Project Genie

Overview

Project Genie is an experimental Google AI tool for generating and exploring interactive virtual worlds, initially launched through Google Labs for eligible U.S. Google AI Ultra subscribers. Over time, its positioning became more concrete: rather than only creating abstract worlds from prompts, Genie added grounding in Google Maps Street View, enabling users to turn real locations or text prompts into navigable 360° environments.

For AI Product Managers, Project Genie is notable because it demonstrates a compelling product pattern: combining foundation-model generation with proprietary geospatial data to create a new user experience layer. It is a useful case study in multimodal product design, data-grounded generation, and how platform assets like Street View can be transformed into differentiated AI features rather than remaining standalone data products.

Key Developments

• 2026-02-01: Google AI included Project Genie in a weekly roundup, describing it as a launch for dynamic world-building alongside other releases such as Gemini updates, AlphaGenome code, D4RT, and Gemini 3 Flash features.
• 2026-02-05: Google AI launched Project Genie as an experimental prototype for U.S.-based Google AI Ultra subscribers aged 18+, focused on building and exploring custom interactive worlds.
• 2026-02-14: Google DeepMind launched Project Genie on Google Labs, emphasizing immediate access for U.S. Google AI Ultra subscribers to create AI-powered projects.
• 2026-05-23: Google DeepMind expanded Project Genie with Google Maps Street View, allowing users to transform real U.S. locations into immersive, interactive worlds.
• 2026-07-11: Google AI launched Street View grounding in Project Genie, letting users generate and explore interactive 360° virtual environments from text prompts or real-world starting points using Google Maps Street View data.

Relevance to AI PMs

• Study grounded generative UX design: Project Genie shows how generative interfaces become more useful when anchored to trusted, structured datasets. PMs can apply this pattern to domains like commerce catalogs, enterprise knowledge, mapping, or CAD assets.
• Understand differentiation through proprietary data: Genie is a strong example of how exclusive platform data such as Street View can create defensible AI experiences. PMs should evaluate what unique datasets their organization can pair with models to produce hard-to-copy features.
• Learn rollout strategy for experimental AI products: The progression from prototype access via Google AI Ultra and Google Labs to richer grounded capabilities illustrates a staged release model. PMs can use similar sequencing for pricing, eligibility, safety validation, and feature maturation.

Related

• Google DeepMind: Core research and product organization associated with Project Genie’s launch and evolution.
• Google Labs: The experimental product surface where Project Genie was made available to early users.
• Google AI Ultra: Subscription tier that initially gated access to Project Genie for U.S. users.
• Google AI: Broader brand under which Project Genie appeared in launch roundups and product announcements.
• Gemini: Related Google AI model ecosystem; useful context for understanding Google’s broader multimodal product stack.
• Google Maps / Google Maps Street View / Street View: The geospatial and imagery foundation that enabled Genie’s grounded 360° world generation.
• AlphaGenome, D4RT, Gemini 3 Flash: Other Google AI launches mentioned alongside Project Genie, indicating its place within a broader portfolio of experimental and multimodal AI products.