Who Owns Claude? The Quiet San Francisco Company Reshaping the AI Race

It seems like a straightforward question. Claude is owned by whom? When you type it into any search bar, contradictory results appear: Amazon, Google, and an unidentified group of tech behemoths. The reality is more subdued and somewhat unfamiliar. Claude is owned by Anthropic, a business with its headquarters located in a rather modest building on San Francisco’s Howard Street—the kind of address you wouldn’t notice if you weren’t looking for it. There is no campus, no fountain, and no spectacle with glass walls. Just an AI lab worth about $380 billion as of early 2026.

The Amodei siblings, Dario, the CEO, and Daniela, the president, were the most notable of the seven people who founded Anthropic in 2021 after leaving OpenAI. Disagreements over direction were a factor in their departure, though no one who departed has ever publicly stated it. Instead, they developed a research methodology they call Constitutional AI, which trains models against a written set of values rather than just human feedback, and a company that loudly insists on AI safety as a founding principle. How much of this is true philosophy and how much is marketing can be questioned. The philosophy appears to be holding up better than most observers anticipated when looking at the company’s actual decisions over a five-year period.

The ownership confusion begins with the investor list, which reads like a who’s who of capital. Google has contributed billions, including two rounds of $2 billion and $300 million in early 2023. Through cloud and infrastructure agreements, Amazon has been an even bigger backer. Somewhere on the cap table are Salesforce Ventures, Spark Capital, and Zoom Ventures. According to reports in late 2025, Nvidia and Microsoft were investing up to $15 billion in total, with Anthropic agreeing to spend $30 billion on Microsoft Azure compute powered by Nvidia chips. None of these agreements grant any one business control over the model or the business.

That has a structural explanation, which is why the majority of coverage ignores it. In addition to being a Public Benefit Corporation, Anthropic has a Long-Term Benefit Trust that has a unique class of shares called Class T, which gives the trustees the power to choose board directors. The current trustees are Richard Fontaine, Mariano-Florentino Cuéllar, and Neil Buddy Shah—names that most people outside of the policy community wouldn’t be familiar with. Theoretically, their role is to maintain the company’s focus on long-term human welfare rather than the next quarter’s earnings. The question of whether that can withstand actual financial strain is still unanswered.

Observing Anthropic’s operations gives the impression that it’s attempting to thread a needle that no AI company has truly managed before. It has refused Department of Defense demands to lift restrictions on autonomous weapons and domestic surveillance — and absorbed a “supply chain risk” designation as a result. Groups with a majority stake in China, Russia, Iran, and North Korea were denied commercial access. It aired Super Bowl ads in February 2026 promising Claude would stay ad-free, a not-very-subtle jab at OpenAI. These are choices that cost real money, and they tell you something about who’s actually steering the ship.

So who owns Claude? Anthropic does. Anthropic is owned, in the conventional sense, by its founders, employees, and a long list of investors who hold equity but not control. The Long-Term Benefit Trust holds the unusual lever that determines who sits on the board. The establishment is still run by Dario and Daniela Amodei. Mike Krieger, the former Instagram co-founder, runs product and now a new division called Labs. The search box doesn’t want a more complex response. It’s also probably the most honest one available, at least until the next funding round rearranges the picture again.