To understand a company’s character, the fastest route is often to look at who’s running it. Anthropic’s answer is unusual: it was founded by eight people in 2021, and all eight came from the same place: OpenAI.
This piece introduces Anthropic’s founding team and leadership: where CEO Dario Amodei comes from, what role each founder plays, and which executives joined later. If you’d rather start with the company as a whole, begin with What Anthropic is as a company.
To set the tone in one line: this is a team that “walked away together to make AI safer,” and that shared conviction is the starting point for understanding Anthropic.
Dario Amodei: A CEO Who Started as a Scientist
The face of Anthropic is its co-founder and CEO, Dario Amodei.
Born in 1983, he started out as a physicist, with full training in physics at Stanford and Princeton. Early on he spent time at Baidu and Google, then joined OpenAI and rose to VP of Research, deeply involved in developing GPT-2 and GPT-3. In other words, more than a businessman-type CEO, he is closer to a genuine researcher who understands how models are actually built.
At Anthropic, Dario is the company’s primary public voice. He speaks frequently on AI safety and regulation, appears at policy forums, and testifies before Congress, to the point of becoming the representative figure for the “AI safety camp” within the industry, and he was named to the TIME 100 list of most influential people. Notably, he himself says he spends as much as 40% of his time on company culture and direction rather than specific products; for a company expanding at high speed, that allocation of time is itself a signal.
Daniela Amodei: The Other Half Holding Up Operations
If Dario handles the outward-facing role and the vision, then the person holding up the company’s day-to-day is his sister, President Daniela Amodei.
She also came from OpenAI, where she led safety and policy operations. At Anthropic, she oversees business operations and the build-out of the executive team; the regulatory experience she accumulated during her OpenAI years later became a core capability that let Anthropic break into highly regulated industries like finance, healthcare, and government. The sibling pairing, one facing outward and one facing inward, is the most distinctive feature of the company’s leadership structure, and both were named to the TIME 100 list of most influential people.
Eight People Who Walked Away from OpenAI
Anthropic’s founding story is dramatic in its own right: in 2021, eight people left OpenAI together, and their reasons were fairly consistent: they believed OpenAI was commercializing too fast and that its safety mechanisms couldn’t keep up. These eight each went on to play different roles at the company:
| Founder | Role at Anthropic | Background at OpenAI |
|---|---|---|
| Dario Amodei | CEO | VP of Research, core to GPT-2/GPT-3 |
| Daniela Amodei | President | Head of safety and policy operations |
| Jared Kaplan | Chief Science Officer / Head of Responsible Scaling | One of the founders of scaling laws research |
| Jack Clark | Head of Policy | Lead of the policy team |
| Chris Olah | Interpretability research | Pioneer of neural network interpretability |
| Tom Brown | Core technology | One of the lead authors of the GPT-3 paper |
| Ben Mann | Core technology | Early technical backbone |
| Sam McCandlish | Core technology | Referred to as CTO in later filings |
This roster reveals Anthropic’s DNA: at its foundation is a team of “old colleagues” drawn from research, safety, and policy leadership who walked away together out of a shared concern, rather than a hastily assembled group of entrepreneurs. It also explains why safety is written into its charter and its technology; for the related governance design, see Why Anthropic is a public benefit corporation.
Executives Who Joined Later
Beyond the founding team, Anthropic has also brought in heavyweight executives over the past few years to support the functions a large company needs:
- Krishna Rao (CFO): Joined in 2024 from Databricks, taking over the company’s massive financing and financial operations.
- Mike Krieger (Product): Co-founder of Instagram, who joined as Chief Product Officer in 2024; in early 2026 he moved into the newly formed Anthropic Labs, bringing Tom Brown and Ben Mann along to do more experimental product exploration.
- Jan Leike (Alignment research): A major alignment researcher who leads the alignment science direction.
These additions mark Anthropic’s transition from a research-heavy startup into a company with a full set of commercial functions.
A Company with “Walking Away” in Its DNA
Step back to look at this team, and you’ll find that “talent flow” is a recurring theme in Anthropic’s story.
The company itself was born by walking away from OpenAI; later, that talent channel continued to flow in both directions. For example, OpenAI co-founder John Schulman joined Anthropic in 2024 and left a few months later. In early 2026, Mrinank Sharma, a lead on the safety research team, publicly announced his resignation, which briefly sparked outside discussion about the loss of safety talent. These movements remind us that in the small circle of frontier AI, where top talent comes and goes is itself a signal worth watching.
But Anthropic’s core founding team has been quite stable, and that stability, combined with a shared starting point of “walking away for safety,” forms its most distinctive corporate character: a group of people who could have stayed at the hottest company chose to strike out on their own and prove that another path could work.
The Parts Not Yet Laid Out
When discussing people, it’s easy to mistake rumor for fact. Here we honestly mark a few areas that remain unclear:
- CTO handover: Sam McCandlish was referred to as CTO in 2024 filings, and Rahul Patil was listed as CTO in later announcements, but the timing and details of any formal handover have not been fully explained by the company.
- Founder equity: The specific ownership stakes of Dario, Daniela, and others have not been disclosed by the company; outside estimates of individual net worth (such as Forbes once estimating Dario’s net worth at around several billion dollars) are third-party projections, not company disclosures.
- Some executive titles: The exact titles of a few executives after they changed units are not fully clear from official sources.
These gaps remind us that quite a few claims about “who holds how much power at Anthropic” are still at the level of estimation.
Penchan’s Take
A company’s founding team is often the source of its character. That is especially true for Anthropic.
These eight people could have stayed at OpenAI, in what was then the most closely watched position, and shared in the explosion that followed. But they chose to leave together in 2021, to bet on the proposition that “safety can run in parallel with capability.” Whether that choice was right, and whether the bet was worth it, will take many years to see clearly; but it did give Anthropic a different foundation from its rivals from day one. For anyone trying to understand this company, remembering where this team came from is often more useful than remembering its latest model version.
Further reading: What Anthropic is as a company, Why Anthropic is a public benefit corporation, Anthropic vs OpenAI.