← Cases
TaiwanManufacturing2024 – ongoing

TSMC

TEM image quality self-check system

TSMC said an AI image check cut lab engineers' recheck time by 90%

WorkingCompany's own claim

Problem

Lab engineers at TSMC had to recheck transmission electron microscope (TEM) images by hand to confirm image quality.

Approach

In 2024 TSMC set up a TEM image quality self-check system. An AI model helps identify images, and digital tools help operators check image quality before engineers review them.

Results

90%reduction in lab engineers' image recheck time, per TSMC's 2024 sustainability report as quoted by TechNewsCompany's own claim[1]

Takeaway

A quality gate at the point where images are made left engineers with fewer images to recheck. The 90% figure comes from TSMC's sustainability reporting and has no stated baseline.

Categories

Innovation type
Process innovation
AI technology
Computer vision
Evidence
Company's own claim

Sources

Source summaries

[1]TechNews17 Dec 2025

TechNews article on TSMC's sustainability and AI efforts. It quotes the company's report that the TEM image quality self-check system, built in 2024, uses an AI model to help identify images and cut engineers' recheck time by 90%. It gives no accuracy or cost data.

Full details are in the original sources.

Added 2 Nov 2026 · Updated 2 Nov 2026
Open case page ↗