Emulate
Developing AI world models that simulate physical environments for robotics, testing and other real-world applications
The founders
PB
JP
The company
Emulate is developing AI world models — systems that understand, generate and simulate interactive physical environments, from 3D scenes to full interactive worlds. The company says the technology could underpin robotics, simulation, gaming, embodied AI and the training of autonomous agents.
As AI moves beyond language, the race is on to give agents an internal model of how physical environments actually behave, so they can predict outcomes and act rather than just talk. Emulate is betting the next foundation layer is simulated worlds, not more text.
Founded in London in August 2026 by former Google DeepMind researchers Jack Parker-Holder, Matthew McGill and Philip Ball, Emulate emerged from the team behind DeepMind's Genie world-model programme. It launched as an independent company rather than a formal DeepMind spinout.
The founding trio previously worked on Genie, DeepMind's family of generative world models capable of producing videos and interactive 3D environments from prompts — the exact problem Emulate is now pursuing independently.
Who else is building in this space
Funded · $14M
Natural language abstraction for AI-native software development.
Funded · $500K
Scientific instrument signal data for AI training.
Funded · $125K
Voice agents that talk to every employee to map how work gets done and support transformation
Funded · $125K
Frontier training data for AI and robotics.
Thousands more founded every month.
Dealroom tracks new companies as they appear — founders, funding and signals, over the API or in the platform.