Anthropic

Anthropic is an artificial-intelligence company founded in May 2021 by seven former OpenAI researchers, led by the siblings Dario Amodei (chief executive) and Daniela Amodei (president), who left OpenAI over disagreements about its direction and its balance between commercialization and safety. The founders came out of the Effective Altruism movement and built the company explicitly around the thesis that advanced AI should be developed with extreme caution, the same existential-risk concern that runs through the Rationalist Community and the institutions Peter Thiel funded through the Machine Intelligence Research Institute. Its 2022 funding round was led by a 500 million dollar investment from Sam Bankman-Fried months before his FTX exchange collapsed in fraud.¹²

Founding and Governance

The Amodeis and their cofounders, including Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish, and Chris Olah, incorporated Anthropic as a public-benefit corporation and built an unusual governance structure, the Long-Term Benefit Trust, intended to give a body insulated from shareholders the power to appoint directors and to weight safety over profit. The arrangement mirrored the nonprofit-control structure that produced OpenAI's 2023 board crisis. Daniela Amodei is married to Holden Karnofsky, cofounder of the effective-altruism organizations GiveWell and Open Philanthropy, which was an early Anthropic funder; Karnofsky later joined the company.²³

The Long-Term Benefit Trust was formally established in September 2023, holding a special class of stock (Class T shares) that gives it authority to elect and remove Anthropic directors on time- and funding-based milestones, phasing up to a majority of the board within four years. Its five founding trustees were Neil Buddy Shah of the Clinton Health Access Initiative as chair, Jason Matheny of the RAND Corporation, Kanika Bahl of Evidence Action, the alignment researcher Paul Christiano of the Alignment Research Center, and Zach Robinson of Effective Ventures US, several of them figures from the effective-altruism and AI-safety world. Matheny stepped down in December 2023 to avoid conflicts with RAND's policy work, and Christiano left in April 2024 to lead AI safety at the U.S. AI Safety Institute, with the trust adding Richard Fontaine, chief executive of the Center for a New American Security and executive director of the Trilateral Commission, in May 2025, and Mariano-Florentino Cuéllar, president of the Carnegie Endowment for International Peace, in January 2026. Bahl and Robinson left in January 2026, and the former Federal Reserve chairman Ben Bernanke joined in July 2026. With the trust's appointment of Novartis's chief executive, Vas Narasimhan, in April 2026, trust-appointed directors became a majority of the board, which also included Yasmin Razavi, Jay Kreps, Reed Hastings and, from February 2026, Chris Liddell, a former chief financial officer of Microsoft and General Motors and deputy White House chief of staff in Donald Trump's first term.⁴⁵⁶⁷

FTX Funding and the Effective-Altruism Tie

In 2022 Sam Bankman-Fried invested approximately 500 million dollars in Anthropic's Series B through his trading firm Alameda Research, taking about 86 percent of the round and roughly 8 percent of the company; the money came from FTX customer deposits. Dario Amodei has said he kept Bankman-Fried off the board and limited him to non-voting shares because of "red flags." After the FTX fraud was exposed and Bankman-Fried convicted, the FTX bankruptcy estate sold the stake, and Amodei moved to distance himself and the company from the effective-altruism label even as the founders remained products of that milieu. Later funding rounds drew on Google and Amazon, carrying the company's valuation into the tens of billions of dollars.¹⁸⁹

The April 2022 Series B that Bankman-Fried led raised 580 million dollars at a 4-billion-dollar valuation, with FTX cofounder Caroline Ellison also named in the round. The FTX estate disposed of the 8 percent stake in 2024 as Anthropic's value climbed: a March 2024 agreement sold about two-thirds of the shares (29.5 million shares to 24 buyers) for 884 million dollars, with ATIC Third International Investment Company, tied to the Abu Dhabi sovereign fund Mubadala, buying roughly 500 million dollars of it and Jane Street Capital, Fidelity-managed funds, and HOF Capital among the rest; a June 2024 sale of the remaining shares brought in over 450 million dollars. Anthropic's May 2023 Series C, led by Spark Capital with Google, Menlo Ventures, and Salesforce Ventures, had raised 450 million dollars, and the stake FTX paid 500 million dollars for would have been worth several billion at later valuations.¹⁰¹¹

The Funding Rounds

Anthropic's first outside capital was a 124-million-dollar Series A in May 2021 led by Jaan Tallinn, the Estonian programmer who had made his fortune as a founding engineer of Skype, with participation from the investor James McClave, Dustin Moskovitz, Eric Schmidt, and the Center for Emerging Risk Research. The lead and several of the backers came from the same effective-altruism and existential-risk donor world as the founders. The 580-million-dollar Series B that Bankman-Fried led the following year was followed in May 2023 by a 450-million-dollar Series C led by Spark Capital.¹²

From 2023 the valuation climbed through a rapid sequence of rounds funded increasingly by the largest technology firms and the late-stage growth funds. A March 2025 round of 3.5 billion dollars, led by Lightspeed Venture Partners, valued the company at 61.5 billion dollars. In May 2026 it closed a 65-billion-dollar round, the largest in its history, at a 965-billion-dollar post-money valuation, co-led by Altimeter Capital, Dragoneer, Greenoaks, Sequoia Capital, Capital Group, Coatue, and D1 Capital, briefly making it the most valuable private company in the artificial-intelligence industry, ahead of OpenAI. Anthropic reported an annualized revenue run rate of about 47 billion dollars and confidentially filed for an initial public offering in mid-2026.¹³¹⁴

The Amazon and Google Stakes

The two largest strategic investors are Amazon and Google, which are also Anthropic's principal cloud and chip suppliers. Amazon committed up to 4 billion dollars in September 2023, beginning with 1.25 billion, and added a further 4 billion in November 2024 to reach 8 billion dollars in total, becoming Anthropic's primary training partner through Amazon Web Services and its Trainium chips. Google, which had put roughly 300 million dollars into the company for about a 10 percent stake in early 2023, committed up to 2 billion dollars later that year and more than 1 billion again in January 2025, holding a stake reported at around 14 percent and contractually capped near 15 percent.¹⁵¹⁶

The arrangement is partly circular: much of the money the two companies invest returns to them as payment for the computing capacity Anthropic rents to train and run its models. By 2026 the appreciation of the two stakes was large enough that the financial press reported the Anthropic holdings accounted for roughly half of the AI-related investment gains Google and Amazon booked in some quarters, gains the companies had not realized because they had not sold the shares.¹⁶¹⁷

National Security Work

In November 2024 Anthropic, Palantir Technologies and Amazon Web Services announced a partnership to bring Claude to U.S. defense and intelligence agencies on classified networks at Impact Level 6; Palantir's chief technology officer, Shyam Sankar, said it gave them "the tool chain they need to harness and deploy AI models securely."¹⁸ In June 2025 the company introduced "Claude Gov" models "built exclusively for U.S. national security customers," already deployed "by agencies at the highest level of U.S. national security," which "refuse less when engaging with classified information."¹⁹ On July 14, 2025, the Pentagon's Chief Digital and Artificial Intelligence Office awarded Anthropic, Google, OpenAI and xAI prototype agreements worth up to 200 million dollars each.²⁰ In August 2025 Anthropic formed a National Security and Public Sector Advisory Council whose inaugural members included the former deputy CIA director David S. Cohen, the former NSA cybersecurity director Dave Luber, the former acting Secretary of Defense Patrick Shanahan, the former senators Roy Blunt and Jon Tester, the former ODNI general counsel Christopher Fonzone, two former administrators of the National Nuclear Security Administration, and the trust member Richard Fontaine.²¹

The Pentagon Dispute

On February 26, 2026, Dario Amodei published a statement that Claude was "extensively deployed across the Department of War and other national security agencies" for "intelligence analysis, modeling and simulation, operational planning, cyber operations, and more," and that Anthropic had been "the first frontier AI company to deploy our models in the US government's classified networks." The department, he wrote, would contract only with companies that accepted "any lawful use," and had threatened to remove Anthropic from its systems, to designate it a "supply chain risk" and to invoke the Defense Production Act unless it dropped two exclusions, mass domestic surveillance and fully autonomous weapons: "we cannot in good conscience accede to their request."²² President Trump then ordered federal agencies to stop using Anthropic's technology, and on March 5 the department told the company it would be labeled a supply-chain risk; Amazon said its cloud customers could keep using Claude for all work not associated with the department.²³ According to The Intercept, which obtained more than 400 pages of the contract documents through a Freedom of Information Act lawsuit, the break came over amendments that would have allowed deployment on classified networks without prohibitions on domestic spying and autonomous weapons; Secretary of Defense Pete Hegseth's designation was overturned by a federal judge in August 2026, and the department quickly reached classified-network agreements with Google, OpenAI and xAI.²⁴

Political Spending and Lobbying

On February 12, 2026, Anthropic announced a 20-million-dollar contribution to Public First Action, "a new bipartisan 501(c)(4)" founded by Republican and Democratic strategists, noting that "vast resources have flowed to political organizations that oppose these efforts."²⁵ Eight days later TechCrunch reported, citing Bloomberg, that Public First Action was "spending $450,000 to boost" New York Assembly member Alex Bores, sponsor of the state's RAISE Act, in the Democratic primary for the 12th congressional district, where the super PAC Leading the Future, backed by Andreessen Horowitz, Greg Brockman of OpenAI, Perplexity and Joe Lonsdale, had spent 1.1 million dollars on ads attacking him.²⁶ On July 21, 2026, Anthropic gave another 20 million dollars, "bringing our total support to $40 million," and stated that both donations "cannot be used to influence the election of any candidate for federal, state, or local office."²⁷

Anthropic's own lobbying disclosures reported expenses of 720,000 dollars in 2024, 4.04 million in 2025 and 3.53 million in the first half of 2026; the number of outside firms registered to lobby for it grew from two in 2023 to nine in 2026.²⁸