The History of Anthropic: From OpenAI Exodus to Claude

Anthropic's founding story is one of the more interesting chapters in recent tech history. In 2021, a group of researchers left OpenAI — one of the best-resourced AI labs in the world — to start a new company focused on AI safety. That group included Dario Amodei, Anthropic's CEO, and his sister Daniela Amodei, the President, along with several other senior researchers.

The reasons for leaving were never stated officially, but publicly available context paints a picture of philosophical disagreements about how fast to move and how much to prioritize safety research relative to capability development. OpenAI had taken a major investment from Microsoft and was moving fast toward commercialization. The Anthropic founders wanted to move deliberately, with safety as the primary mission, not a feature.

Anthropicraised substantial funding from the start — Amazon became a major investor, alongside Google and several venture capital firms. The investment sizes reflect how seriously the investor community takes AI as a transformational technology, and also how much it costs to actually train competitive models.

The first public version of Claude launched in 2023. It was well-received immediately, particularly for its thoughtful responses and its willingness to engage substantively with difficult questions. Subsequent versions — Claude 2, Claude 3 in its Haiku/Sonnet/Opus variants — expanded capability and added features like larger context windows.

Anthropic's position in the AI landscape is distinctive: it's one of a small number of companies with both the technical capability to build frontier models and a genuine philosophical commitment to doing so responsibly. Whether that combination is sufficient to navigate the risks ahead is the open question at the heart of everything the company does.