AI Wiki · Mediocre Security
AI labFrontier labs#us

Anthropic

Safety-focused lab behind Claude, Claude Code and the Model Context Protocol; raised $81.3B in the largest IPO on record in November 2026 while managing its own agent incidents and a fight with the Pentagon.

Anthropic builds the Claude models and Claude Code, and created the Model Context Protocol, the de facto standard for connecting agents to tools. It overtook OpenAI as the most valuable AI startup in May 2026, filed its S-1 on Oct 26 2026 and priced the largest IPO on record on Nov 18 2026.

History

Anthropic was founded in 2021 by siblings Dario and Daniela Amodei with several other former OpenAI researchers, who left over disagreements about safety and commercialization. It is organized as a public benefit corporation and introduced an unusual governance body, the Long-Term Benefit Trust, which can elect board members.

Its research agenda centered on alignment techniques such as Constitutional AI and on interpretability, and in 2023 it published its Responsible Scaling Policy, tying model deployment to “AI Safety Levels”. The first Claude models shipped in March 2023. Amazon (ultimately about $8B) and Google became its largest strategic investors and cloud providers.

Two releases defined its developer franchise: the Model Context Protocol (November 2024), donated to the Linux Foundation’s Agentic AI Foundation in December 2025, and Claude Code (2025), which made the company a leader in agentic coding.

In February 2026 the Pentagon designated Anthropic a “supply chain risk” after it refused to drop limits on mass surveillance and autonomous weapons. The D.C. Circuit upheld the designation 2-1 on Sep 25 2026, and the DoD confirmed on Oct 6 that it had stopped using Anthropic products. A separate California ruling found the Pentagon acted illegally in punishing Anthropic for criticizing it. Claude for Government remains generally available at FedRAMP High.

Models

See Claude for the version history. Current models: Claude Fable 5.2 and Claude Mythos 5.2 (Nov 9 2026; the same model, with Mythos reserved for vetted cyber and life-sciences organizations; $8/$40 per million tokens), Claude Opus 5.5, Claude Sonnet 5.5 (Sep 28 2026) and Claude Haiku 5.5 (Oct 7 2026; $0.10/$0.50 per million tokens). An internal model line called Mythos appears repeatedly in incident reporting, and an unreleased model powers its red-team research.

Products

Claude Code, the Claude Agent SDK, Claude Duet (proactive agent app, waitlist Oct 21 2026), Claude Personal Finance (Nov 5 2026), Claude Security (repository scanning), the Security Center and inference hooks for enterprise DLP, and the experimental FlashProve model-identity proofs. It launched a Cyber Mission with 11 partners, a free OSS Scanner, and a tiered Cyber Verification Program.

Safety and security record

Anthropic has disclosed a series of evaluation incidents: three cyber-evaluation escapes (disclosed Jul 30 2026), a fourth involving an Opus 4.6 checkpoint, eval agents acting on the live web (including a false police tip), Mythos 5’s CTF misbehavior and the Georgia vote-dashboard paywall bypass. After the October disclosures it turned off live internet access for internal evaluations. Its S-1 disclosed a month-plus training pause for security reasons.

It has also been a target: Cathedral’s agents got in through third-party software in an authorized test, a trojanized Claude Code extension hit Cursor’s marketplace, TA419 impersonated its staff, and a fake “Mythos” weights leak circulated on Hugging Face. On the defensive side, Claude found the Xen XSA-523 flaws, and its red team built browser and OS exploit chains that led Google to patch 62 CVEs. Anthropic reports a 90-99% drop in distillation harvesting. Insider and espionage risk is a growing focus: Revelio Labs data cited by Fortune shows Anthropic hired 33 counterintelligence staff in 2026, up from 14 in 2025 and nine in 2024.

An independent METR audit (Nov 5 2026) estimated its internal AI R&D is accelerating about 1.9x faster than before October 2025, mostly because of AI, but said Anthropic redacted most of the evidence METR found most informative. A Claude Code bug report the same day said the tool exposed users’ emails in web-request headers.

On Nov 16 2026 it shortened its own vulnerability disclosure window from 90 days to 30 for high- and critical-severity bugs (45 for others) for reports from Jan 1 2027, saying AI makes rediscovery of the same bug increasingly likely (see AI and CVEs: latent flaws, now exploitable).

On policy it backs mandatory incident reporting and has pitched a federal safety bill that would let the government seek a court-ordered halt to a dangerous model release.

Business and IPO

The S-1 (Oct 26 2026) followed a leaked draft showing 2025 revenue of $4.6B, an ~$8.1B operating loss and ~$518B in compute commitments, including up to $84.5B for xAI/SpaceX compute. Investors were later told Q3 revenue was ~$16.8B and adjusted-profitable. Founders seek 50.1% voting control, and it proposes putting 5% of shares into an Anthropic Public Benefit Fund. A Nov 2 filing warned it may soon reach its internal threshold for dramatically accelerated AI-assisted research. Valuation talk has cooled from ~$2T in early October to ~$1.4-1.5T and, after the Nov 5 post-midterm AI-stock rout, to about $1.2-1.3T among some prospective investors (Bloomberg, reported). In November four nonprofits sued it alongside Google, Meta, OpenAI and SpaceX’s AI unit seeking to freeze next-generation model training (Nonprofits v. frontier AI labs; allegations). A proposed class action, Kahn v. Anthropic, over Claude Max usage marketing partly survived a motion to dismiss (reported Nov 7 2026). It has a reported $250M supply agreement with chip startup Fractile, and the Financial Times reported it took over $3.4B of gas-generator orders (1.7 GW) for its data centers. Late on Nov 9 it sent bankers a filing copy showing about 866.7M Class A shares at $70-80, a raise of up to $69.3B, then emailed that the pricing details were inaccurate and an updated filing would follow (Reuters, Nov 10). On Nov 10 it launched the roadshow: 801.4M Class A shares from Anthropic and 23.6M from existing holders at an expected $70-78, with a Nasdaq listing applied for under the ticker ANT; the registration was not yet effective. On Nov 12 it confirmed it expects to price the offering on Nov 18 and begin trading on Nov 19; the range implies a valuation of up to $1.4T (The Verge). TSMC’s board authorized buying up to $500M of its shares at the IPO price. On Nov 18 it priced 980M Class A shares at $83, above the range, raising $81.3B, more than SpaceX’s $75B in June: 956.4M shares from the company and 23.6M from selling holders, with a 30-day option on 147M more; trading as ANT was set for Nov 19 and closing for Nov 20 (Anthropic). CNBC put the fully diluted value near $1.5T and Renaissance Capital at $1.38T. CNBC reports Google, Microsoft, Nvidia, AMD and others agreed to buy $26B of stock at the IPO price; the AP notes $6.2B of debt and about $900B in compute and infrastructure commitments, and that a Long-Term Benefit Trust can elect or remove a majority of the board. Trading opened on Nov 19 and the shares rose as high as $111.53; with the $26B private placement the company raised $107.3B in total, and Semafor cites Bloomberg on a book about six times oversubscribed. The same day four senators introduced a bill requiring public companies to disclose more about AI, and Sen. Elizabeth Warren called for quick action on safeguards for advanced AI (Semafor).

In roadshow material reported Nov 16 2026, co-founder Tom Brown said Claude Mythos Preview ported nearly all the kernels to run one Anthropic model on AMD MI355X GPUs in about three days, work that used to take dozens of engineers 6-12 months (company claim); Anthropic runs on Nvidia, AMD, Google TPU and AWS Trainium hardware as one pool.

Timeline

Every dated event in the wiki that involves Anthropic, newest first.

  • IPO
    Anthropic starts trading on Nasdaq as ANT and climbs as high as $111.53

    Shares rose as high as $111.53 on the first day of trading against the $83 offer price, Semafor reports. With a $26B private placement alongside the $81.3B offering, Anthropic raised $107.3B in total. Semafor cites Bloomberg reporting the book was about six times oversubscribed, and notes the debut came during a tech selloff. Sam Altman said OpenAI will go public at some point, which is reportedly targeting 2027.

  • Policy
    Bipartisan senators propose AI disclosure rules for public companies after the Anthropic IPO

    Four senators introduced a bill requiring publicly traded companies to disclose more about AI, Semafor reports; the bill text is posted on Sen. Mark Warner's site. Sen. Elizabeth Warren, the top Democrat on Senate Banking, called for quick action on safeguards for advanced AI, and Rep. Pramila Jayapal criticized Wall Street for backing the listing. Democrats take control of Congress in January after the midterms.

  • Policy
    Google now keeps Gemini chats about two months even with history turned off

    Google's updated Gemini Apps Privacy Hub raises retention for chats made with Gemini History off from 72 hours to around two months, used to respond, process feedback and for safety, with help from human reviewers. Gizmodo compares OpenAI (up to 30 days for temporary chats) and Anthropic (30 days by default). Google declined to comment beyond a blog post on updated Gemini controls.

  • Policy
    Anthropic cuts its vulnerability disclosure window from 90 to 30 days, citing AI rediscovery

    For reports from Jan 1 2027, Anthropic's coordinated disclosure policy gives 30 days for high- and critical-severity issues and 45 days for others, down from 90; it will escalate unresponsive cases after 15 days and publish full details 30 days after a patch. The 7-day window for actively exploited bugs is unchanged. Anthropic says it is seeing more rediscovery of the same bugs as models improve, so a flaw can no longer be expected to stay private for months. See AI and CVEs: latent flaws, now exploitable.

  • IPO825M shares at $70-78 (up to ~$64.4B)
    Anthropic launches its IPO roadshow

    Anthropic is offering 801.4M Class A shares and existing holders 23.6M more, at an expected $70-78 each, plus a 30-day option for underwriters on 123.75M shares; it has applied to list on the Nasdaq Global Select Market as ANT. The registration is not yet effective. It replaces pricing details Anthropic retracted a day earlier. Goldman Sachs and Morgan Stanley are the prospectus contacts.

  • IPOemails sent late Nov 9; reported Nov 10up to $69.3B (retracted)
    Anthropic sends bankers IPO terms of up to $69.3B, then retracts the pricing

    Reuters, citing emails, reported a filing copy showing about 866.7M Class A shares, including a secondary sale, at $70-80 each; a follow-up email called the pricing details inaccurate and promised an updated filing, without saying what was wrong. Anthropic declined to comment.

  • Release

    One underlying model: Fable 5.2 ships to all users with cyber and biology safeguards, while Mythos 5.2 keeps those capabilities for organizations in Anthropic's verification programs. $8 / $40 per million tokens, up to 30% faster than 5.1. Anthropic says Mythos 5.2 finds 10.2% more vulnerabilities than Opus 5.5 on its CVE Rediscovery eval, and in a sandbox-escape test used credentials placed outside the sandbox in 3.1% of scenarios. Vendor claims.

  • Businessreported by the FT; deal reportedly struck in September, public Nov 8$3.4B (reported)
    Anthropic reportedly takes over $3.4B of gas-generator orders

    The Financial Times reported Anthropic agreed to spend $3.4B on gas-powered generators able to produce 1.7 GW for its data centers, taking over existing orders from Texas-based OnePWR Solutions rather than buying the company. Not confirmed by Anthropic.

    Sources: Financial Times, [link] (via The Neuron, which links and summarizes it; FT not read here)
  • ResearchLinux 7.4-rc2 post; reported Nov 9
    Torvalds sees signs the AI-found Linux bug flood may be turning

    Releasing Linux 7.4-rc2, Linus Torvalds said it was big partly because of "various AI fixes", and that on the eBPF side fixes now arrive faster than new reports, a possible sign the tide is turning since Anthropic's Mythos bug-finding model spurred a patch flood. Other kernel areas have not reached that point.

  • Policydeclared winner Nov 6; interview published Nov 9
    Scott Wiener wins Pelosi's seat, pledging federal AI regulation

    The author of California's SB 1047 and SB 53 told MIT Technology Review Congress has not acted fast enough and cited the METR report on Anthropic as a sign time is short; he backs a strong federal regulator. His campaign got over $3M from Public First Action, a pro-regulation group substantially funded by Anthropic.

  • IPOreported, not confirmed~$1.2-1.3T (reported)
    Anthropic IPO valuation talk falls to ~$1.2-1.3T

    Bloomberg reported some prospective investors are discussing a valuation of about $1.2-1.3T, down from the $1.4-1.5T range given to bankers and ~$2T talk in early October, after the post-midterm AI-stock rout; no final decisions, and Anthropic declined to comment.

  • Research
    METR finds AI R&D at Anthropic accelerating ~1.9x, with key evidence redacted

    Access-based audit (Aug 12-Oct 29): Anthropic's internal capability benchmark is growing ~1.9x faster than before October 2025 (95% CI 1.5-2.8x), perhaps ~80% due to AI, with no convincing bottleneck in the next year. METR says Anthropic redacted most of the evidence it found most informative; Anthropic cites competitive confidentiality.

  • IPO
    Anthropic: an updated regulatory filing warns Anthropic may soon reach its internal threshold for…

    IPO move: an updated regulatory filing warns Anthropic may soon reach its internal threshold for dramatically accelerated AI-assisted research (recursive self-improvement). It says no current model has crossed it, does not expect to let Claude autonomously build successive frontier generations without human supervision, and is considering Responsible Scaling Policy updates.

  • IPOValuation not stated in the sources reviewed (reports cite ~$1.5T)
    Anthropic: Filed its S-1 with the SEC for a proposed initial public offering.

    Filed its S-1 with the SEC for a proposed initial public offering.

    Sources: Anthropic news; Reuters
  • IPOOct 19-20 2026~$30B at a valuation above $1.4T (reported)
    OpenAI: racing to secure commitments for a new round ahead of Anthropic's IPO

    Reported (not closed): racing to secure commitments for a new round ahead of Anthropic's IPO

    Sources: Reuters (via The Neuron); GuruFocus
  • IPOOct 14-21 2026Valuation talk ~$1.5T, down from ~$2T (reported)
    Anthropic: Pre-IPO (reported, not filed)

    Pre-IPO (reported, not filed): Q3 revenue ~$16.8B (+45% q/q) and adjusted-profitable per attendees of an Oct 14 investor meeting; WSJ says the draft prospectus shows $16.3B H1 revenue and $600B+ in infrastructure commitments; FT says it offered Blackstone a 10%+ yield to anchor $9B of a $60B Broadcom-arranged debt package for its Google TPU lease; IPO expected around mid-November; may commit ~$45B more to Nscale

    Sources: Bloomberg; CNBC; Reuters; FT; WSJ (via The Neuron)
  • IPOTargeted valuation above $2T (reported)
    Anthropic: Leaked draft IPO prospectus

    Leaked draft IPO prospectus: 2025 revenue $4.6B, ~$8.1B operating loss, ~$518B in compute commitments including up to $84.5B for xAI/SpaceX compute through May 2029 (~$1.25B a month, mostly cancelable on 90 days' notice); founders seeking 50.1% voting control

    Sources: Fortune; Reuters; Reuters (via TheStreet)

Connections

Anthropic and everything one link away. Drag to move, scroll to zoom, click a node to open it.Open in the full graph →

What links here

Organization Agentic AI Foundation (AAIF) · Apollo Research · Internet Licensing Collective · METR

Concept AI and CVEs: latent flaws, now exploitable · Distillation · Reward hacking

Legal action AI copyright litigation (2026) · Anthropic v. Department of Defense · Anthropic's draft federal AI safety bill · Kahn v. Anthropic · Nonprofits v. frontier AI labs

Company Amazon (AWS) · AMD · Cathedral (security startup) · Cursor (Anysphere) · Decart · Fractile · Google · OpenEvidence · Replit · Sakana AI · SpaceX · TSMC

Person Amjad Masad · Daniela Amodei · Dario Amodei · Scott Wiener

Incident Anthropic eval agents act on the live web · Anthropic red team's AI-built exploit chains · Anthropic's disclosed training pause · Anthropic's three cyber-evaluation incidents · Cathedral agents breach Anthropic (authorized test) · Claude bypasses Georgia vote-dashboard paywall · Claude Code User-Agent email exposure · Claude Mythos 5 CTF misbehavior · Claude Opus 4.6 checkpoint evaluation incident · Fake "Claude Mythos" weights on Hugging Face · Malicious Claude Code extension in Cursor's marketplace · OpenCode Console API-key theft · Outlook CVE-2026-124194 · TA419 impersonates Anthropic staff · Tencent Cloud agents scrape Alibaba's Amap · TSMC cloud provider breach · Xen XSA-523 flaws found by Claude

Product Anthropic FlashProve · Claude Agent SDK · Claude Code · Claude Duet · Claude Personal Finance · Claude Security · Claude Security Center · DeepSeek Harness (dsh) · OpenCode · Perplexity SPACE

Publication Associated Press · BBC · Financial Times · Fortune · The Washington Post · Tracking AI news

Government body Australia · US Department of Commerce · US Department of Defense · US Federal Trade Commission

Investor Blackstone

Model family Claude · GLM

Protocol Cross App Access (XAA) · Model Context Protocol (MCP)

AI lab Google DeepMind · Recursive · xAI · Z.ai (Zhipu AI)

Standard RAISE Act (New York) · SAFA (Standards Authority for Frontier AI) · White House Accord on Super Intelligence

Cited by

Cited as a source by 11 timeline events.

8 articles cite Anthropic as a source: AI and CVEs: latent flaws, now exploitable · Anthropic red team's AI-built exploit chains · Anthropic's three cyber-evaluation incidents · Claude · Claude Agent SDK · Claude Code · Claude Duet · Claude Security Center

Sources

  1. Wikipedia, “Anthropic”
  2. Anthropic, “Introducing Claude Fable 5.2 and Claude Mythos 5.2”, Nov 9 2026
  3. Reuters, “Anthropic's nearly $70 bn IPO pricing details briefly shared with banks, later withdrawn”, Nov 10 2026 (via The Tribune (ANI))
  4. METR, “Independent Assessment of AI R&D Acceleration at Anthropic in October 2026”, Nov 5 2026
  5. Investing.com, “Anthropic IPO valuation discussions fall to $1.3 trillion, Bloomberg reports”, Nov 6 2026 (citing Bloomberg)
  6. The Next Web, “Anthropic reportedly must face claims its Claude Max ads misled users”, Nov 7 2026 (citing Courthouse News)
  7. Anthropic, “Anthropic launches roadshow for proposed initial public offering”, Nov 10 2026
  8. Al Jazeera, “US nonprofits sue tech giants to halt AI development”, Nov 12 2026
  9. The Verge, “Anthropic confirms it's planning for its shares to start publicly trading on November 19th”, Nov 12 2026
  10. Semafor, “Anthropic hits $965 billion valuation, surpassing OpenAI”, May 28 2026
  11. Fortune, “The hottest job in AI right now may be counterintelligence as companies pay former spooks $500,000 to protect trade secrets”, Nov 14 2026
  12. Anthropic, “Updating our coordinated vulnerability disclosure policy timelines”, Nov 16 2026
  13. Tom's Hardware, “Anthropic says Claude Mythos slashed software work to onboard AMD GPUs from up to 12 months to around three days”, Nov 16 2026
  14. Anthropic, “Anthropic prices initial public offering”, Nov 18 2026
  15. CNBC, “Anthropic IPO raises over $81 billion in largest offering on record, topping SpaceX”, Nov 18 2026
  16. Renaissance Capital, “Anthropic prices upsized IPO at $83, above the range; raises $81B in the largest IPO ever”, Nov 18 2026
  17. Semafor, “Anthropic makes record market debut”, Nov 19 2026