Google’s Genie world model can now simulate real streets with Street View

2026-05-19 · Source: TechCrunch · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Intermediate, short

Summary

Google DeepMind has integrated Street View data with Project Genie, its general-purpose world model, enabling the generation of diverse, interactive environments based on real-world locations. This new feature, announced at Google I/O 2026, allows users to simulate specific street scenes and modify conditions like weather, offering immersive experiences for various applications. Google has amassed over 280 billion Street View images across 110 countries over two decades, providing a rich dataset for Genie. The latest iteration, Genie 3, released for research preview last August and accessible to Google AI Ultra subscribers since January, is already used in Waymo's simulators to train self-driving cars for rare events. While current simulations offer "video game quality" rather than photorealism and lack physics awareness, the system demonstrates strong spatial continuity, accurately remembering and simulating environments during 360-degree turns. This experimental integration is rolling out to U.S. Ultra users today, with global access following in weeks.

Key takeaway

For research scientists and robotics engineers developing autonomous systems, this integration of Street View with Project Genie offers a powerful new tool for creating highly specific, real-world-anchored simulation environments. You can now train agents on rare events or test deployments in diverse geographical and weather conditions, significantly accelerating development and validation cycles. Be aware that current simulations are not yet photorealistic or physics-aware, requiring careful consideration of their applicability for high-fidelity testing.

Key insights

Google DeepMind integrates Street View with Project Genie to create interactive, real-world-anchored simulations for diverse applications.

Principles

• Real-world data enhances AI simulation fidelity.
• Spatial continuity is key for immersive virtual environments.

Method

Project Genie leverages Google Street View's 280 billion images to generate interactive, modifiable simulations of real-world locations, enabling scenario testing for robotics and user exploration.

In practice

• Simulate rare events for robot training.
• Preview locations under different environmental conditions.

Topics

• Project Genie
• Google Street View
• World Models
• Robotics Simulation
• Autonomous Driving

Best for: Computer Vision Engineer, Research Scientist, AI Scientist, Robotics Engineer, Tech Journalist

Related on AIssential

• Google pairs its Genie world model with Street View to create explorable AI worlds based on real places — Google's Genie model uses Street View data to create interactive AI worlds for agent training and simulation.
• Simulate real-world places with Project Genie and Street View — Project Genie now integrates Street View to create imaginative, interactive virtual worlds grounded in real-world locations.
• ‘World models’ are AI’s latest sensation: what are they and what can they do? — World models create interactive, physics-aware virtual environments for training AI, overcoming generative AI's physical world limitations.
• TAI #190: Genie 3 World Model Goes Public — Genie 3 enables real-time interactive world generation, advancing both game development prototyping and embodied AI research.
• NVIDIA Research Advances Robotics From Simulation to the Real World — Simulation-to-real transfer is foundational for achieving generalizable, reliable embodied autonomy in robotics.
• World Models in Plain English — World models aim to spatially compress and predict real-world environments, moving beyond text-based learning.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by TechCrunch.