From Safety Concerns to Billions: The Defectors Who Built Anthropic

The rise of Anthropic began not with a quest for wealth, but with a dramatic exit. In late 2020, siblings Dario and Daniela Amodei walked away from their high-ranking vice presidency roles at OpenAI, citing deep frustrations over how the company approached artificial intelligence safety. Fearful that AI capabilities were advancing far quicker than our ability to control them, the duo staged a defection, bringing along five colleagues who shared their concerns regarding OpenAI’s pivot toward aggressive commercialization. This core group included researchers Tom Brown, Jared Kaplan, Sam McCandlish, and Chris Olah, as well as policy expert Jack Clark. Together, they sought to create a sanctuary for safer development, though that mission is now coinciding with one of the largest financial windfalls in corporate history.

As Anthropic prepares to go public on November 19, these seven founders are positioned to become some of the wealthiest individuals in the tech industry. With each holding roughly a 1.4 percent stake in the company, current valuations suggest their individual holdings are worth upwards of 14 billion dollars. Should the initial public offering hit its higher targets and drive the company’s market valuation toward 1.4 trillion dollars, those personal fortunes could soar to between 17 and 18 billion dollars each. While critics and former associates note that such massive incentives inevitably influence decision-making, insiders maintain that the venture was born out of a genuine desire to prevent others from mishandling a powerful global technology.

The leadership team reflects a diverse blend of academic rigor and strategic grit. CEO Dario Amodei brings a background in computational neuroscience from Princeton and experience at Google Brain and Baidu, viewing AI as a tool to solve complex biological mysteries. His sister Daniela, serving as president and board chair, complements this technical depth with her experience in political campaigning and risk management at Stripe. Meanwhile, Chief Compute Officer Tom Brown adds an entrepreneurial edge; despite admitting he once struggled with linear algebra in college, Brown became a pivotal figure in machine learning research and played a primary role in training GPT-3 before joining the exodus to form Anthropic.