Imagine an AI that doesn’t just read the internet but learns from doing – making decisions, solving problems, and improving through real-world interaction. That’s the vision behind David Silver’s new venture, Ineffable Intelligence, which just secured what could be Europe’s largest seed funding round at $1 billion. But this isn’t just another AI funding story; it’s a signal that the industry is pivoting from data-hungry models to experience-driven systems that could reshape how businesses deploy artificial intelligence.
The Reinforcement Learning Revolution
Silver, the British researcher behind DeepMind’s AlphaGo and Google’s Gemini models, left Google last year to pursue what he calls “superhuman intelligence” through reinforcement learning. Unlike today’s large language models that train on massive text datasets, reinforcement learning systems learn through feedback and interaction with their environment. In a paper co-authored with Richard Sutton, Silver predicted that “experience will become the dominant medium of improvement and ultimately dwarf the scale of human data used in today’s systems.”
This approach could solve some of AI’s most pressing business challenges. Current models require enormous computational resources and struggle with complex, multi-step tasks. Reinforcement learning agents, by contrast, could independently complete sophisticated workflows based on simple instructions – potentially revolutionizing everything from customer service to supply chain management.
The Global Funding Frenzy
Silver’s $1 billion seed round, led by Sequoia Capital with potential investments from Nvidia, Google, and Microsoft, comes amid unprecedented AI investment. According to TechCrunch, nearly 20 U.S.-based AI startups raised $100 million or more in just the first two months of 2026. Anthropic secured $30 billion, Runway raised $315 million, and ElevenLabs pulled in $500 million – all while European players like France’s Mistral AI made strategic acquisitions to build sovereign infrastructure.
“This isn’t just about throwing money at AI,” says one industry analyst who requested anonymity. “These investments represent specific bets on different approaches to artificial intelligence. Silver’s focus on reinforcement learning versus Anthropic’s large language models or Mistral’s infrastructure play shows the market is maturing and diversifying.”
The Infrastructure Bottleneck
But here’s the catch: all these AI ambitions are colliding with a severe hardware crisis. Phison CEO Khein-Seng Pua warns that memory chip shortages could extend until 2030, driven by hyperscalers buying up available supply regardless of cost. DRAM prices have jumped roughly 7x in the last year, and small manufacturers may collapse by 2026 as they struggle to secure components.
This creates a paradox for companies like Ineffable Intelligence. While reinforcement learning promises more efficient AI systems, the infrastructure to run them is becoming increasingly scarce and expensive. As TechCrunch reports, memory management is becoming “a huge part of AI infrastructure costs,” with companies like Anthropic implementing complex caching strategies just to keep models running affordably.
The Business Implications
For enterprises, this means several things. First, AI deployment costs are likely to remain high as hardware shortages persist. Second, companies that master memory optimization and infrastructure efficiency will gain competitive advantages. Third, the shift toward experience-based learning could create new opportunities for businesses with proprietary data or unique operational environments.
Ben Bajarin of Creative Strategies notes that we’re moving from the “training era” to the “inference era,” which demands different approaches. This is evident in Meta’s multibillion-dollar deal for Nvidia’s next-generation chips as the social media giant plans to nearly double its AI infrastructure spending to $135 billion this year.
The European Angle
Silver’s London-based venture represents more than just another AI startup – it’s part of Europe’s push to establish technological sovereignty. With Mistral AI acquiring Paris-based Koyeb to accelerate its cloud ambitions and investing $1.4 billion in Swedish data centers, European companies are building infrastructure to compete with U.S. giants.
Floriane de Maupeou of Serena notes that such combinations “will play a key role in building the foundations of sovereign AI infrastructure in Europe.” This matters for businesses operating in regulated industries or those concerned about data sovereignty and geopolitical risks.
What Comes Next
The $1 billion question is whether reinforcement learning can deliver on its promise. While Silver’s track record at DeepMind suggests technical credibility, building “superhuman intelligence” remains an ambitious goal. The success of Ineffable Intelligence will depend not just on algorithms but on navigating hardware constraints, attracting talent, and finding practical business applications.
For now, the massive funding round signals that investors believe experience-based AI represents the next frontier. As companies grapple with the costs and limitations of current models, approaches that learn through interaction rather than data ingestion could offer a path forward. But they’ll need to solve the infrastructure puzzle first.