Skip to main content
Alphabet Folds Robotics Company Intrinsic Into Google to Scale Physical AI
Industry·CNBC / Intrinsic·1 min read

Alphabet Folds Robotics Company Intrinsic Into Google to Scale Physical AI

Alphabet is absorbing its robotics software company Intrinsic into Google, ending its five-and-a-half-year run as an independent 'Other Bets' subsidiary. Intrinsic spun out of Alphabet's moonshot factory X in 2021 with a mission to make industrial robotics programming accessible — its flagship product Flowstate lets users build robotic applications through a web-based platform without writing thousands of lines of code. Inside Google, Intrinsic will integrate with Gemini AI models, Google Cloud, and DeepMind to accelerate what the company calls 'physical AI' for manufacturing and logistics. The move follows Intrinsic's November 2025 joint venture with Foxconn to deploy AI robots in U.S. electronics assembly. CEO Wendy Tan White said joining Google lets them unlock physical AI for a much broader set of manufacturing businesses. The consolidation signals Google is getting serious about competing with Amazon and Tesla in robotics — and that the 'Android for robots' vision Intrinsic has been building may finally get the scale and AI infrastructure it needs.

GoogleAlphabetIntrinsicFlowstatePhysical AIManufacturingFoxconnDeepMind
Read original source

Related News

View All News →
Intrinsic Rejoins Google to Push Physical AI Into ManufacturingIndustry

Intrinsic Rejoins Google to Push Physical AI Into Manufacturing

Five years after spinning out from Alphabet, robotics startup Intrinsic is folding back into Google. The move pairs Intrinsic's Flowstate platform — a web-based environment for building robotic applications — with Google's Gemini AI models and cloud infrastructure. CEO Wendy Tan White stays on, and existing partnerships including Foxconn continue. Intrinsic acquired the Open Source Robotics Corp. (the company behind ROS) in 2022, which now also falls under Google. The deal follows Google DeepMind's recent collaboration with Boston Dynamics on Atlas, signaling Google is making a serious push into physical AI and industrial robotics.

·The Robot Report
XPENG to Build First Humanoid Robot Mass Production Facility in GuangzhouIndustry

XPENG to Build First Humanoid Robot Mass Production Facility in Guangzhou

Chinese EV maker XPENG announced plans to build what it calls the industry's first humanoid robot mass production facility, located at the Guangtang Sci-Tech Innovation City in Guangzhou's Tianhe District. The 110,000-square-meter park will house production lines for XPENG's IRON humanoid — a highly anthropomorphic robot featuring a human-like spine, biomimetic muscles, flexible skin, dexterous 22-DOF hands, all-solid-state batteries, and three Turing AI chips delivering 2,250 TOPS of compute. XPENG signed a strategic agreement with the Tianhe District government to develop the site, leveraging seven years of robotics R&D and its existing smart EV manufacturing expertise to tackle the humanoid production bottleneck.

·Gasgoo
ZaiNar Raises $100M to Build a Location Layer for Physical AI and RobotsIndustry

ZaiNar Raises $100M to Build a Location Layer for Physical AI and Robots

ZaiNar emerged from nine years of stealth with over $100 million in funding and a valuation above $1 billion. The Belmont, California company built a platform that turns existing wireless networks — 5G, Wi-Fi, private cellular — into a continuous sensing system that knows where everything is, without cameras, GPS, satellites, or extra hardware. The core breakthrough is sub-nanosecond time synchronization across radio signals, which translates to sub-meter positioning accuracy indoors, outdoors, and through walls. For robots, this solves a fundamental problem: current AI systems lack the real-time, centralized location data needed to move reliably in physical spaces. GPS fails indoors, cameras drift without heavy compute, and UWB beacons cost thousands per facility. ZaiNar claims its approach works with infrastructure already deployed worldwide, potentially giving home and industrial robots a universal positioning layer that works everywhere.

·The Robot Report