Emulate, a London startup founded barely three months ago, has raised a seed round of up to $700 million co-led by Index Ventures and Lightspeed Venture Partners, with Nvidia's investment arm NVentures and game-engine maker Unity joining alongside earlier backer Creandum, according to reporting confirmed by KuCoin News. The round values Emulate at $3.7 billion as of November 11, 2026. The company was incorporated in August 2026 by three former Google DeepMind researchers — Jack Parker-Holder, Matthew McGill and Philip Ball — who helped build DeepMind's Genie world models, and it still has no public product, website or disclosed headcount.
What exactly is a world model?
A world model is an AI system trained on video and interactive data to learn the rules the physical world runs on: how objects fall, bend, break or collide, and how a scene changes after a given action. Instead of generating a single image or a block of text, a world model can simulate an entire interactive environment — forecasting, for instance, how a glass shatters when dropped or how a robot arm needs to move to avoid clipping a shelf.
That makes world models useful well beyond games or video generation. A robotics team can run a warehouse-picking strategy through thousands of simulated attempts inside a world model before ever testing it on a physical robot, catching failures that would otherwise mean broken hardware or damaged inventory. Genie, the DeepMind research line Emulate's founders worked on, generates interactive 3D environments from a short prompt — the technical seed that Emulate is now trying to turn into a standalone company and, eventually, a product.
Why would investors write a $700 million check for a company with no product?
Because the market they're betting on is still forming and the researchers are scarce. Industry estimates cited alongside the round put the broader market for AI software that trains robots and autonomous systems — sometimes grouped under the "physical AI" label — growing from $1.5 billion in 2026 to $15.24 billion by 2032. Investors who want exposure to that curve have few places to put large checks: most teams with deep world-model research experience are still inside labs like DeepMind, OpenAI or Meta rather than running independent startups.
Nvidia's and Unity's involvement signal two different bets. Nvidia wants more workloads — training and running world models is compute-intensive — running on its chips. Unity, whose engine already renders the 3D environments many robotics and gaming companies use for simulation, is positioning itself as infrastructure for whatever Emulate eventually ships.
What's next for Emulate?
The company has given no public timeline for a product launch, and reporting on the round has consistently noted the absence of a disclosed website or team size — unusual transparency gaps for a company commanding a $3.7 billion valuation. Emulate joins a small but growing list of well-funded world-model startups competing for talent and compute alongside in-house efforts at DeepMind, Nvidia and Chinese labs building similar robot-simulation platforms. The next concrete signal to watch for is any announcement of a working demo, a named robotics or gaming partner, or hires that reveal whether Emulate is building a general-purpose simulator or something narrower aimed squarely at robot training.