← Cases
HungaryManufacturing2023 – ongoing

Flex

Deep-learning predictive maintenance for PCBA production

Flex reported ~33% higher operator productivity after introducing predictive maintenance

WorkingCompany's own claimOfficial ↗

Problem

Errors in circuit-board assembly could pass through several production stages before detection. Existing machine-data analysis explained failures after they occurred.

Approach

Flex introduced deep-learning models at its Zalaegerszeg factory in Hungary in 2023. The models analyzed historical machine data to predict and locate anomalies or maintenance needs before production problems developed.

Results

~33%increase in line-operator productivity reported by FlexCompany's own claim[1]
20-30%improvement in line utilization through reduced PCBA downtime, reported by FlexCompany's own claim[1]
25-30%increase in engineering-team productivity reported by FlexCompany's own claim[1]

Takeaway

Earlier fault detection addressed both machine downtime and the rework caused by defects discovered downstream.

Categories

Business function
Production and quality
Innovation type
Process innovation
Evidence
Company's own claim

Sources

Source summaries

[1]Flex2024

Flex described predictive maintenance introduced at its Hungarian PCBA factory in 2023. It reported gains in operator productivity, line utilization and engineering productivity.

Full details are in the original sources.

Added 1 Nov 2026 · Updated 1 Nov 2026
Open case page ↗