Google DeepMind

Supply chain relationships

• Yaskawa Electric Corporation

  Outbound - Compute & Control - Yaskawa’s Agentic Robot System development integrates Google DeepMind’s Gemini Robotics ER 1.6 reasoning model

  Source

• Carolina Parada - Senior Director & Head of Robotics
• Demis Hassabis - Co-Founder & CEO
• Lila Ibrahim - Chief AI Readiness Officer

• Google DeepMind broadly releases Gemini 4 Argon

  2026-10-13 - Research Publication - Google reports an 87% score on its Embodied Reasoning benchmark for physical and spatial problems. In an internal test, Argon reconstructed a human cornhole video, generated simulation code and transferred a skill to an Apptronik Apollo humanoid that made seven consecutive shots. The robotics developer release includes examples for simulation reconstruction, robot control, physical understanding and research reproduction.

  Source
• Google DeepMind introduces Gemini Robotics On-Device for local robot control

  2025-06-24 - Research Publication - Google DeepMind introduced Gemini Robotics On-Device as an efficient VLA model optimized to run locally on robotic devices. The model supports general-purpose dexterity, fast task adaptation and SDK-based evaluation in MuJoCo with small demonstration sets. Local inference is central for practical robot autonomy in latency-sensitive or disconnected environments.

  Source
• Google DeepMind introduces RT-2 vision-language-action model

  2023-07-28 - Research Publication - RT-2 was presented as a robotics model trained on both web-scale vision-language data and robot action data. The release described improved generalization on emergent robotic skills compared with RT-1 and visual-language baselines. The event is sourced to Google DeepMind's official RT-2 research blog.

  Source

Events

• 2026-10-13 - Research Publication - Google DeepMind broadly releases Gemini 4 Argon

  Google reports an 87% score on its Embodied Reasoning benchmark for physical and spatial problems. In an internal test, Argon reconstructed a human cornhole video, generated simulation code and transferred a skill to an Apptronik Apollo humanoid that made seven consecutive shots. The robotics developer release includes examples for simulation reconstruction, robot control, physical understanding and research reproduction.

  Source
• 2026-07-15 - Partnership - Yaskawa integrates Gemini Robotics into MOTOMAN NEXT

  The system combines vision, path planning and force sensing to turn a general instruction into physical robot actions. Yaskawa describes automatic recovery from dropped objects and links to production-management systems for tasks such as parts replenishment. The system remains under development; specific applications and availability will be announced after preparations are complete.

  Source
• 2026-06-30 - Product Launch - Apptronik opens expanded Robot Park and unveils Apollo 2

  The nearly 90,000-square-foot Austin site anchors a network that also includes Google DeepMind, Mercedes-Benz and GXO locations. Apptronik described a mix of teleoperation and autonomous execution to gather data for Gemini Robotics models. Apollo 2 had already served as a training platform for more than a year; its data would inform the planned Apollo 3 commercial generation.

  Source
• 2026-06-09 - Market Signal - Google DeepMind selects 15 startups for European Robotics Accelerator

  Google DeepMind selected 15 early-stage European robotics startups for the inaugural three-month accelerator. The equity-free program provides technical mentorship, product guidance, Google AI infrastructure and access to Gemini Robotics models. The cohort spans industrial automation, construction, marine robotics, recycling, medical microrobotics, sensing, teleoperation and humanoid systems.

  Source
• 2026-06-09 - Award - Staer joins Google DeepMind Robotics Accelerator

  Staer was one of 15 early-stage European robotics companies selected for the cohort. The programme runs from June through September 2026 and includes dedicated Google mentorship. Participants receive hands-on access to Gemini Robotics models and Google Cloud infrastructure.

  Source
• 2026-04-11 - Demonstration - Innate runs 25 MARS robots at Y Combinator RoboHacks

  Innate and Iterate co-hosted the two-day robotics hackathon at Y Combinator's San Francisco office. More than 120 builders developed projects using at least 25 general-purpose MARS robots. NASA, Google DeepMind and other technology organizations supported the event's speakers, tracks and prizes.

  Source
• 2026-03-24 - Partnership - Agile Robots and Google DeepMind form strategic research partnership for industrial robotics

  Agile Robots and Google DeepMind say they have formed a strategic research partnership to advance next-generation AI robotics. The collaboration combines Google DeepMind's Gemini Robotics foundation models with Agile Robots' industrial robotics platform. The companies say the work will focus first on industrial and manufacturing use cases where reliability, adaptability, and scale matter.

  Source
• 2026-03-04 - Market Signal - Google DeepMind launches robotics accelerator for European startups

  Google DeepMind launched a three-month robotics accelerator for startups across Europe. The program is designed to help robotics startups bridge cutting-edge technology and market-ready ventures. Applications were scheduled to close on March 25, 2026, with the cohort tied to Google DeepMind robotics support and startup infrastructure.

  Source
• 2026-01-05 - Partnership - Boston Dynamics and Google DeepMind partner to bring Gemini Robotics to Atlas

  Joint research was planned at both companies, using a fleet of new Atlas robots. The work combines Gemini Robotics foundation models with visual-language-action approaches for humanoid behaviour. Initial applications target industrial tasks, beginning with automotive manufacturing.

  Source
• 2025-11-19 - Executive Hire - Google DeepMind hires former Boston Dynamics CTO Aaron Saunders for robotics hardware push

  Google DeepMind hired Aaron Saunders, former CTO of Boston Dynamics, as vice president of hardware engineering. The hire signaled a deeper push into robotics hardware and physical AI alongside Google DeepMind robotics-model work. Saunders brought senior humanoid and mobile-robotics experience from Boston Dynamics into Google DeepMind's robotics organization.

  Source
• 2025-09-25 - Demonstration - Apptronik Apollo demonstrates Gemini Robotics 1.5 skill transfer

  Tasks presented only to ALOHA 2 during training also worked on Apollo and the bi-arm Franka platform. Gemini Robotics 1.5 was available to selected partners, while Robotics-ER 1.5 was offered to developers through the Gemini API.

  Source
• 2025-06-24 - Research Publication - Google DeepMind introduces Gemini Robotics On-Device for local robot control

  Google DeepMind introduced Gemini Robotics On-Device as an efficient VLA model optimized to run locally on robotic devices. The model supports general-purpose dexterity, fast task adaptation and SDK-based evaluation in MuJoCo with small demonstration sets. Local inference is central for practical robot autonomy in latency-sensitive or disconnected environments.

  Source
• 2025-03-12 - Demonstration - Apptronik Apollo shown with Gemini Robotics

  The vision-language-action model was trained primarily on ALOHA 2 data and could also control a Franka-based bi-arm platform. A separate Robotics-ER model provided spatial reasoning that developers could connect to their existing low-level controllers.

  Source
• 2024-12-19 - Partnership - Apptronik partners with Google DeepMind on humanoid AI

  Apollo was described as a 5-foot-8-inch, 160-pound platform designed for physically demanding industrial work. The agreement followed Apptronik’s commercial engagements with GXO and Mercedes-Benz during the same year.

  Source
• 2024-10-16 - Demonstration - FrodoBots runs AI-versus-gamers Earth Rover Challenge at IROS 2024

  The competition ran October 16-17 during IROS 2024 in Abu Dhabi with five AI teams and five human gamers. Teams remotely controlled sidewalk robots on navigation missions in Abu Dhabi, Singapore, Taipei, Stockholm and other cities. FrodoBots collaborated with Yield Guild Games and researchers from Google DeepMind and academic institutions on the benchmark.

  Source
• 2023-07-28 - Research Publication - Google DeepMind introduces RT-2 vision-language-action model

  RT-2 was presented as a robotics model trained on both web-scale vision-language data and robot action data. The release described improved generalization on emergent robotic skills compared with RT-1 and visual-language baselines. The event is sourced to Google DeepMind's official RT-2 research blog.

  Source

Relationships

• Agile Robots

  Technology Partner - Agile Robots and Google DeepMind partnered to bring embodied AI intelligence to industrial robotics platforms.

• Apptronik

  Research Partner - Apptronik works with Google DeepMind on Apollo 2 data collection for Gemini Robotics model development.

• Yaskawa Electric Corporation

  Verified AI Integration - Yaskawa integrated Gemini Robotics ER 1.6 into its MOTOMAN NEXT agentic system.

• Enchanted Tools

  Software Platform - Enchanted Tools integrated Google Gemini AI into its Mirokaï humanoid robot platform for embodied intelligence.

• Boston Dynamics

  Atlas Deployment Partner - Boston Dynamics committed Atlas fleets for Google DeepMind deployment in 2026.

• Boston Dynamics

  Gemini Robotics Collaboration - Google DeepMind Gemini Robotics; public materials reference Boston Dynamics among partners.

• Agile Robots

  AI Robotics Collaboration - Google and Agile Robots are collaborating on AI robotics systems.

• FrodoBots

  Research Co Organizer - FrodoBots co-organized the IROS 2024 Earth Rover Challenge with researchers including Google DeepMind staff.

Comparable profiles

• Honda Robotics

  stack adjacency - Research Lab Layer

Shared workflows

• Stanford University

  workflow adjacency - Robot Learning Research

• Airion

  workflow adjacency - Robot Learning Research

• Simate

  workflow adjacency - Robot Learning Research

Context tags

• Robot model supply chain - Customer Environment
• General AI robot learning lab - Peer Group
• Robot learning research - Workflow
• Research lab - Stack Layer