Anthropic, PBC

Open full sized image

Anthropic launched Claude, its family of large language models, in 2023.

© Davide Bonaldo/SOPA Images—LightRocket/Getty Images

Anthropic PBC is an artificial intelligence (AI) company headquartered in San Francisco. Founded in 2021 by former OpenAI researchers, the company is the developer of the Claude AI family of large language models (LLMs).

Anthropic emphasizes a safety-focused approach to AI research and development and is organized as a public benefit corporation (PBC)—a for-profit venture aimed at making a positive public impact. It applies this principle through a training approach known as “constitutional AI,” a technique that trains AI systems to follow a structured set of ethical and safety guidelines. Anthropic also has a Long-Term Benefit Trust that helps ensure the company follows its mission of developing AI responsibly. The trust has considerable influence over the company’s board of directors.

Claude models are known for their long-context processing, structured reasoning capabilities, and comparatively cautious responses, which—in principle and design—aim to reduce harmful or noncompliant outputs. Claude is used by both consumers and businesses.

In 2021, Anthropic was founded by a group of former OpenAI researchers, including siblings Dario Amodei, the company’s chief executive officer, and Daniela Amodei, its president. Other cofounders included Jared Kaplan, a theoretical physicist and AI researcher, and Chris Olah, known for his work in mechanistic interpretability, a field focused on understanding how neural networks process and represent information internally.

Open full sized image

Dario Amodei is a leading voice in the debate over how quickly—and safely—AI should advance.

© Chance Yeh—HubSpot/Getty Images

Anthropic completed training its Claude 1 model in 2022, prior to the public release of OpenAI’s ChatGPT. The model was initially made available in limited access settings before a broader release in 2023. Company leadership cited the need for additional safety testing and refinement before expanding availability.

Open full sized image

Daniela Amodei cofounded Anthropic with her brother, Dario Amodei, in 2021.

© Kimberly White—Wired/Getty Images

Just months before Claude’s public launch, Anthropic published the model’s constitution in May, outlining the ethical and operational values guiding Claude’s performance.

Covering a detailed set of protocols and rationales, the core values underpinning Claude’s constitution can be broadly summarized along four categories: that the model is broadly safe, broadly ethical, compliant with Anthropic’s specific guidelines, and genuinely helpful to operators and users.

Anthropic also worked with Palantir (PLTR) and AWS to make Claude available to U.S. defense and intelligence agencies. The company restricted its use for autonomous weapons, malicious cyber activity, and domestic surveillance.

Anthropic also continued developing its AI models, releasing Claude 4 in May 2025. Claude 4 included Opus and Sonnet, which offered improved reasoning, speed, and coding capabilities.

After the release of Claude Gov in June 2025, Anthropic received a $200 million contract from the U.S. Department of Defense, formalizing its involvement in national security despite the company’s internal policies restricting its use for certain defense and military operations.

Anthropic also entered a multiyear, $200 million partnership with Snowflake (SNOW), a cloud-based data company. The deal embedded Claude models within Snowflake’s data cloud, making the product available to more than 12,000 business customers worldwide.

Anthropic’s funding rounds also accelerated, with the company raising $3.5 billion in March 2025 and another $13 billion in September, bringing its valuation to about $183 billion. Anthropic also faced copyright infringement controversies over material used to train its models. That same year, Anthropic agreed to a $1.5 billion settlement in a class-action lawsuit brought by authors over the use of pirated books to train its models.

By early 2026, tensions between Anthropic and U.S. defense officials over the company’s AI restrictions intensified. Anthropic’s policies prohibited the use of its systems for fully autonomous weapons and mass surveillance of U.S. citizens. Defense officials argued that the restrictions could hinder national security operations, particularly large-scale data analysis.

The dispute escalated after Anthropic declined to relax its safeguards. The Pentagon responded by moving to designate the company a “supply chain risk,” a step that threatened existing and future government contracts, including the $200 million defense agreement awarded in 2025. The contract was subsequently terminated as negotiations broke down.

In August 2026, a U.S. district court ruled that the government had failed to show that Anthropic posed a national security risk and had illegally punished the company for criticizing the Pentagon’s AI policies. The court ordered the government to withdraw its designation of Anthropic as a supply chain risk.

But in September, a federal appeals court upheld a separate supply-chain risk designation issued under a different law. The court rejected Anthropic’s arguments that the designation violated its constitutional rights. Anthropic requested a rehearing before the full appeals court, leaving the dispute unresolved.

Anthropic’s valuation continued to climb in 2026. In May, the company raised $65 billion in a funding round that valued it at $965 billion, up from $183 billion following its September 2025 round.

In June, Anthropic filed an S-1 registration statement with the U.S. Securities and Exchange Commission, taking its first formal step toward an initial public offering. The company made its registration statement public in October and the next month launched its offering of 825 million shares at an expected price of $70 to $78 apiece. At the midpoint of that range, Anthropic would be valued at roughly $1.3 trillion. The company applied to list its shares on Nasdaq under the ticker symbol ANT.

Anthropic’s IPO filings also provided a more complete look at its finances. Revenue rose to nearly $4.6 billion in 2025 from $386 million in 2024. Revenue reached $33.2 billion in the first nine months of 2026, while the company reported an operating loss of about $1.3 billion. Anthropic also disclosed $900.9 billion in purchase and lease commitments, nearly all of it related to the computing infrastructure needed to develop and operate its AI models.

The IPO also brought greater attention to Anthropic’s unusual approach to corporate governance. Under its proposed structure, the company’s seven cofounders would collectively control 50.1% of the voting power on matters generally submitted to shareholders. Separately, Anthropic’s Long-Term Benefit Trust would have authority to appoint a majority of the board of directors, giving it considerable influence over the company’s direction.

Anthropic also established a Public Benefit Fund to support charitable initiatives focused on addressing the effects of AI. The company contributed 5% of its shares to the fund and said it plans to make additional contributions as its stock price rises.