English 
As semiconductor packaging starts getting more complex, the inspection part is becoming kind of a bigger bottleneck, than the manufacturing side, somehow. The challenge is not just finding defects. It is finding them early enough and without depending on a shrinking pool of highly skilled specialists.
Omron is expanding its collaboration with NVIDIA to tackle exactly that problem.
By combining Omron’s AOI and AXI inspection systems with NVIDIA Omniverse, the company is building a real-time digital twin environment for semiconductor substrate inspection. The goal is higher inspection accuracy and a process that relies less on individual expertise.
One major application is the visualization of component warping caused by heat during manufacturing. These distortions can lead to soldering defects but are often difficult to detect using conventional methods. Through physical simulations inside the digital twin, manufacturers can identify these issues earlier and adjust thermal profiles before defects emerge.
こちらもお読みください： 日本におけるAIインフラの整備が加速する中、Nvidiaは冷却と電力供給の面で三菱重工業に期待を寄せています
オムロン is also merging surface inspection data from AOI systems with internal structural data from AXI equipment to expose hidden relationships between visible defects and their root causes.
The partnership is also bringing AI agents into inspection workflows. Using エヌビディア’s Video Search and Summarization technology alongside visual language models and Graph RAG, engineers can ask questions in natural language and receive explanations based on similar historical cases and inspection data.
For an industry battling complexity and talent shortages at the same time, that shift matters.
日本語