Dario Amodei began as a physicist and studied the brain and ways of detecting disease before turning to artificial intelligence. After working at Baidu, Google and OpenAI, he co-founded Anthropic in 2021; its Claude service competes with ChatGPT. His career has been shaped both by a desire to accelerate scientific progress and by concern about the safety of new technology.
From physics to medical research
Amodei was born in San Francisco in 1983 and grew up with his younger sister, Daniela Amodei, in the city’s Mission District. Their father was of Italian origin and their mother was American. After attending Lowell High School, he studied physics at the California Institute of Technology.
As a student, he was already concerned about the social consequences of science. At the start of the Iraq War in 2003, he used a university newspaper article to urge fellow students to protest the US invasion. He argued that scientists should consider the implications when their work could be used for war.
He later transferred to Stanford University, where he earned a physics degree, before continuing his studies at Princeton University. His doctoral research examined networks of nerve cells and physical and mathematical models of brain activity. As a postdoctoral researcher at Stanford University’s medical school, he studied proteins and methods that could help detect cancer.
A personal loss also shaped his interests: his father died following a hepatitis C infection, a few years before the arrival of drugs that now successfully treat many patients. Amodei has said in interviews that faster medical progress might have saved his father’s life. The experience contributed to his belief that computers and AI could accelerate drug discovery and improve understanding of disease.
From Baidu to OpenAI
Amodei left conventional research in 2014. Machine-learning researcher Andrew Ng recruited him to the US research center of Chinese technology company Baidu. There he worked on speech recognition and recognized early that larger models, more training data and greater computing power could substantially improve AI performance.
He then joined the Google Brain research group, where he worked on neural networks. In a 2016 paper with other researchers, he examined five problems that could cause AI systems to behave in unintended or dangerous ways. These included how to prevent a system from causing harm while carrying out a task or fulfilling an objective in a way that conflicts with human intentions.
Also in 2016, Amodei joined OpenAI, which had been founded the previous year with the involvement of Sam Altman and Elon Musk, among others. He later helped direct development as vice president of research. GPT-2 and GPT-3, technological predecessors of ChatGPT, were developed under his leadership. Introduced in 2020, GPT-3 proved capable of handling complex text and programming tasks.
Amodei also led safety research, including work on methods that use human feedback to improve model responses. But he and several colleagues disagreed with company leadership over the pace of commercialization and the attention given to safety. Amodei left OpenAI in 2020.
Anthropic and Claude
Dario and Daniela Amodei founded Anthropic in January 2021 with five former OpenAI colleagues. The founders included Tom Brown, Jack Clark, Jared Kaplan, Sam McCandlish and Christopher Olah. Dario took responsibility for research, technological development and long-term strategy, while Daniela directed daily operations, organizational structure and business activity.
Anthropic raised $124 million in 2021 from investors including Jaan Tallinn, a co-founder of Skype. It operates as a public benefit corporation, a structure that requires management to consider specified social goals alongside owners’ financial interests. In Amodei’s view, the structure could help preserve safety principles even when they slow development or reduce profits. The source says the company remains loss-making.
Claude became widely available in March 2023. Anthropic has focused on business customers from the outset; its later Claude Code service supports tasks including writing code, finding bugs and testing. Amazon and Google invested in the growing company and provided computing capacity, while Microsoft and other technology companies also became partners.
The source claims that Anthropic’s revenue was $4.6 billion in 2025 and $17 billion in the third quarter of 2026, with annualized revenue based on October reaching $81 billion.
Pentagon dispute and stock-market claims
According to the source, a dispute arose in 2026 between Anthropic and the US government over military use of the technology. The company sought restrictions preventing Claude from being used for mass surveillance of US citizens or to control fully autonomous weapons systems. Amodei held talks with Defense Secretary Pete Hegseth, but they did not reach an agreement. The source says the government blacklisted the company and that Anthropic challenged the decision in court.
The source also reports a stock-market listing in November 2026. It claims Anthropic sold $81.3 billion in shares and that the transaction reached $107.3 billion including a parallel private fundraising. The stated $83 offer price implied a fully diluted valuation of $1.5 trillion. Shares reportedly opened at $111.50 and rose to $115, briefly lifting the company’s value above $2 trillion. The source estimated Amodei’s wealth at more than $27 billion during the trading day.
According to the report, public trading brings a recurring dilemma in Amodei’s career into sharper focus: how to reconcile investors’ expectations of growth and profit with his position that development should be slowed when safety risks demand it.



