TheTechno Optimist.

Project SuncatcherGoogle DeepMind—Google announces TPU 8t and TPU 8i at Cloud Next, splitting training and inferenceGemini RoboticsGoogle DeepMind—Google DeepMind releases Gemini Robotics 2, enabling full-body humanoid control with 22-DoF handsAntigravityGoogle DeepMind—Google announces Antigravity 2.0 with multi-agent orchestration and voice commands at I/O 2026Tensor Processing Unit (TPU)Google DeepMind—Google says Ironwood, its seventh-generation TPU, will be generally available on Google Cloud in the coming weeksGoogle DeepMind—Google DeepMind expands Synth-ID tool globally to identify AI-generated contentVeoGoogle DeepMind—Google releases Veo 3.1 with richer audio and more editing control in FlowGeminiGoogle DeepMind—Google and SUNY launch AI research platform for university researchersProject SuncatcherGoogle DeepMind—Google reveals Project Suncatcher: orbital AI data centers powered by solar, two prototype satellites planned for early 2027Genie world modelsGoogle DeepMind—Project Genie adds Street View, letting users generate explorable versions of real placesGeminiGoogle DeepMind—Google releases Gemini 3 ProTensor Processing Unit (TPU)Google DeepMind—Google unveils eighth-generation TPUs, split into TPU 8t for training and TPU 8i for inferenceGeminiGoogle DeepMind—Google introduces Gemini agent platform for universal work across devicesAlphaEvolveGoogle DeepMind—Google DeepMind's AlphaEvolve becomes generally available, finds 23% speed improvement in Gemini kernelGemma 4Google DeepMind—Google releases Gemma 4, open-weight models from 2B to 31B, with Apache 2.0 licenseGenie world modelsGoogle DeepMind—Project Genie opens Genie 3 world generation to Google AI Ultra subscribersAlphaFoldGoogle DeepMind—Google DeepMind folds its AlphaFold team into other projects after John Jumper leaves for AnthropicVeoGoogle DeepMind—Google unveils Gemini Omni at I/O, pairing Gemini's reasoning with Veo video generationGeminiGoogle DeepMind—Google launches the Gemini 3.5 family at I/O with 3.5 Flash; 3.5 Pro promised for the following monthGeminiGoogle DeepMind—Google releases Gemini 3.6 Flash, 3.5 Flash-Lite and 3.5 Flash Cyber

Gemini Robotics

by Google DeepMindGBUS

Google DeepMind's vision-language-action model for robotic control. Gemini Robotics 2 enables whole-body movement and complex manipulation.

DeployStage 4 of 5

Gemini Robotics 2 released Aug 2026. Full-body control, 22-DoF hands.

Updated Aug 2026Checked 10 Oct0 updates this week

Milestones

Next · Deployment on Apptronik Apollo 2 and other humanoids
  1. Gemini Robotics 1 demonstrates multimodal robotic control2024Complete.
  2. Gemini Robotics 1.5 adds motion transfer for embodiment adaptationSep 2025Complete.
  3. Gemini Robotics 2 released with full-body controlAug 2026Complete.
  4. 22-DoF hand control on Apptronik Apollo 2Aug 2026Complete.

Most important, last 3 months

Upcoming

  1. Dec 2026Deployment on Apptronik Apollo 2 and other humanoidsin about 7 weeks (next)
  2. 2027Continued robotics applications and refinement
Primer