Anthropic is a San Francisco AI company founded in 2021 that builds the Claude family of language models and describes itself as an AI safety and research company. On October 26, 2026 it publicly filed to list its shares, disclosing $16.3 billion of revenue in the first half of 2026. This profile summarizes who founded the company, how it is structured, what it sells, who funds it and what its IPO paperwork reveals, using the company’s own statements and press coverage of the filing.
Origins and founders
Anthropic was started in 2021 by a group of researchers who had previously worked at OpenAI. The seven co-founders named in the filing, as reported by TechCrunch, are Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah. Dario Amodei is chief executive and Daniela Amodei is president. Their stated aim was to do frontier AI research with safety as a central design goal rather than an add-on, an approach the company calls building “reliable, interpretable, and steerable AI systems.”
The company’s early research output reflected that focus, including work on interpretability, preference-based training and a method it calls Constitutional AI, in which a model is trained to follow written principles. For a primer on the underlying field, see our explainer on AI alignment.
A public benefit corporation with a trust
Anthropic is a public benefit corporation, a legal form that lets directors weigh a stated public mission alongside shareholder returns. In 2023 it also created a Long-Term Benefit Trust, a body of independent trustees with the power to elect a portion of the board; the company describes the trust on its website. According to the S-1 as reported by TechCrunch, the trust will appoint most of the board after the listing, and the seven founders will form a “Founder LLC” holding 50.1% of voting power. Anthropic says the structure is meant to protect its mission, while critics point out that voting power will not match economic ownership.
Products: Claude and the business behind it
Claude launched publicly in 2023 and has since grown into a family of models offered through consumer apps, an API and cloud marketplaces. Developer tools such as Claude Code, an agentic coding product, have become a major part of its pitch. Earlier this month we covered Claude Duet, an always-on agent in research preview.
The business is weighted toward companies. TechCrunch reports that as of the end of September Anthropic had more than 11,900 customers spending at least $100,000 a year, about 1,900 spending at least $1 million and roughly 250 spending $10 million or more. The Guardian adds that the five largest customers account for 25% of the business, a concentration the company lists as a risk.
Money and compute
Anthropic has raised funding from large technology companies and investors, and it relies on several cloud and chip providers for compute. Reporting on the filing lists multi-year commitments with Google, Microsoft, Amazon and SpaceX; our chart of the week breaks down the $518 billion total reported earlier. The Guardian says Anthropic spent nearly $9 billion on AI infrastructure in the first half of 2026, up from $2.3 billion a year earlier.
Financially, the S-1 shows $16.3 billion in first-half 2026 revenue against $1.2 billion in the same period of 2025, and an operating loss of $693 million. The much larger $80.7 billion net loss was driven mainly by non-cash charges tied to convertible debt, per TechCrunch. Outlets have reported different 2025 loss figures, and we have not reconciled them against the filing itself.
Safety policies and oversight
Anthropic publishes a Responsible Scaling Policy that ties the safeguards it applies to the capabilities of its models, and it works with outside evaluators. We have reported on its arrangement giving METR employee-like access, and our explainer on who tests frontier AI places that in context. The company has also drawn scrutiny: the FTC sent civil investigative demands to it and OpenAI this month, as we reported.
How it compares
OpenAI is the closest rival in scale and product breadth; see our OpenAI profile for the contrast between the two origins. The S-1 names OpenAI and Google as direct competitors, according to The Verge. Anthropic’s pricing and enterprise focus differ from consumer-first rivals, but both depend on the same scarce resource: compute.
What to watch next
- The IPO price range, amount to be raised and listing date, none of which has been announced.
- Whether first-half revenue growth persists in the second half.
- How public investors respond to the Founder LLC governance structure.
- Regulatory outcomes from the FTC demands and the proposed AI safety legislation.
- Further disclosures on customer concentration and compute contracts.
Key takeaways
- Anthropic was founded in 2021 by seven researchers, several of them ex-OpenAI.
- It is a public benefit corporation with a Long-Term Benefit Trust.
- Revenue was $16.3 billion in the first half of 2026, mostly from business customers.
- The founders will hold 50.1% of voting power through a Founder LLC after listing.
- Compute commitments and customer concentration are the main risks it discloses.
Frequently asked questions
Who founded Anthropic?
Seven co-founders: Dario Amodei, Daniela Amodei, Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah. The company was founded in 2021.
What is Claude?
Claude is Anthropic’s family of AI models, available through apps, an API and cloud platforms. It is used for chat, writing, analysis and coding.
Is Anthropic publicly traded?
Not yet. It has filed an S-1 with the U.S. SEC and plans to list under the ticker “ANT,” but no price or date has been announced.
