Google DeepMind launches Genie 4 world model with native audio, real-time interaction

The significance here is less the model itself than the distribution path: DeepMind is routing a world model through consumer subscriptions, an API and a studio tool at once, which positions generative environments as a platform layer rather than a research demo. Native audio and real-time dialogue matter because they close the gap between simulated worlds and something that feels operable, which is where adjacent RWA and on-chain simulation narratives tend to draw their claims. The trusted-tester phase is the constraint worth watching — how quickly access widens, and whether early partners build on the API or stay in Genie, is the open question.