Seven former OpenAI employees — including siblings Dario Amodei (former VP of Research) and Daniela Amodei — leave to found Anthropic, citing concerns about OpenAI’s direction on safety and its pivot toward commercialization under Microsoft’s influence.

Cumulative: $124MLed by Jaan Tallinn (Skype co-founder) and Dustin Moskovitz (Facebook co-founder). Valued at ~$550M. Framed as funding for AI safety research. Also includes Eric Schmidt (former Google CEO), James McClave (Jane Street, crypto-adjacent trading firm).

Cumulative: ~$1BGoogle takes a 10% stake in Anthropic for $300M. Cloud infrastructure deal included — Anthropic begins building on Google Cloud alongside AWS.

Cumulative: ~$700MLed by Sam Bankman-Fried / FTX, which contributed $500M of the $580M total. Caroline Ellison also involved. Valued at $4B. Jane Street’s James McClave participates again. Summer 2022: Anthropic trains the first version of Claude but doesn’t release it publicly, citing safety.

First public release of Claude. Marketed heavily on “Constitutional AI” and safety-first principles. Positions Anthropic as the ethical alternative to OpenAI/ChatGPT.

Cumulative: ~$1.5BLed by Spark Capital. Investors include Google, Salesforce Ventures, Zoom, Menlo Ventures, Microsoft, Sound Ventures, SK Telecom, HOF Capital. Google invests an additional $500M (Oct 2023) and commits to $1.5B more over time.

Cumulative: ~$2.75BAmazon takes minority stake. Initial $1.25B with commitment to $4B total. Anthropic designates AWS as primary cloud provider. Claude available to AWS customers via Bedrock. This is the deal that makes Anthropic structurally dependent on Amazon infrastructure.

Cumulative: ~$7.6BLed by Menlo Ventures. Includes participation from DAMAC Group (Dubai) and investors linked to Saudi Arabia (MIS). Also Qualcomm (San Diego). Valuation reaches ~$18.1B. First entry of Gulf capital.

Cumulative: ~$10BRemaining $2.75B from Amazon’s pledged $4B commitment delivered. Anthropic now deeply embedded in AWS ecosystem. Claude models trained on AWS custom Trainium chips.

Bankrupt FTX sells its ~$800M in Anthropic shares to a consortium including ATIC Third International Investment Company (Abu Dhabi sovereign-linked) and Jane Street Global Trading. The tainted crypto money is laundered into sovereign and trading firm capital.

Key hires: Jan Leike, John Schulman, Durk Kingma — all from OpenAI. Reinforces the “we’re the safety company” narrative at a moment when OpenAI is visibly abandoning safety commitments.

Anthropic partners with Palantir Technologies and AWS to provide Claude models to U.S. intelligence and defense agencies. Claude deployed in Palantir’s Impact Level 6 (IL6) classified environment — data classified up to SECRET level. Palantir CTO calls it “asymmetric AI advantage” for “critical missions.” Anthropic’s head of sales calls it “responsible application of AI.”

Cumulative: ~$14BAmazon invests another $4B, doubling its total to $8B. Single largest investor in Anthropic. Also in November: Microsoft and Nvidia announce partnership deal with Anthropic.

Cumulative: ~$18BLed by Lightspeed Venture Partners ($1B). Participants: Bessemer, Cisco, D1 Capital, Fidelity, General Catalyst, Jane Street, Salesforce Ventures. Google adds another $1B. Revenue crosses $1B annual run-rate.

Claude Opus 4 and Claude Sonnet 4 released. Model Context Protocol (MCP) connector introduced. Claude Code goes from research preview to general availability with IDE integrations. Product-market fit accelerates among developers.

Anthropic announces Claude Gov, a government-specific product. In use at multiple U.S. agencies. Positions Anthropic as a government AI vendor alongside its “responsible AI” branding.

Anthropic signs a $200M contract with the Department of Defense. Claude becomes the first frontier AI model approved for use on classified networks. Pentagon agrees to abide by Anthropic’s Acceptable Use Policy — including restrictions on autonomous weapons and mass surveillance.

Cumulative: ~$34BCo-led by ICONIQ Capital, Fidelity, Lightspeed. Significant investment from the Qatar Investment Authority (sovereign wealth fund). Gulf capital now a major presence. Revenue reportedly ~$10B annual. Anthropic also blocks sales to entities majority-owned by Chinese, Russian, Iranian, or North Korean entities.

Anthropic secures access to up to 1 million Google TPUs. Partnership projected to bring 1+ gigawatt of AI compute online by 2026. Google’s total investment in Anthropic now exceeds $2B.

Microsoft commits up to $5B and Nvidia up to $10B investment in Anthropic. Claude models now available across AWS (Bedrock), Google Cloud (Vertex AI), and Microsoft Azure (Foundry) — all three major clouds. Every hyperscaler is now an Anthropic stakeholder.

Defense Secretary Pete Hegseth issues AI strategy memorandum directing all DoD AI contracts adopt standard “any lawful use” language. This directly targets Anthropic’s contractual restrictions on autonomous weapons and mass surveillance. Negotiations between Pentagon and Anthropic begin to break down.

Newest frontier model. Framed as advancing “agentic workflows” and “professional knowledge work.” Simultaneously: Blackstone invests $200M in Anthropic, building on earlier participation — total stake reaches $1B. Blackstone is now one of Anthropic’s largest non-venture, non-sovereign backers.

Cumulative: ~$64BSecond-largest private tech funding round in history. Led by GIC (Singapore sovereign wealth fund) and Coatue. Co-led by D.E. Shaw Ventures, Dragoneer, Founders Fund (Peter Thiel), ICONIQ, MGX (Abu Dhabi). Includes Microsoft, Nvidia, Sequoia Capital. Run-rate revenue: $14B. 8 of Fortune 10 companies now use Claude.

Second model in the 4.6 family. Both models target complex agentic workflows — “the kind of tasks that benefit from computer control.”

After Anthropic refuses to remove its two red lines (no mass surveillance of Americans, no fully autonomous weapons), Hegseth designates Anthropic a supply chain risk — a label previously reserved for foreign adversaries. Trump posts on Truth Social: “WE will decide the fate of our Country — NOT some out-of-control, Radical Left AI company.” GSA removes Anthropic from USAi.gov. Six-month phase-out period begins — but Claude continues to be used in Iran military operations.

Sam Altman announces OpenAI’s deal with the DoD within hours of Anthropic’s blacklisting. xAI also cleared for classified use. The competitive dynamic becomes nakedly visible: Anthropic’s principled stand creates a market opening its rivals immediately fill.

Anthropic reports 1M+ daily new users. Claude surpasses ChatGPT in Apple App Store rankings for the first time. The Pentagon fight becomes a brand accelerant. Public sympathy as business strategy.

Sues in Northern District of California and D.C. Circuit Court. Alleges First Amendment and due process violations. Argues designation exceeds statutory authority. Dozens of researchers from OpenAI and Google DeepMind file amicus brief in personal capacity supporting Anthropic. WSJ reports Claude was used in U.S. strikes during conflict with Iran and in the raid that led to Maduro’s arrest.

The Information reports Anthropic is negotiating a joint venture with Blackstone ($1.15T AUM) and Hellman & Friedman to embed Claude across their portfolio companies. Palantir-style model: AI technology + consulting/implementation services. Thousands of companies across real estate, healthcare, energy, manufacturing. Talks briefly paused due to Pentagon dispute, then resumed.

Claude can now control your Mac — clicking, scrolling, navigating applications autonomously. Available for Pro and Max subscribers as “research preview.” Paired with Dispatch (mobile task assignment). The “Her” endgame accelerates: Claude as operating system layer between user and digital life.