Equity instead of a licence: reading the DeepMind–EVE deal
Neither exactly: in 2026 Google DeepMind took a minority stake in Fenris Creations, the studio behind EVE Online, and runs an offline version of the game on its own servers for research into long-horizon planning and memory. That is an ownership position, not a rental. It tells studios that the deepest simulations can attract capital rather than a fee — but it is not the deal most studios should wait for, because an equity arrangement is slow, exclusive by nature, and available to almost nobody.
The interesting part of the DeepMind deal is not that a lab wanted a game. It is the instrument. A licence rents access for a term; a stake buys a position in the studio that keeps making the thing. Choosing the second says the buyer expects to need not just this version of the world, but the people who can change it.
That is a reasonable thing to want. A training environment is never finished. A lab that trains against a world for two years will want new scenarios, different physics, a map that behaves unlike the shipped game. Owning part of the studio makes that conversation cheap. Licensing makes it a change request.
Why most studios should still want the licence
An equity deal is slow, it is exclusive in practice, and it is available to roughly nobody. It also changes what a studio is: once a lab holds a stake, the roadmap has a second audience. For a studio that wants a new revenue line rather than a new shareholder, the licence is the better instrument, and it has one property equity does not — you can do it more than once.
A game licensed to one lab for a defined use can be licensed to the next lab on its own isolated copy. Each of those is a separate contract with its own fees. The second one is cheaper to deliver than the first, and the studio keeps the difference.
What it signals for pricing
A lab willing to take an equity position in a studio has told the market what a deep simulation is worth to it, and the number is not small. Studios negotiating a licence should read that as a floor on seriousness rather than a comparable: it means the buyer has a budget and a thesis, not that your title is worth a stake.
The practical lesson is narrower and more useful. Depth of simulation, not release date, is what attracted that money. A back-catalogue title with real systems underneath it is closer to this market than a recent hit without them. Which games fit goes through what that means in practice.