NVIDIA's GEAR Lab unveiled ENPIRE on June 17, 2026, a robot-learning framework in which AI agents run experiments on real robot hardware by themselves. Frontier coding agents like Codex and Claude Code learned precise manipulation with no human intervention, inserting a GPU into a motherboard at a 99% success rate. ASAP summarizes this result, released with Carnegie Mellon and UC Berkeley.

What was unveiled

ENPIRE is an agentic robotics framework that lets coding agents run research directly on real robots. NVIDIA GEAR Lab released it with Carnegie Mellon and UC Berkeley on June 17, 2026, hitting a 99% success rate on high-precision tasks like GPU insertion, pin organization, and zip-tie cutting. It was tested with Codex (GPT-5.5), Claude Code, and Kimi Code variants across fleets of 1, 4, and 8 robots.

How it teaches itself

The core of ENPIRE is that the agent closes the learning loop itself. A Policy Improvement module generates, revises, and tests policy code using rewards, videos, execution traces, and failure analysis. Critically, agents read research papers online and propose algorithmic hypotheses on their own, meaning the agent takes on the researcher's role rather than following human-designed training.

Why it matters

ENPIRE is an attempt to close the research loop for physical AI. Robot learning has required humans to design rewards and code, but ENPIRE hands hypothesis, experiment, and verification to the agent to run autonomously. NVIDIA said it plans to open-source the framework, opening a path for outside companies and labs to set up self-running robotic labs.

Wrap-up

ENPIRE shows that coding agents can autonomously do research in the physical world, not just the digital one. The 99% GPU-insertion success rate shown on June 17, 2026 is not a mere demo but points to self-running robotic labs that repeat research and experiments without human intervention. The direction is clear: agents are taking on hypotheses, not just execution.

Source: NVIDIA GEAR Lab ENPIRE release (2026-06-17, NVIDIA, Carnegie Mellon, UC Berkeley); reporting (Tom's Hardware, NVIDIA Blog, TechTimes).