Emulate, a London AI company incorporated in August by three former Google DeepMind researchers, is negotiating a seed round of up to $700m at a post-money valuation of about $3.7bn. Index Ventures and Lightspeed Venture Partners are leading. The company has no product, no revenue and no website.
Emulate was founded last month by Jack Parker-Holder, Matthew McGill and Philip Ball, all of whom worked on DeepMind's Genie world-model programme. Parker-Holder led the development of Genie 2, and all three contributed to Genie 3, the system capable of generating interactive environments from text prompts.
The company is building world models: AI systems designed to simulate how physical environments behave and how actions change them, rather than predicting language in the manner of the large models behind ChatGPT and Gemini. The obvious commercial application is robotics, where a developer could test thousands of movements in a simulated warehouse before deploying hardware.
The round's terms have not been finalised and could still change. On the figures reported, a $700m cheque at a $3.7bn post-money valuation would hand new investors close to 19 per cent of the company, leaving the three founders and any employee pool to share the balance — an unusually low dilution rate for a seed round, and a reflection of how much leverage frontier research teams currently hold over capital.
That structure is the story. A conventional UK seed round trades 15 to 25 per cent of the business for single-digit millions. Emulate is proposing to trade a similar slice for a sum larger than most European Series C rounds, before it has shipped anything.
Forrester's vice president and principal analyst Charlie Dai, speaking to City AM, described the pricing as a bet on "future capability creation, not present revenue or commercial traction". Venture firms, he said, are now writing large cheques to research leaders with a track record in foundation models before any product launch, particularly in simulation and world understanding.
Emulate is the latest in a lengthening run of heavily capitalised British AI labs founded by DeepMind alumni, a pattern that has turned the London lab into a reliable source of deal flow for growth investors. It also lands in a contested category. World Labs, founded by Fei-Fei Li, has taken in upwards of $1bn for spatial intelligence work, and a cluster of companies is now aiming at simulation and robotics rather than competing head-on with the major language-model labs.
For the UK ecosystem, the round is a second data point of the same kind in a month. Capital is concentrating at the top, and increasingly it is arriving before there is a company to speak of.
Original source: City AM — https://www.cityam.com/month-old-deepmind-spinout-nears-4bn-on-future-capability-bet/
