Robotics

Today's Stories

1. 1

   Anthropic unveils Model Hardware Standard for physical AI

   Anthropic released Model Hardware Standard (MHS) as a research preview on Thursday. It is a framework to connect LLMs like Claude to physical devices, including manufacturing equipment and microscopes. MHS lets companies integrate AI into equipment in "hours or minutes," instead of "weeks, if not months" of custom work. It is model-agnostic and works with any LLM, not just Claude, and is built on Anthropic's Model Context Protocol (MCP), an open standard from 2024.

   Anthropic developed MHS with HHMI Janelia Research Campus. Partners include Genentech, Carnegie Mellon, QuEra, Universal Robots, Amazon Web Services, and others. Not all existing equipment can connect without a programming interface, so Anthropic is working with device manufacturers to build compatible products.

2. 2

   Hyundai's US EV plant runs 1,400 robots, 2:1 vs human

   Hyundai's new US EV factory operates 1,400 robots, giving it a robot-to-human ratio of 2:1—more than triple the industry average, the article reports. The figure highlights how far leading manufacturers are pushing automation in EV production, signaling that high robot density is becoming a competitive factor in the sector, the article suggests.

   The plant's reliance on robots may hint at future labor strategies and automation trends in the broader auto industry, though the article notes no explicit plans or projections.

3. 3

   Robust AI deploys 100th Carter robot, names new CEO

   Robust AI has deployed its 100th Carter robot and named former Fox Robotics CEO Marin Tchakarov as its new CEO. CTO Rodney Brooks says a robot only counts as deployed when customers rely on it in daily operations and the company has the support infrastructure to keep it running. He distinguishes working deployments from "humanoid deployment theater."

   Scaling from 100 robots to 1,000 requires more service and support staff, according to Brooks.

4. 4

   Meta Tests Robots in Data Centers, Tech Layoffs Near 140,000

   Meta is testing robots from Watney Robotics, Kinova, and ABB at its data centers in Altoona, Iowa, and New Albany, Ohio. The robots can swap networking cables, power-cycle servers, and reseat hardware. One Meta worker said a successful cable-swapping robot could replace up to 80% of some technicians' workloads. U.S. tech companies have cut nearly 140,000 jobs in 2026, according to a Financial Times analysis. While not all layoffs are AI-related, Meta's experiment shows automation moving from software to physical tasks, potentially reducing labor costs as its capital expenditures reach $31.1 billion in Q2 2026.

   Meta expects $130 billion to $145 billion in capital expenditures for 2026. Robots remain slower than humans and struggle with complicated cabling, so the technology isn't ready to replace technicians yet, but the potential is significant.

5. 5

   Anthropic launches Model Hardware Standard, cutting device integration time to hours

   Anthropic has introduced the Model Hardware Standard (MHS), a spec that gives AI agents a unified interface to control physical devices like robotic arms and lab equipment. It is being released as a research preview for select labs and manufacturers, with an open-source release planned later. The company says MHS cuts integration time for lab equipment from weeks or months down to hours or minutes. The standard builds on Anthropic's Model Context Protocol (MCP) for software and extends that approach to machines, potentially making it easier for AI agents to work in labs and factories.

   In a test at quantum computing company QuEra, a Claude-developed control program succeeded in 695 out of 700 attempts, a 99.3 percent success rate, running entirely on its own. However, Anthropic notes that Claude still struggles with physical cause and effect, and all three partner examples haven't been independently verified.

6. 6

   Brain Corp CTO on scaling 50,000+ robot fleet

   John Black, CTO of Brain Corp, discussed scaling robot fleets and building trust in public spaces on The Robot Report Podcast. Brain Corp has commercialized over 50,000 autonomous mobile robots (AMRs), which have logged more than 25 million hours of safe operation. Black's experience highlights the challenges and importance of deploying large fleets of robots in real-world environments, where safety and reliability are critical. His background as a TÜV Rheinland-certified functional safety engineer underscores the emphasis on safety standards.

   The podcast also covered news including XPeng's humanoid robot unit Dogotix raising $900M, Generalist adding $200M in funding, NVIDIA's Jetson Orin Nano 2 doubling inference performance for edge robotics, and Symbotic taking No. 1 in warehouse automation per Interact Analysis.