.png)
Intel
Manufacturing · International
Up to $2M scrap savings/year
Intel detects production defects inline with AI, catching problems during manufacturing, not after
Edge AI inspects each wafer in under 90 seconds. Up to $2M scrap avoidance per year.
Friction
In assembly and test factories, offline inspection meant defects were discovered late. One process problem could damage multiple wafer lots before detection.
Breakthrough
Inline inspection at micron level in the production line, with edge AI that can stop tooling. Detect the deviation during processing, not after.
Impact
Up to $2M/year scrap avoidance. Higher quality. Engineers freed from repetitive offline inspection.
Unlock the full analysis with breakthrough, impact, what made it smart and its technical approach bellow!
Problem
In assembly and test factories, offline inspection meant defects were discovered late. One process problem could damage multiple wafer lots before being caught.
What made it smart
Moving inspection from offline to inline, in the production flow, not after it. Edge AI processes approximately 20GB of data per wafer within 90 seconds and can trigger an immediate process stop.
Technical approach
High-resolution cameras make continuous images. ML model analyses images at the edge and triggers alarm or stop. Architecture: edge compute and HPC to handle approximately 20GB per wafer within 90 seconds. Detection capability in the order of 6 to 12 micrometres.
Strategic lesson
The value of quality AI is proportional to how early in the process it intervenes. Detect at source, not at outcome.
Reflection question
At what point in your process do you find out something went wrong? And how much does everything that happened before that point cost you?