Mitsubishi Electric just connected a humanoid robot to a CNC machine with natural language
Mitsubishi Electric announced a system that lets a humanoid robot work a CNC machine tool using only natural language instructions from a human supervisor ... no custom programming per robot model, no bespoke communication protocols between the CNC controller and the robot.
Just tell the system what needs to be done, and the AI coordinates the rest.
The company is calling it the industry's first machining system to enable seamless collaboration between humanoid robots and CNC machine tools through physical AI.
The system was first announced on October 21, 2026, and is being publicly demonstrated this week at JIMTOF 2026 in Tokyo.
What the system actually is
There are 3 integrated components.
- The first is an AI Machining Task Planner that accepts natural-language instructions from an operator, breaks them into step-by-step machining routines, and schedules tasks across both the CNC machine and the humanoid robot. No per-robot programming is required, and the planner handles the sequencing.
- The second is a physical AI-enabled CNC controller, based on Mitsubishi's M800/M80 series, with the AI embedded at the kernel level. The controller exchanges operational data with the robot in natural language and coordinates execution in real time.
- The third is the humanoid robot itself. Mitsubishi has not named a specific robot model or manufacturing partner. The system is designed to be robot-model agnostic, which is part of the point: you should be able to swap in a different robot without rebuilding all the integration work from scratch.
The AI backbone across all three components is Mitsubishi's Maisart platform.
What the robot does at the machine
The humanoid takes over the tasks that skilled machine-tool operators currently perform manually:
- workpiece setup
- equipment cleaning
- removal and transport of finished products
- inspection
- troubleshooting
These are the operator-side tasks that surround the CNC cutting process itself. The machine does the cutting; the robot does everything the human used to do around it.
The stated goal is fully unmanned machining, where a single natural-language instruction from a human initiates an entire production sequence that runs without further supervision.
The integration problem
The technical barrier Mitsubishi is addressing is not widely understood outside the manufacturing sector, but it is real.
Integrating any robot with CNC machinery today requires writing robot-specific communication functions and control sequences, essentially custom software that bridges the robot's control system to the CNC controller's control system. Every time you change the robot model, you redo that work. Every time the CNC software updates, you may break the integration.
The cost and complexity of that work is one of the main reasons CNC shops have been slow to adopt automation beyond simple, fixed-function loaders.
Mitsubishi's approach replaces that custom-integration layer with a natural-language interface and a robot-model-agnostic planner. The CNC and the robot talk to each other through the AI layer, not through bespoke machine-to-machine protocols. In principle, this means a shop that installs the system today can migrate to a different humanoid robot in three years without rebuilding the entire integration stack.
The timeline: 2030
Target practical deployment is 2030. That is a long runway.
No performance data has been disclosed, and there are no published efficiency numbers, cycle time improvements, or uptime figures from real production environments.
What we're seeing right now is a demonstrated prototype being shown at a trade fair, with a commercialization target four years out.
Essentially Mitsubishi Electric, a 130-year-old industrial conglomerate with deep relationships across Japanese and global manufacturing, is saying that it is looking at humanoid integration with its CNC controller business that is already installed in factories worldwide. It's basically future-proofing itself.
The 2030 target also puts this squarely in the window when humanoid robots should reach meaningful deployment scale in manufacturing globally.
Sources
Written by
I tell stories with data and build startups. I've raised angel and VC funds, led software companies and startups, and chronicled the rise of the mobile economy. Currently I write for Forbes, host several podcasts, and am starting up HumanoidMegaHub. I was TUNE's mobile economist and Singular's VP of Insights, built the research team at VentureBeat, and managed teams creating software for partners like Intel and Disney. In past lives I built websites, e-learning platforms, and mobile apps, and, a long time ago, was named to Folio's top 100 of the media industry's "most innovative entrepreneurs and market shaker-uppers.”