A London startup building AI world models is reportedly raising $700 million, with Nvidia’s investment arm and Unity joining the round. If completed on the reported terms, it would give a young company with no public product a striking roster of strategic backers.
Watch Desk analysis
What happened
MLQ.ai says Sifted reported that NVentures, Nvidia’s investment arm, and software maker Unity joined Emulate’s $700 million funding round, valuing the startup at $3.7 billion. Emulate, Nvidia and Unity have not announced the investment, according to the report.
Emulate was founded in August by former Google DeepMind researchers Jack Parker-Holder, Matthew McGill and Philip Ball. The founders worked on DeepMind’s world-model team, and Parker-Holder was a principal researcher on Genie 3, according to the account. Emulate has no public product yet.
Why it matters
World models aim to learn how environments behave, so AI systems can simulate them. That could matter well beyond games: simulations can help train robots and AI agents without relying entirely on costly, slow real-world trials. The reported investors bring two relevant angles, AI computing hardware and software for building interactive 3D worlds.
The money is also a signal of how much investors are willing to put behind this area of AI. But a large round is not evidence that a useful product is nearly ready. Emulate has not publicly launched one, so the work and its eventual applications remain to be seen.
Our read
This is a substantial reported bet on a field with plausible uses in robotics and AI training, not a product announcement dressed up as one. The combination of Nvidia and Unity is interesting because it connects computing infrastructure with interactive environments. The next test is whether Emulate can turn that backing into a working system people can evaluate.
What to watch
- Whether Emulate or its reported investors confirm the round and its terms.
- What Emulate builds, and when it offers a product or research results for scrutiny.
- Whether its world models demonstrate useful transfer to robotics or AI-agent training.
Discussion spark: Should investors back world-model companies this heavily before they have a public product, or should meaningful funding wait for evidence that the technology works?
Sources and evidence
- Nvidia and Unity join Emulate's $700 million round for DeepMind alumni building world models, Sifted reports – MLQ.ai (11 November 2026, 16:18 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.