The Anthropic IPO: Three Numbers, One of Which Is a Price
The Setup
Anthropic filed a confidential S-1 draft with the SEC on 1 June 2026 and is reported to be targeting an October listing on Nasdaq, with Goldman Sachs, JPMorgan and Morgan Stanley underwriting and a raise reported above $60B. Revenue run-rate was reported at roughly $4.7B as of May.
Anthropic has not confirmed the terms. Everything below the S-1 filing itself is sourced reporting, and a confidential draft is explicitly not a commitment to list — companies withdraw at this stage routinely.
The Concept
A private company does not have a valuation. It has several, produced by different mechanisms with different reliability:
A round price is what an investor paid on a date, for preferred stock that usually carries liquidation preference and other protections. It is a real transaction, and it is stale the moment it closes.
A secondary mark is where existing shares change hands between holders. Thinner, but continuous and closer to a market.
A target is what someone hopes the offering will price at. It is a forecast wearing the costume of a number.
The Read
Sort the three that way and they stop contradicting each other:
- $965B — the post-money on the Series H that closed in May. A real price, for preferred stock, four months old.
- $1.05–1.15T — where secondaries have been trading. The closest thing to a live market, and roughly 9–19% above the May round.
- $2T+ — what investors are reported to be targeting for the listing. Roughly 2.1x the last round and ~1.8x the secondary market.
That last gap is the thing worth staring at. A target roughly double the most recent transaction is not a valuation — it is a claim about demand that will be tested on one day in October. It may well clear; sentiment for scarce AI assets has repeatedly absorbed step-ups this size. But it should be read as the least reliable of the three, not the headline, and the widely-circulated pairing of "$965B valuation" with an October listing quietly mixes a closed May round with a future event.
The revenue figure carries the same caution in a different form. At a $4.7B run-rate, $2T is roughly 425x revenue; even the secondary range is above 220x. Those multiples are not absurd on their face for a company growing at this rate — they are simply doing all the work, which means the entire thesis rests on forward growth rather than on anything visible today.
The Action
Wait for the public S-1. The confidential draft tells you almost nothing; the public filing carries audited financials, the cap table, share classes, use of proceeds, and the risk factors the company is legally obliged to disclose. Every number above is reporting. That document is testimony.
Read the risk factors first, not last. For this company the ones that matter are concentration of compute suppliers and cloud partners, model-training input costs, litigation exposure, governance and share-class structure, and competitive pressure — including from Chinese labs. Those sections are where a filing is least promotional.
If you hold exposure indirectly, price it now rather than on listing day. Anyone holding the large public shareholders, or a fund with a secondary position, already owns a piece of this outcome. That is a position to size deliberately before a listing reprices it, not after.
What to Watch Next
- The public S-1 flip. Timing tells you whether October is real; a filing that does not surface by early September makes the window difficult.
- Whether the secondary market converges on the target. Secondaries moving toward $2T would corroborate the target. Staying near $1.1T means the gap is the underwriters' problem to close.
- Share-class structure. Whether public buyers get votes materially changes what is being sold.
- The comparison set. A listing at this size reprices every private AI mark by reference, whether or not those companies did anything.
The Counter
The strongest case against the skepticism above: private-round prices systematically understate fast-growing companies, because they are struck months earlier with investor protections attached. A $2T target may not be a stretch from $965B so much as an acknowledgement that the May round was already stale when it closed. Secondary markets are also thin and often trade at a liquidity discount, so $1.05–1.15T may understate rather than anchor.
And a revenue multiple is the wrong lens if growth is steep enough — a run-rate captured in May is itself a lagging figure for a company at this stage. Anchoring on 425x assumes today's revenue is the right denominator, which is precisely what the bull case denies.
The honest position is that nobody outside the company can currently distinguish those readings with public information. That is not a reason to pick one; it is a reason to wait for the filing.
Key Terms
- Confidential S-1 — a draft registration filed privately with the SEC. Permits SEC review before public disclosure; does not commit the company to listing.
- Post-money valuation — company value immediately after a round closes, including the new money. The headline number from a funding announcement.
- Secondary market — trading in existing private shares between holders, rather than new shares issued by the company.
- Liquidation preference — the right of preferred holders to be paid before common. It is why a preferred round price and a per-share common value are not the same thing.
Primary Sources
- SEC — Anthropic confidential S-1 draft, filed 1 June 2026 (the only primary document; terms not public)
- Reuters via Yahoo Finance — investors target $2 trillion IPO valuation in October
- TheStreet — prospectus and registration filing
- The Wall Street Journal — reporting on IPO timing, underwriters and the May round (subscription)