Google’s Veo 3 video model could head to video games.
DeepMind CEO Demis Hassabis hinted Tuesday evening that Veo 3, Google’s video-generating AI, might power playable game worlds.
The prompt came on X, where a user wrote, “Let me play a video game of my veo 3 videos already,” and asked, “playable world models wen?” Hassabis replied: 
now wouldn’t that be something.
Next morning, Logan Kilpatrick, lead product for Google AI Studio and Gemini API, teased with:
🤐🤐🤐🤐
No official announcements yet. A Google spokesperson told TechCrunch there’s nothing to share now.
Veo 3 remains in public preview. It generates video plus audio, from speech to soundtracks. It simulates physics for realistic movement but isn’t a full world model.
World models predict how environments evolve with player actions. Google aims to turn Gemini 2.5 Pro into such a brain-simulating model. DeepMind already launched Genie 2, which can create endless playable worlds. Google also formed a team for AI that simulates the real world.
Veo 3 looks more suited for in-game cinematics, like cutscenes or trailers. It’s passive and needs to evolve into an interactive simulator to work for gameplay.
Google might link Veo with Genie later to tackle real-time, consistent, and controllable game worlds—the real challenge in game AI.
Google’s entering a hot space. Microsoft, Scenario, Runway, Pika, and OpenAI’s upcoming Sora model are all chasing game-ready AI worlds.
Watch Google’s next moves. Its deep pockets and reach could change how AI and game worlds merge.
