures that would re-anchor how public markets value frontier AI capacity.

Monday's filing coincided with a risk-on session for AI-linked equities. Nasdaq futures rose, Nvidia gained more than 3% in early trading, and Bitcoin briefly dipped below $80,000 as investors rotated between crypto beta and AI infrastructure names. Anthropic's debut will therefore land in a market already debating whether AI capex is a bubble or the new baseline for corporate investment.

By Tuesday, October 27, secondary chatter focused on how ANT would trade relative to mega-cap cloud names and whether index funds could absorb a float large enough to matter for Nasdaq composition rules. Index providers typically need weeks of notice, so a mid-November pricing would still allow year-end inclusion debates to play out in public.

What competitors and partners reveal

Anthropic stated that no single customer accounts for more than 10% of revenue and that 79% of revenue comes from customers outside the top five, addressing concentration risk that often scares public investors. Google purchased 38.6 million G-2 non-voting redeemable shares, while an April agreement commits Amazon to provide up to $20 billion in financing. Those relationships show that hyperscalers are simultaneously competitors, customers, and creditors to the same labs they hope to keep inside their ecosystems.

OpenAI's decision to remain private for now does not insulate it from ANT's pricing discovery. If Anthropic's opening valuation embeds a premium for safety governance, OpenAI's eventual listing—or lack thereof—will be judged against that benchmark. Enterprise procurement officers, meanwhile, are reading the customer concentration statistics as evidence that multi-vendor AI strategies are already the norm rather than the exception.

Implications for the AI industry

A successful Anthropic listing would validate the API-first, safety-branded model at a scale OpenAI has not yet tested in public markets. It would also force every other frontier lab to prepare investor narratives around alignment spending, eval infrastructure, and liability for model misuse. Regulators reading the risk factors will see explicit acknowledgment that advanced AI could pose catastrophic or existential risks if not appropriately governed—language that may shape future disclosure norms for the sector.

For developers and enterprises, the IPO is less about the opening print than about continuity: will long-term compute contracts survive quarterly earnings pressure? Will safety budgets expand or compress when Class A shareholders demand margin improvement? The S-1 does not answer those questions, but it makes them impossible to ignore.

Anthropic has raised $141.1 billion in capital as of June 30, according to the filing. Listing on Nasdaq as ANT would convert a private balance sheet into a public experiment in funding frontier research without pretending the risks are ordinary software risks. The next two weeks of roadshow revisions will determine whether investors agree—and whether the ticker becomes a daily barometer for how much safety premium the market is willing to pay.