On 2 November Anthropic filed an amended registration statement (S-1/A) with the US Securities and Exchange Commission disclosing third-quarter 2026 revenue of about $17 billion and a net loss of $23.3 billion. Purchase and lease commitments, almost all for compute, stand at $895.4 billion. Passages of the amendment quoted by several outlets also say that AI-accelerated AI research may 'imminently' reach a threshold in the company's Responsible Scaling Policy.

Slower growth and the cost of compute

According to Fortune and Forbes, revenue rose about 47% from $11.6 billion in the second quarter and roughly twelvefold from $1.4 billion a year earlier. Operating expenses were $17.5 billion, turning a $974 million operating profit in the prior quarter into a $567 million operating loss. The filing explains that short-term compute 'currently costs significantly more' than multi-year capacity, and it booked about $2 billion of non-cash charges for an employee donation-matching programme. Most of the $23.3 billion net loss stems from fair-value accounting on convertible notes.

'Substantially all' of the $895.4 billion in commitments is compute and infrastructure, with about $448 billion falling due by the end of 2030 and about $70 billion in the twelve months to September 2027. Liquidity stands at $146.6 billion. Media reports describe a Nasdaq listing in late November at a valuation of $1.4–1.75 trillion, which could not be confirmed from the filing itself.

The safety-policy threshold and a Senate inquiry

According to passages quoted by several secondary reports, the amendment says internal estimates for future models suggest the company 'may imminently' meet the RSP's definition of 'dramatically accelerated progress attributable to AI', strengthening the original filing's 'could advance' to 'may have already advanced'. It reportedly also says the company is considering RSP changes that would require human review and approval of specific R&D decisions. The filing itself could not be opened for this edition, so the wording rests on press accounts.

The disclosure came a day before the 3 November deadline that Senator Elizabeth Warren set in her 27 October letter to CEO Dario Amodei. The letter asks how the company mitigates loss of control, autonomous cyberattacks and bioweapons misuse, and states: 'I am deeply troubled by the risks raised by Anthropic itself.' Securities disclosures by frontier AI firms are beginning to function as documents that show outsiders how safety policies are being operated.

Sources

Primary sources

  1. 1

    Anthropic, PBC — Form S-1/A

    Primary sourceU.S. Securities and Exchange Commission (EDGAR) · 2 Nov 2026

    U.S. Securities and Exchange Commission (EDGAR) · External site

Reporting

  1. 2

    Anthropic's growth slowed and costs jumped in the third quarter, updated IPO filing shows

    ReportingFortune · 2 Nov 2026

    Source checked 3 Nov 2026

    Fortune · External site
  2. 3

    Anthropic Hit $17 Billion In Third Quarter Revenue Ahead Of IPO—But AI Spending Commitments Hit $900 Billion

    ReportingForbes · 2 Nov 2026

    Source checked 3 Nov 2026

    Forbes · External site
  3. 4

    Anthropic Q3 Results: $17 Billion Revenue, $23 Billion Loss, $895 Billion in Commitments

    ReportingTrending Topics · 2 Nov 2026

    Source checked 3 Nov 2026

    Trending Topics · External site
  4. 5

    Anthropic Warns Future AI Models May Soon Trigger 'Dramatically Accelerated' Research Pace

    ReportingHuggingNews · 2 Nov 2026

    Source checked 3 Nov 2026

    HuggingNews · External site
  5. 6

    Sen. Warren probes AI safety, market risks in Anthropic IPO

    ReportingCFO Dive · 28 Oct 2026

    Source checked 3 Nov 2026

    CFO Dive · External site

Read this edition

From developments to evidence and analysis

Background

No explicit relationship is recorded.

Evidence

No explicit relationship is recorded.

Follow-ups / later versions

No explicit relationship is recorded.

Related questions

Cite this article

AI & Security Observatory(2026-11-02)「Anthropic、上場申請書の修正版で売上170億ドル・純損失233億ドルを開示 計算資源の購入契約は8,954億ドル」『AI & Security Observatory』、第1.0版、最終更新2026-11-03。https://ai-security-observatory.org/ja/brief/brief-20261103-07/

AI & Security Observatory (2026-11-02). Anthropic's amended IPO filing discloses Q3 revenue of $17bn and a $23.3bn net loss, with $895.4bn in compute purchase commitments . AI & Security Observatory, version 1.0. Updated 2026-11-03. https://ai-security-observatory.org/en/brief/brief-20261103-07/

Version 1.0 · Publication date 2 Nov 2026 · Updated 3 Nov 2026

Revision history
  1. Earlier revisions have not been recorded.