AIToday
RoboticsTop Companies' AI MovesTop Companies AIPublished: Nov 6, 2026, 06:30 JST

Mitsubishi Electric shows physical AI at CEATEC 2026

Send on LINE
Mitsubishi Electric shows physical AI at CEATEC 2026

3 Key Points

  1. What happened

    At CEATEC 2026, Mitsubishi Electric demonstrated a screw-tightening AI trained on about 2,000 images per screw hole that needs no fixtures or specialist knowledge, and a task-general sound separation AI (TUSS) that won the CEATEC AWARD Minister of Internal Affairs and Communications Prize, with fiscal 2027 targeted for physical AI implementation and product application.

  2. Why it matters

    The demonstrations show Mitsubishi Electric turning hands-on production know-how into deployable AI, which could make factory automation more flexible and less dependent on specialist setup, though broad rollout appears to hinge on further integration work.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

Mitsubishi Electric's exhibit sits within a broader push it frames around three business pillars: smart energy, automation, and defense & space. These are areas where a stoppage can have major social impact, so the company argues that physical AI must absorb unexpected events and keep operating, rather than merely follow pre-set rules. Its approach rests on two elements: an action model that reproduces skilled workers' judgment, and integration with existing sensors and equipment.

The company also used the event to connect that factory-floor work to its digital platform Serendie. In a stage discussion with Nozomi Networks CEO Andrea Carcano, Mitsubishi Electric executive officer Nobuo Asahi described how Serendie is meant to deepen customer relationships after hardware delivery by turning usage data into solutions. Carcano noted that cyber risk has moved from being one issue among many to what he described as the biggest barrier to AI adoption, with 40% of executives citing cybersecurity as their most important challenge in one survey.

Alongside the screw-tightening and sound-separation demonstrations, Mitsubishi Electric outlined an unmanned factory concept it calls an intelligent autonomous operation factory, organized around a six-level autonomy maturity scale. It also described next-generation data centers targeting PUE1.1, with AI proposing cooling settings and an eventual goal of autonomous control, and highlighted a collaboration with Elephantech, including an additional 4 billion yen investment in 2026.

FAQ
What does the screw-tightening AI need to work?
It is trained on about 2,000 images per screw hole, so it requires neither fixing jigs nor specialist knowledge, and in a demo it followed and tightened a screw even when the receiving part was deliberately moved.
When will the sound separation AI be used in products?
Mitsubishi Electric aims to implement TUSS in physical AI and apply it to products during fiscal 2027.
How does Mitsubishi Electric plan to reach PUE1.1 in data centers?
One approach is AI that suggests optimal air-conditioner settings to reduce airflow and power consumption, with autonomous control as the eventual goal, alongside partnerships with Foxconn, ZutaCore, and NVIDIA.
Top Companies AIRead Original Article
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleJDSC launches Logi-one Engine for cross-logistics AI optimization