- Former DeepMind researcher David Silver raised $1.1 billion in seed funding for Ineffable Intelligence at a $5.1 billion valuation.
- The London-based startup wants to build “superlearners” — AI that discovers knowledge through trial and error, not by ingesting human data.
- Silver pledged to donate all equity proceeds to charity, calling superintelligence “a huge responsibility” that must serve humanity.
David Silver, the researcher behind AlphaGo and AlphaZero, just raised the largest seed round in AI history. His startup, Ineffable Intelligence, pulled in $1.1 billion at a $5.1 billion valuation, according to the company’s newly launched site. That figure puts a months-old London lab into pentacorn territory — companies valued above $5 billion before shipping a product.
The round was led by Lightspeed Venture Partners, with backing from a roster of investors betting that the next breakthrough in AI won’t come from bigger training datasets. It’s a contrarian wager at a moment when OpenAI, Google, and Meta are spending hundreds of billions to scrape, license, and generate ever more human data. Silver thinks that entire approach is a dead end.
“Human data is like a kind of fossil fuel that has provided an amazing shortcut,” Silver told WIRED. “You can think of systems that learn for themselves as a renewable fuel — something that can just learn and learn and learn forever, without limit.”
What Ineffable Intelligence Is Actually Building
Ineffable’s core thesis is straightforward: the current LLM paradigm — scrape the internet, train on human output, fine-tune with human feedback — produces systems that can only ever replicate what humans already know. Silver calls this a fundamental ceiling. Reinforcement learning, by contrast, lets AI agents discover knowledge through trial and error inside simulations, building their own understanding from scratch. AlphaGo taught itself to play Go without any human game records. AlphaZero did the same for chess, Go, and shogi. Ineffable wants to scale that philosophy far beyond board games — to science, engineering, and economics.
“I think of our mission as making first contact with superintelligence,” Silver said. “By superintelligence I really mean something incredible. It should discover new forms of science or technology or government or economics for itself.” He compares the current state of AI to biology before Darwin — lots of observations, no unifying law. Ineffable’s goal is to find that law for intelligence itself.
The company’s site makes the ambition explicit: “If successful, this will represent a scientific breakthrough of comparable magnitude to Darwin: where his law explained all Life, our law will explain and build all Intelligence.” That’s a lot to claim for a startup that hasn’t disclosed a product, a timeline, or a specific architecture. Silver has reportedly recruited top researchers from DeepMind and other frontier labs, but details about what they’re actually building remain thin.
Why Investors Are Betting Billions on a Seed Round
A $1.1 billion seed round is unprecedented, even by the inflated standards of 2026 AI funding. But it follows a clear pattern: star researchers with track records are attracting nine-figure checks before writing a line of code. Recursive Superintelligence, co-founded by DeepMind’s former principal scientist Tim Rocktäschel, raised $500 million earlier this year with enough demand to stretch toward $1 billion. The common thread isn’t a product — it’s a person.
Ravi Mhatre, a cofounder and partner at Lightspeed Ventures, which is backing Ineffable, called Silver “a world-class researcher” whose career is “basically a single, coherent argument for being able to scale intelligence without human priors.” That argument has been playing out since 2016. AlphaGo learned Go from scratch. AlphaZero mastered three games without human data. MuZero learned without even being told the rules. Each result proved that reinforcement learning can produce superhuman performance without the data treadmill that LLMs depend on.
The question Ineffable has to answer is the same one that has haunted reinforcement learning for a decade: how do you go from confined environments like board games to the mess of the real world? Silver says he sees a path through simulations — placing AI agents in virtual environments where they can learn to achieve goals and collaborate. He’s cagey about what those simulations look like, which is either strategic restraint or an admission that the hardest problems haven’t been solved yet.
There’s also the matter of what Silver personally stands to gain — or rather, what he won’t. He pledged to donate all equity proceeds from Ineffable to charity. “It’s a huge responsibility to build a company focusing on superintelligence,” he told WIRED. “I think this is something that has to be done for the benefit of humanity, and any money that I make from Ineffable will go to high-impact charities that save as many lives as possible.” A billionaire-equivalent pledge before the company has a product is either remarkably principled or remarkably convenient for recruiting talent in a competitive market. It’s probably both.
Silver met DeepMind CEO Demis Hassabis at a chess tournament when they were kids. He joined DeepMind in 2011 and left only because he wanted a lab fully dedicated to reinforcement learning — “not just a corner of another place dedicated to LLMs,” as he put it to WIRED. OpenAI published its own AGI framework this week, a five-principle document built around LLM scaling. Silver thinks that entire path leads to a ceiling. He just raised enough money to prove it — or to discover that the real world is harder than Go.
