Google Teases Gemini Robotics ER-1.6

Beyond the Smartphone: Google Teases “Gemini Robotics ER-1.6” Ahead of I/O 2026

Google is making it clear that its vision for artificial intelligence extends far beyond the screens of our smartphones and laptops. Just one day before the official kickoff of Google I/O 2026, internal leaks, session schedules, and industry whispers have revealed that Google is preparing a massive push into embodied AI with a new robotics model codenamed Gemini Robotics ER-1.6.

While last week’s Android Show focused entirely on mobile breakthroughs like Gemini Intelligence, Googlebooks, and smart glasses, the main I/O stage is set to showcase how Gemini will act as the physical brain for the next generation of automation hardware.

What is Gemini Robotics ER-1.6?

The ER-1.6 model represents Google’s latest breakthrough in embodied AI—the practice of embedding massive multimodal language models into physical robotic systems. Rather than relying on rigid, pre-programmed code to tell a robot how to move its joints, Gemini Robotics allows machines to understand the physical world using natural language, computer vision, and real-time spatial reasoning.

Key Expected Core Capabilities:

Visual-Spatial Understanding: Seamlessly parsing messy, real-world environments (like a warehouse, a messy kitchen, or a laboratory) to identify objects, judge distances, and calculate safe execution paths.
Natural Language Instruction: Allowing human operators to issue complex, un-programmed commands like, “Sort the damaged tax documents from the intact ones and place them in the secure bin,” which the robot can break down into autonomous physical actions.
Advanced Real-World Generalization: Moving past repetitive assembly-line tasks. The ER-1.6 framework is built to adapt on the fly when obstacles change, an object slips, or a tool isn’t where it belongs.

The Strategic Shift: Google’s Long-Term Hardware Play

For investors and developers alike, the introduction of a dedicated robotics model layer signals that Alphabet is looking to establish Gemini as the fundamental operating layer for physical automation.

   [ Cloud / Desktop ]       ──>  Gemini 4.0 & Workspace Studio
   [ Mobile Ecosystem ]      ──>  Gemini Intelligence & Aluminium OS
   [ Physical Automation ]   ──>  Gemini Robotics ER-1.6

By presenting an interconnected ecosystem that spans from on-device smartphone chips up to industrial hardware machinery, Google is positioning its AI architecture to compete directly against specialized robotics frameworks being built by rivals like OpenAI and physical-first automation companies.

What Else is on the Horizon for the Keynote?

While Gemini Robotics is capturing the attention of the automation sector, Sundar Pichai’s main keynote address today is expected to reveal several major software and media model upgrades:

Model / Platform	What It Brings	Expected Impact
Gemini 4.0 Upgrade	The next core model iteration, aiming to match recent logic benchmarks.	Enhanced coding and complex reasoning pipelines.
Gemini Omni	A standalone, multimodal video generation and editing engine built directly into chat.	High prompt fidelity, built-in background music, and fast rendering.
Gemini Spark	Google’s always-on, proactive assistant framework.	Automates email sorting, meeting preparation briefs, and data aggregation.
Genie 3 & Gemma 4	Advanced interactive world simulators and lightweight, open-source model updates.	Gives developers deep flexibility for edge-compute workflows.