DeepMind Co-Founder David Silver Raises 1.1 Billion Dollars for AI Research Without Human Training Data

David Silver, the DeepMind researcher who led the development of AlphaGo, has raised $1.1 billion for a new AI research venture focused on building systems that learn without relying on human-generated training data.

The approach marks a significant departure from the method used by most large language models, which are trained on vast datasets of text and other content produced by humans. Silver's work at DeepMind demonstrated that AI systems could achieve superhuman performance in complex domains like chess and Go by learning through self-play and reinforcement learning rather than human examples.

The new venture aims to apply similar principles more broadly, developing AI that can reason and solve problems from first principles rather than pattern-matching against human-generated content.

The $1.1 billion raise reflects continued investor appetite for AI research that could produce capabilities beyond what current large language models offer. It also signals growing interest in approaches that sidestep concerns about the quality, bias, and copyright status of human-generated training data.

Silver's work at DeepMind produced AlphaZero, which mastered chess, shogi, and Go from scratch, and AlphaFold, which predicted the three-dimensional structure of proteins and accelerated drug discovery research.

Details about the new venture's specific research agenda and timeline have not been publicly disclosed. The fundraise was reported in late April 2026.