What was disclosed
On 10 November 2026, Anthropic filed Amendment No. 2 to its Form S-1 registration statement (File No. 333-299633) and an accompanying issuer free writing prospectus. The documents describe a concurrent private placement of Class N common stock: immediately after the IPO closes, and subject to certain conditions, seven companies will buy, “severally and not jointly,” an aggregate of $26.0 billion of Class N stock at a price per share equal to the IPO price. The sale is contingent on completion of the offering.
The filing lists the buyers and, at the assumed $74.00 midpoint of the $70.00 to $78.00 price range, the approximate amounts:
| Buyer | Approximate amount at $74.00 | Shares at $74.00 |
|---|---|---|
| Google LLC | $10.0 billion | 135,135,135 |
| NVIDIA | $10.0 billion | 135,135,135 |
| AMD | $2.5 billion | 33,783,783 |
| Broadcom | $1.0 billion | 13,513,513 |
| Micron Technology, Inc. | $1.0 billion | 13,513,513 |
| Microsoft Corporation | $1.0 billion | 13,513,513 |
| TSMC Partners, Ltd. | $500 million | 6,756,756 |
Because the purchase price tracks the final IPO price, share counts will change if the offering prices above or below the midpoint. The filing says these placement shares are not registered in the offering and will not be subject to market standoff agreements with Anthropic or lock-up agreements with the underwriters. Anthropic has applied to list on the Nasdaq Global Select Market under the symbol “ANT”; the company is offering 801,351,354 shares and selling stockholders 23,648,646.
How this relates to earlier arrangements
Reporting on the filing notes that the Google commitment follows a March 2026 arrangement under which Anthropic gave Google an option to buy up to $10 billion of Class N stock in a private placement concurrent with the offering at the IPO price; the amended filing now lists Google as a buyer. Nvidia’s purchase was first reported in September as a possible anchor investment while talks were under way, and CNBC reports that the amended prospectus says Nvidia has agreed to buy another $10 billion of Class N stock right after the offering. Coverage differs on how the commitments relate to earlier public pledges, so readers should rely on the filing text rather than on secondary summaries for those links.
Most of the buyers are also suppliers or partners across the AI hardware and cloud chain, which is why the structure is being read as a combination of investment and commercial relationship. That reading is commentary rather than something the filing states.
Why research offices might care
- Vendor concentration and dependencies. Institutions that license Claude through enterprise agreements, or build research workflows on it, now have a clearer picture of who holds large, non-voting stakes in the vendor. Procurement and research-security reviews can reasonably note the relationship between a model provider and its chip, cloud and memory suppliers when assessing continuity risk.
- Compute and capital for the research ecosystem. Frontier-model development is capital-intensive, and these commitments are one input to how much compute Anthropic can fund. Whether that changes academic access terms, credits or research programmes is not stated in the filing.
- Disclosure and conflict-of-interest practice. Researchers and offices working with AI vendors that also have investor-suppliers may want to confirm that existing conflict-of-interest and sponsored-research disclosure policies cover such indirect relationships.
- Timing is conditional. The IPO range is an estimate, trading has been reported as planned to begin on 19 November, and the placement closes only after the offering completes. Terms can still change before pricing.
For what the same filing says about Anthropic’s own assessment of AI-driven R&D acceleration, see CASRAI’s earlier report: Anthropic’s IPO filing now says it has met its own threshold for dramatically accelerated AI progress.
This post reports disclosed facts and is not investment advice.
