The most interesting number in AI funding right now isn't a valuation. It's a multiplier. According to Bloomberg, venture capitalists are fighting so hard to invest in Emulate that they're buying into the same fundraising round at escalating prices — driving the startup's valuation up roughly 15-fold in days. "Venture capitalists are fighting to invest in AI startup Emulate, driving its valuation into the billions — even though the company is only three months old and doesn't yet have a website or product."

Read that again. The same round, at multiple prices. Index Ventures and Lightspeed Venture Partners got in first at around a $200 million valuation, according to people familiar with the matter, and Creandum AB then bought in at a higher figure. That's not a fundraise anymore — it's an auction where the founders get to sell the same shares twice and let the market argue about the spread.

The founders, to be fair, have the credentials VCs are bidding on. Jack Parker-Holder, Matthew McGill, and Philip Ball worked on DeepMind's Genie world models — the systems that generate interactive, navigable 3D worlds from a text prompt. Emulate, founded just three months ago, is reportedly building AI systems that model and simulate physical-world dynamics: the world-model bet that AI's next frontier isn't chatbots, it's machines that understand how the physical world behaves.

This is bigger than one startup. Emulate is the third major AI lab to spin out of DeepMind in 2026: David Silver's Ineffable Intelligence raised roughly $1 billion at a $5 billion valuation, and Recursive Superintelligence raised more than $600 million at around $4 billion. As former DeepMind scientist Janusz Marecki told Bloomberg: "Valuations used to be 10 times revenue. Now it's 200 times, 300 times." Credentials have decoupled from companies — the résumé is the asset class.

The skeptical read, and it's a fair one: when a seed round reprices 15x in days with no product, the leverage sits entirely with the founders, and the late bidders are paying a peak-hype price with no evidence to grade against. If world models turn out to be a five-year research problem rather than a two-year product one, the investors who paid $200 million and the investors who paid billions are holding very different lottery tickets in the same drawing.

Watch next: whether the round actually closes at a billion-plus valuation, and whether Index and Lightspeed's early low-price shares make the latecomers' math look brilliant or absurd. And note the bar Emulate has to clear — the founders' previous employer already ships their prior work: Project Genie is publicly available to subscribers. They're not just selling the future. They're competing against their own past.

— via Bloomberg