A former DeepMind researcher has raised $55 million in a pre-seed round that values the company at $300 million, according to TechCrunch, before the startup has shipped a product, signed a customer or, by most accounts, settled on a final name. The round, led by a top-tier venture firm, is a bet on a person rather than a business.
Pre-seed rounds of this size were unheard of a few years ago. The stage is meant for founders testing an idea, typically raising a few million dollars from friends, angels and early-stage funds. A $55 million pre-seed implies the investors believe the researcher can build something transformative, and it compresses what usually takes years into a single wire transfer.
The deal is the latest extreme in a market that has been inflating the price of AI talent. Top researchers have become the scarcest resource in the technology industry, and venture firms have concluded that the fastest way to own the next big AI company is to fund the people most likely to start it. Names from OpenAI, Google DeepMind and Anthropic now command valuations at formation that used to require a working product.
The inflation of early-stage AI deals has been building for years. Mistral, the French AI lab, raised its first round of more than $100 million at a valuation above $2 billion within weeks of its founding, before releasing a product, and Character.AI raised $150 million at a billion-dollar valuation on the strength of its founder’s reputation. Each deal pushed the envelope further, and each became the benchmark for the next. A $300 million valuation for a pre-seed would have been dismissed as fiction in 2023; today it is a data point.
DeepMind’s pedigree carries particular weight in this market. The Alphabet-owned lab produced the research that underpins much of modern AI, and its alumni are among the most sought-after founders in the industry. The researcher at the center of Thursday’s deal worked on some of the lab’s most prominent projects, according to people familiar with the matter, though the specifics of the work have not been disclosed, and the new company’s direction remains vague beyond a stated interest in advancing AI research.
The contrast with the rest of the investing world could hardly be starker. The same week the pre-seed was announced, BP said it would close its corporate venture arm, a unit that had operated for two decades and backed hundreds of startups across energy and technology. The oil giant’s retreat is part of a broader pullback by traditional energy companies, which have been systematically exiting technology venture investing after years of middling returns.
The two stories describe the same market from opposite ends. In AI, capital is flooding toward individuals with the right pedigree, often before they have built anything. In legacy industries, capital is flowing out of venture investing entirely, as companies decide the returns don’t justify the risk. Both dynamics are visible in the same week’s headlines, and both say something about where the money thinks the future is.
BP’s exit from venture investing tells a different story about the same week. The company’s venture unit, one of the oldest corporate funds in the energy industry, backed everything from solar startups to software companies over its two decades, but its returns never matched the industry’s, and the energy transition it was designed to hedge has moved more slowly than expected. BP is not alone: several European energy majors have wound down or shrunk their venture operations in recent years, concluding that the technology bets of tomorrow belong in funds run by specialists, not in the treasuries of oil companies.
The coincidence of the two announcements is the kind of juxtaposition the markets love and investors ignore at their peril. Capital is not neutral; it flows toward where it expects the highest returns, and this week it moved in two directions at once: into a researcher who has not built anything, and out of a corporate fund that built a portfolio of hundreds of companies. The first bet assumes the future belongs to AI’s creators. The second bet assumes the future of energy belongs to whoever produces it cheapest, with technology bought, not invented, on the side.
For the researcher at the center of the deal, the money comes with expectations that are hard to overstate. A $300 million valuation on day one means the company must eventually justify that number in revenue, and the investors who wrote the checks will want to see progress on a timeline measured in quarters, not years. The freedom that a pre-seed round is supposed to provide comes with a clock attached.
Whether the bet pays off is unknowable, which is the point. Venture capital is a business of outliers, and the firms writing these checks are playing for outcomes where one success covers a hundred failures. The question the market is answering with these deals is not whether the researcher will succeed, but whether the price of the ticket has become so high that even the winners struggle to produce a return.
For now, the startup has what most companies spend years pursuing: money, attention and a valuation that puts it among the most richly priced young companies in the world. What it doesn’t have is a product. That comes next, and the industry will be watching.
