The supply question behind Anthropic’s IPO

The supply question behind Anthropic’s IPO

How a US mega-IPO turns private ownership into public-market supply—and what Korean-market investors should compare.

2026-11-04

Liquidity & Value

For an investor in a private company, an IPO changes who can buy the shares. It does not give every existing shareholder liquidity on the same day.

Anthropic’s proposed Nasdaq listing makes that distinction unusually visible. Its initial S-1, filed on 26 October, was followed by an amendment on 2 November. The intended ticker is ANT. The amended filing still leaves the offer price, share count and split between new shares and selling-shareholder stock blank. A public filing is a step towards an offering, not confirmation that it has been completed. [1] [2]

For investors who follow Korean IPOs, this is a useful case study in two questions: how the listing process works, and how much stock can reach the market after trading begins.

Rapid growth does not settle the supply question

Anthropic reported $33.2bn of revenue for the first nine months of 2026, including $17.0bn in the September quarter. That quarter also produced a $567m operating loss. Revenue growth and the terms on which new investors can own that growth are separate parts of the analysis. [2]


Anthropic quarterly revenue in 2026: Q1 $4.731bn, Q2 $11.555bn and Q3 $16.956bn.

Unaudited quarterly revenue, rounded to two decimals in USD billions. These are reported quarters, not annualised forecasts. Source: S-1/A, Quarterly Consolidated Statements of Operations, p. 161. [2]

The draft prospectus does not yet establish the size of the freely tradable float. A reported fundraising target divided by a reported valuation is an illustrative ratio, not a verified float calculation. Existing-holder sales, share conversions, restricted holdings and the final capitalisation table all matter.

How the US process differs from Korea’s

Both markets combine disclosure requirements, exchange admission standards and investor demand. In the US, SEC review concerns the registration statement; Nasdaq separately assesses its listing requirements. SEC effectiveness is not an endorsement of the company’s value. In Korea, the exchange’s preliminary listing review and the securities-registration process are separate stages. [3] [4] [5]


Simplified US and Korean underwritten IPO processes, including disclosure and exchange review, bookbuilding, pricing, allocation and trading.

Illustrative process comparison, as of 4 November 2026. Workstreams can overlap; timing and requirements vary by issuer, market and transaction. Sources: SEC, Nasdaq and FSC. [3] [4] [5]

The practical differences are most visible in allocation and resale. US investors may obtain an IPO allocation through participating brokers, with access depending on the deal and intermediary. Korea has a distinct public-subscription process alongside institutional bookbuilding. A strong subscription ratio measures demand for an allocation; it does not establish how long those investors will hold their shares. [3] [5]

Lock-ups also need to be read in context. US underwriting agreements commonly restrict insider sales, but the terms vary. In Korea, mandatory holding periods for specified existing shareholders must be distinguished from the holding commitments made by institutions receiving IPO allocations. Neither market can be reduced to a single universal six-month rule. [6] [7]

Korea’s framework is changing: the FSC’s implementation plan scheduled preliminary demand assessment and cornerstone-investor arrangements for 13 November 2026. This note is dated before that planned start; the diagram describes the conventional sequence. [8]

Read the holder groups before reading the calendar

Anthropic’s filing sets out 23 release rows, tied to the prospectus date and subsequent earnings announcements. Those are conditional triggers, not 23 confirmed calendar dates. The four principal holder groups do not follow the same schedule. [2]


Four Anthropic holder groups have distinct release triggers. Non-founder insiders may release 10% initially and a further 35% after FY2026 results; other groups have staged or longer restrictions.

Each percentage applies to the relevant group’s own locked securities. Group sizes and aggregate shares released are not supplied in the blank schedule. Non-trading-day adjustments, other resale restrictions and potential waivers must also be considered. [2]

For the specified “other 180-day” cohort, the scheduled early tranches add to 86% of that group’s holdings. That is not 86% of Anthropic. Nor should it be plotted as a fixed amount of selling before day 180: the earnings dates and earlier final-release provisions determine the actual sequence. The underwriters may also waive restrictions in circumstances described in the filing.

The investment question is therefore the number of shares becoming eligible relative to the tradable float—and how much holders actually choose to sell. A percentage without its denominator cannot answer it.

Unlocks and new issuance are different sources of supply


Three sources of market supply: new IPO shares, existing shareholder sales after restrictions end, and conditional future share issuance.

LASQUAD illustration. This separates transaction mechanisms; it does not forecast selling volumes or prices. Anthropic-specific arrangements are described in its S-1/A. [2]

Anthropic also has conditional financing arrangements with strategic partners. Existing Amazon notes, future funding capacity and a concurrent private-placement option should not be added together as if they were all newly issued, immediately tradable shares. Google’s optional IPO placement reduces the remaining capacity under its agreement. Broadcom-linked notes may convert at a discount to a trailing market-price average, subject to their terms. [2]

There is a further distinction between share classes. Converting an already outstanding Class N share into one Class A share changes its route to public trading; it does not itself create a second economic share. New shares issued on a financing draw or option exercise raise a different dilution question.

The relevance for Korean investors

A listed frontier-AI company would give global investors another way to express an AI view. It could compete for portfolio allocations while also supporting demand for computing infrastructure. Those channels can coexist. The IPO alone would not establish that selling in Korean semiconductor shares was caused by capital moving into Anthropic.

Our focus is the link between a company’s capital needs and a shareholder’s exit. For a Korean pre-IPO investment, the equivalent work is to map the primary-versus-secondary offering mix, ownership rights, mandatory holding periods, institutional commitments and subsequent financing. The first trading day is one event in that process.

What we will watch next

  • Final offering terms: price, new shares, existing-holder sales and the resulting float.

  • Holder-group sizes: the figures needed to translate release percentages into share counts.

  • Prospectus and earnings dates: the inputs needed for an actual calendar, with any waivers tracked separately.

  • Financing activity: which facilities are drawn, whether notes convert and when the resulting shares can be resold.

The useful comparison is how clearly investors can trace ownership, dilution and the path to liquidity.

Research as of 4 November 2026. The offering is proposed and its terms may change. General information only; not an offer to buy or sell securities or a personal investment recommendation.

Sources

  1. Anthropic — proposed IPO announcement, 26 October 2026

  2. Anthropic S-1/A, 2 November 2026 — quarterly accounts; shares eligible for future sale; underwriting; financing agreements

  3. SEC — Investor Bulletin on IPOs

  4. Nasdaq — Listing Qualifications review

  5. FSC — IPO pricing and institutional participation reforms, 23 April 2026

  6. SEC — IPO lock-up agreements

  7. KRX — KOSDAQ listing review guide, 2025

  8. FSC — proposed implementation framework for cornerstone investors, 30 July 2026

Sources

Primary filings and official references are linked in the article above. Research as of 4 November 2026.

Updated

2026-11-04

Private company research. Human conversations.

Profiles are based on public information. Inclusion does not indicate an available investment opportunity or an offer to buy or sell securities.