SAKI has developed an automated X-ray inspection system for the increasingly large circuit boards used in AI servers. Its 3Xi-ZS1EX can handle boards up to 1,000 by 690 mm and combines 3D imaging with AI-based image analysis to spot internal defects.
Watch Desk analysis
What happened
The system is designed to inspect complex printed circuit board assemblies, where warping, tall components and hidden defects can make quality checks trickier. SAKI says it compensates for board-surface warping and uses an AI-based algorithm to reduce image noise, helping detect small voids and other defects.
The report in ASSEMBLY describes the 3Xi-ZS1EX as applying inspection parameters automatically to provide consistent results across complex assemblies. The system’s stated maximum board size is 1,000 by 690 millimetres.
Why it matters
AI servers depend on large, complicated circuit boards, and a defect buried inside one is not much easier to find just because the board has a shiny new job title. Automated inspection could help manufacturers examine those assemblies without trading away accuracy or production speed.
The practical interest is in the combination: 3D X-ray imaging, compensation for warping and AI-based image analysis aimed at revealing small defects. SAKI describes what the system is designed to do; the supplied report does not provide independent performance results or comparisons with other inspection systems.
Our read
This is a useful piece of AI infrastructure news at the unglamorous end of the supply chain. The model is not writing poems or booking meetings; it is helping inspect the boards that make data-centre hardware work. If the system delivers the consistency and defect detection SAKI describes, that is a meaningful manufacturing application, not AI sprinkled on a product brochure.
For operators, the useful questions are how it performs on real production lines and whether it can keep pace with the growing size and complexity of AI-server boards. A machine that spots more defects but slows the line to a crawl would be a rather expensive magnifying glass.
What to watch
- Independent performance results for defect detection and inspection speed.
- How the system handles warped boards and tall components in production.
- Whether manufacturers adopt it for the larger assemblies used in AI servers.
Discussion spark: When inspecting increasingly complex AI-server boards, should manufacturers prioritise catching more hidden defects or keeping inspection fast enough for production?
Sources and evidence
- Source update (2 November 2026, 12:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.