Google announced that Project Genie can now simulate real-world streets by integrating with Google Maps Street View during the Google I/O 2026 conference [1].
This development marks a shift in generative AI by grounding synthetic environments in actual geographic data. By combining a world model with existing mapping infrastructure, Google allows users to explore immersive simulations of real places rather than purely imaginary landscapes.
The announcement took place on May 19, 2026 [1], in Mountain View, California. The project, developed in part by the DeepMind team, enables AI-generated worlds to be anchored to specific real-world locations [1]. This creates a hybrid environment where the AI can render interactive spaces based on the visual and spatial data captured by Street View.
Google said the integration is intended to support a variety of high-impact applications. These include the development of robotics, where AI agents can be trained in realistic digital twins of city streets, and the gaming industry, which can now create levels based on actual geography [1]. Travel applications are also a primary focus, as the technology allows for more immersive previews of destinations.
Project Genie functions as a world model that generates interactive environments. By linking this capability to the vast archive of Street View, the system can simulate the physics and layout of real streets [1]. This allows for a level of interaction that traditional 360-degree imagery cannot provide; users can move through and interact with the simulated space in real time.
The company said this grounding in real-world data helps ensure that the generated simulations remain accurate to the physical locations they represent [1].
“Project Genie can now simulate real-world streets by integrating with Google Maps Street View”
The integration of Project Genie with Street View represents a move toward 'spatial intelligence,' where AI does not just generate images but understands and simulates the 3D physics of the real world. For robotics and autonomous systems, this provides a scalable way to create training environments that mirror reality without the risk of physical crashes. For consumers, it transforms Google Maps from a static reference tool into an interactive simulation engine.
