Reflection AI, a New York-based startup founded by ex-Google DeepMind researchers, is raising over $1 billion to develop open-source large language models that rival Meta, DeepSeek, and Mistral. The company, which recently launched its coding agent “Asimov,” aims to position itself as the leading U.S. provider of open-source AI at a time when demand for flexible, customizable models is surging. With major backing from Lightspeed, Sequoia, and CRV, Reflection AI is betting on open-source as a competitive edge against proprietary AI giants.
