Okuma Corporation · Year ended March 2026

Okuma brings bearing diagnosis down to about three minutes

Okuma makes machine tools, including lathes and machining centers, that cut metal into parts. It supplies manufacturers in industries such as automotive and aerospace, and also develops the CNC controls that direct its machines’ movements.

A worker beside a lathe, with its rotating shaft, bearing and connected machine control brought into view
DATA VIEW

Okuma’s spindle-bearing diagnostic-time comparison

Published May 2026. About one hour is expressed as 60 minutes. The company compares a process including measurement, data handling and assessment, not total factory downtime.

Okuma’s spindle-bearing diagnostic-time comparison
Conventional diagnostic process~60 min
Diagnosis completed inside the CNC~3 min

Source: Okuma, AI machine-diagnosis development, page 2 (Japanese). The document does not provide detailed measurement conditions or the number of machines tested for this comparison.

Use information inside the control, without adding a sensor

From a bearing check to a maintenance decision

A schematic of the work described in the May 2026 release.

  1. Attach, measure, assess

    Attach, measure, assess

    A person attaches a sensor to the spindle-bearing area, uses an instrument and organizes the results. This is the conventional process described by Okuma.

  2. Use the CNC’s internal information

    Use the CNC’s internal information

    For the new turning-spindle diagnosis, the machine’s own control completes measurement, data handling and assessment without an additional sensor.

  3. Use the result in maintenance

    Use the result in maintenance

    Early detection can support planned parts replacement. The diagnosis does not itself perform a repair.

The cutaway and parts are explanatory drawings, not the dimensions or exact internal layout of a particular model. The sensorless step refers to turning spindles; milling-spindle-bearing diagnosis uses a vibration sensor.

Source: Okuma release, physical pages 1–2 (Japanese)

Test field data in the company’s own inspection process

CONCEPT MAP

Use machine-control information in maintenance

An editorial summary of relationships described by the company.

Machine tools and in-house CNCCombine knowledge of the machine and its controls
  1. Internal informationUsed for turning-spindle diagnosis
  2. Field dataTraining, validation and in-house inspection
  3. Maintenance decisionsFrom early detection to planned parts replacement

Keep model coverage separate from company-wide sales

THE VIEW AFTER THE NUMBERS

What the numbers suggest

Bringing spindle-bearing diagnosis down to about three minutes meant putting data handling and assessment inside the machine as well as measurement. Okuma’s in-house CNC controls the motion that cuts metal and helps check the machine’s condition.

Share this article

Share on X

This is not investment advice or a stock recommendation. We reorder company-published primary-source facts to make the shape of the business easier to see.

Primary sources used

AI machine-diagnosis development (Japanese)

Okuma Corporation · 2026-05-08

Pages used: p.1 / p.1-2 / p.2

Open source →

FY2025 Results Presentation (Japanese)

Okuma Corporation · 2026-05-19

Pages used: p.4

Open source →

Financial Results for the Year Ended March 2026 (Japanese)

Okuma Corporation · 2026-05-12

Pages used: p.1

Open source →

NEXT COMPANY

Azbil Corporation · Year ended March 2026

Look at the next company

Azbil’s test building alternates cool and warm air

Azbil combines sensors, valves and control systems to adjust airflow and when heating and cooling equipment operates. Alongside testing technology in its own building, it offers services that use customers’ operating data to propose improvements and assess the results. In the year ended March 2026, BA sales rose 5.1% despite lower new-building sales, and segment profit grew 18.6%. Control, renovation and service work continue after a building is completed.

See the company through the numbers →
~20 minutesCool/warm-air switching interval at No.103: checked October 2026