nvidia-ineffable-intelligence-ai-superlearners-reinforcement-learning

Nvidia has entered a major new partnership with UK-based AI startup Nvidia, joining forces with Ineffable Intelligence to develop a new generation of artificial intelligence systems that learn from experience instead of relying only on human data.

The collaboration comes just months after Ineffable Intelligence launched in late 2025. The startup was founded by former Google DeepMind reinforcement learning lead David Silver, now a professor at University College London University College London. The company is positioning itself at the center of what it calls “superlearners,” AI systems designed to continuously improve through trial and error.

Nvidia CEO Jensen Huang said the partnership represents a major shift in how AI will evolve. He explained that the next phase of AI will move beyond static knowledge and toward systems that constantly learn from real-world experience. He added that Nvidia is working closely with Ineffable to design infrastructure for large-scale reinforcement learning.

Unlike traditional models trained on massive datasets of human-generated text, Ineffable focuses on reinforcement learning. This approach allows AI systems to learn by interacting with environments, making mistakes, and improving over time. The company believes this method can unlock new levels of intelligence that current models cannot reach.

The partnership will use Nvidia’s advanced computing platforms, including its Grace Blackwell chips and Vera Rubin systems, to power large-scale training pipelines. Engineers from both sides will work together to build infrastructure that can support continuous learning models at scale.

Silver described the shift as a fundamental change in AI research. He said most current systems already know what humans know. However, the real challenge lies in building machines that discover new knowledge on their own through experience rather than imitation.

Investor interest in the sector continues to surge. Ineffable Intelligence recently raised a record $1.1 billion seed round, backed by major global investors including Sequoia Capital, Lightspeed Venture Partners, DST Global, and Index Ventures. The round also included participation from Nvidia and other strategic backers.

The momentum reflects a broader wave across the AI industry. New research labs led by former engineers from companies like OpenAI, DeepMind, Anthropic, and xAI are attracting billions in funding within months of launching. These startups are racing to define the next stage of artificial intelligence development.

On the same day, another startup, Recursive Superintelligence, founded by a former DeepMind engineer, announced a $650 million raise. Meanwhile, AMI Labs secured $1 billion in funding earlier this year, showing how aggressively investors are betting on next-generation AI systems.

In London, Ineffable Intelligence is becoming one of the most closely watched players in this new AI race. Its focus on reinforcement learning places it in direct contrast with language-model-heavy approaches that dominate today’s AI landscape.