Nvidia announced an engineering partnership with Ineffable Intelligence, a London-based startup founded by former DeepMind reinforcement learning lead David Silver. The company, which raised a record $1.1 billion seed round in April, will work with Nvidia to build AI systems that learn through trial and error rather than human data. Nvidia CEO Jensen Huang called them “superlearners.”
What Is Ineffable Intelligence?
Founded in late 2025 by UCL professor and former DeepMind reinforcement learning team lead David Silver, Ineffable Intelligence is pursuing superintelligence through a fundamentally different approach than most AI labs. Instead of training on human-generated data, the company focuses on reinforcement learning — AI models that learn from their own experience and discover new knowledge independently.
What Does the Nvidia Partnership Entail?
Nvidia and Ineffable will enter into an engineering-level collaboration to build infrastructure for large-scale reinforcement learning. Engineers from both companies will team up, using Nvidia’s Grace Blackwell chips and Vera Rubin platform to create a pipeline that can feed reinforcement learning systems at scale.
Why Is This Approach Different From Other AI Labs?
Most leading AI models are trained on human data — text, images, and code created by people. Ineffable’s approach trains on “rich forms of experience that are quite distinct from human language and other human data,” potentially requiring novel model architectures and training algorithms. This could lead to systems that discover genuinely new knowledge rather than just pattern-matching existing human knowledge.
Key Takeaways
- David Silver’s Ineffable Intelligence raised a record $1.1B seed round in April
- Nvidia partnership focuses on large-scale reinforcement learning infrastructure
- Approach fundamentally different from human-data-trained models
- Will use Nvidia’s Grace Blackwell chips and Vera Rubin platform
- Jensen Huang: “The next frontier of AI is superlearners”
- Part of a wave of new AI labs founded by ex-Big Tech researchers
Frequently Asked Questions
What is reinforcement learning and why does it matter? Reinforcement learning is a training method where AI models learn from trial-and-error experience rather than labeled data. It’s how AlphaGo learned to beat the world champion at Go.
How does Ineffable’s approach differ from standard RL? The company is building systems that can learn continuously from experience at massive scale, potentially discovering knowledge that humans haven’t yet found.
When will we see results from this partnership? No specific timeline has been announced. The partnership is focused on building infrastructure first, with model capabilities to follow.