Google Folds Intrinsic into Core for Physical AI

🔑 Enhanced Key Takeaways

• •Intrinsic acquired Robots-as-a-Service company Vicarious and Open Robotics subsidiaries in 2022 to bolster its AI robotics capabilities.[3]
• •In October 2025, Intrinsic partnered with Foxconn on a joint venture for general-purpose intelligent robots targeting full factory automation in electronics manufacturing.[6]
• •Intrinsic released its Intrinsic Vision AI model (IVM) in late 2025, enabling high-precision tasks without costly depth sensors via AI perception.[1][6]
• •In January 2026 at CES, Google collaborated with Boston Dynamics to integrate Gemini Robotics model into the Atlas humanoid.[2]

🛠️ Technical Deep Dive

• •Intrinsic's platform includes Flowstate, a web-based development environment and simulation engine with reusable 'skills' (building blocks of robotic behaviors) for app creation across diverse hardware.[1][3][5]
• •Intrinsic Vision AI model (IVM) supports high-precision perception and grasping without expensive depth sensors, used in collaborations like NVIDIA and Foxconn for AI-driven hardware execution.[1]
• •The platform acts as a unifying software layer compatible with various robot arms, sensors, cameras, and AI models, allowing learning via demonstration and simulation instead of manual coding.[3][4]

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline