Anthropic Files S-1 to List on Nasdaq as ‘ANT,’ Beating OpenAI to Public Markets

Anthropic Files S-1 to List on Nasdaq as ‘ANT,’ Beating OpenAI to Public Markets

Anthropic made its S-1 public with the US Securities and Exchange Commission on October 26, 2026, laying out plans to list on Nasdaq under the ticker 'ANT' and becoming the first frontier AI lab to place a prospectus in front of public investors. The filing, which follows a confidential submission earlier in the year, details the Claude product line, the company's dependence on third-party cloud capacity and the governance role of its Long-Term Benefit Trust, while leaving the offering size and

Anthropic publicly filed its S-1 registration statement with the US Securities and Exchange Commission on October 26, 2026, setting out plans to list on Nasdaq under the ticker 'ANT'. The first public version leaves the offering size, price range and final syndicate roster blank. The document follows a confidential submission made earlier under the same SEC process, and marks the first time outsiders can read the company's full financial and risk disclosures.

The prospectus opens with the company's stated mission — to 'revolutionize every area of human endeavor' — and lists its main product lines: the Claude model family, a developer-facing API and the Claude Code coding tool. The risk section centres on three issues: heavy reliance on third-party cloud compute, direct competition with OpenAI and Google, and uncertainty around model safety and regulation. The filing also explains how

Three pressures explain the timing. The first is capital structure: the $13 billion round in September 2025 pushed the post-money valuation to about $183 billion, leaving limited private capital willing to keep buying at a premium, while employee equity and early funds need a liquidity exit. The second is cost structure: multi-year compute contracts with Google Cloud and AWS are fixed obligations whose duration private money struggles to match, and only public markets can supply longer-dated capital. The third is competitive tempo: after OpenAI completed its public benefit corporation restructuring in 2025, the market broadly expects a listing in 2026 or 2027, and whichever company lists

Liu Gong's take: the real significance of Anthropic's S-1 is not that it beats OpenAI to the queue, but that frontier-model cost structure is being opened to public scrutiny for the first time. Within six months — by April 2027 — ANT will almost certainly be trading, and its first two earnings calls will be dominated by compute commitments and gross margin rather than benchmark leaderboards.

From a technical standpoint, competition has shifted from training to inference economics. Anthropic's revenue leans on enterprise APIs and coding tools, workloads that burn many tokens, run at high frequency and face steady price compression; consumer subscriptions, by contrast, can hold margin through brand premium. On the publicly reported numbers, the September 2025 round valued Anthropic at roughly $183 billion against about $7 billion of annualised revenue, a sales multiple near 26x; OpenAI's secondary-market trades imply about $500 billion against roughly $13 billion annualised, near 38x. Both are called frontier labs, yet the gap is about 30% — and that gap is precisely the price of billable enterprise revenue versus consumer narrative. The same rule holds in industrial quality inspection, where line customers pay not for the strongest model but for a fraction of a cent saved per thousand inferences and a version they can lock.

On the business side, revenue concentration cuts both ways. Claude Code is reported at a $1 billion-plus annualised run rate, and coding is the easiest capability to monetise today, but it is capped by the number of developers willing to pay; once GitHub Copilot, Cursor and Google compete at the same price point, the fight compresses both acquisition cost and margin. Enterprise API stickiness comes from migration cost, giving Anthropic more predictable net revenue retention than a consumer subscription — at the price of weaker bargaining power against a handful of large customers.

For the industry, this listing is a pricing anchor. CoreWeave priced at $40 in March 2025, below its range, showing public markets will only conditionally fund asset-heavy AI stories; Circle and Figma doubling on debut showed that asset-light, cash-generative names are still chased. Anthropic sits between the two: it builds no data centres but carries long-term compute contracts. If ANT holds its offer price, the window opens for OpenAI and xAI; if it breaks, the $183 billion private mark gets repriced immediately, and the transmission will be faster than any private secondary discount.

In the industrial vision inspection work Liu Gong follows, one variable is easily missed: a growing share of manufacturers want models deployed inside the plant, pushing inference from cloud to edge. That erodes the long-run slope of per-call public API billing and pushes Anthropic toward higher-priced private licences. The gap between the inference-cost line and the cloud-commitment line in the S-1 will say more about the company's real position than any safety statement.

  • Valuation multiple: on reported figures, Anthropic's September 2025 round valued it near $183 billion against roughly $7 billion annualised revenue, about 26x sales; OpenAI's implied $500 billion against roughly $13 billion annualised is near 38x — a discount of about 30% for Anthropic.
  • Enterprise share: a mid-2025 Menlo Ventures report put Anthropic at about 32% of enterprise LLM API spending, OpenAI near 25% and Google near 20%, the base for its revenue predictability.
  • Single-product scale: Claude Code is reported above a $1 billion annualised run rate, the fastest-growing line and one of the few businesses the S-1 can break out separately.
  • IPO benchmark: CoreWeave priced at $40

For the industry, the near-term (three months) effect runs through pricing and queueing: ANT's roadshow will reset the reference base for every AI listing through 2027, and banks will reprice AI risk premia within days. Over twelve months, if gross margin holds in an explainable band after listing, public markets will again pay up for compute-heavy companies; if not, the valuation pressure transmits straight back into private rounds and makes the next lab's fundraising harder.

For ordinary users, the near term (three months) is close to invisible: Claude's subscription tiers and free allowances will not move because of a listing, though reporting pressure may push more aggressive paid

On the competitive side, OpenAI has not publicly responded to Anthropic's filing, though the market read it as an accelerant for its own listing timetable. Google stayed silent; it is simultaneously one of Anthropic's largest cloud suppliers and a direct Gemini competitor, a dual role that makes its position in ANT's pricing awkward. Analyst opinion splits on two points. Bulls stress the recurring nature of enterprise API revenue and a roughly 32% enterprise share, arguing that 26x sales against OpenAI's 38x leaves room for a re-rating. Bears point at the multi-year compute contracts, arguing that fixed obligations will eat gross margin fast once revenue growth decelerates.

The developer

Questions worth exploring:

  • How public markets price inference economics: the link between compute commitments, gross margin and net revenue retention
  • How Anthropic's early listing reshapes OpenAI's IPO window and private-market valuations
  • How on-premise enterprise deployment erodes per-call public API billing over time

Further reading: 2026 Frontier LLM Comparison: A Practical Guide — a durable, evergreen reference. Bookmark it.

This item comes from The Verge AI (published 2026-10-26 13:32:49). Wenfeng Radar provides a Chinese/English summary and commentary on overseas AI developments; the original content is owned by its author.


Daily aggregation of overseas AI developments with in-depth analysis. Bookmark us to never miss an important signal; back to homepage.

Republishing: feel free to quote or repost with a link back to the original article on Wenfeng Radar. · Business & partnership: see our media kit
Daily briefing + Sunday review · Unsubscribe anytime by replying · Your email stays private

Leave a Reply

Your email address will not be published. Required fields are marked *

中文EN
Contact · Privacy · Disclaimer · AI Disclosure · Weekly Review · Media & Business
Email: wenfengleida@wenfengleida.cn