Intel's Seeq AI Cuts Fab Anomaly Detection Time by 30%
The eight-month deployment also saved about $7.5 million a year in a single use case and cut technicians' workload by 15-20%, according to etnews.com.
Intel has reported that an eight-month deployment of the agentic AI platform Seeq across its semiconductor fab utility equipment has cut anomaly detection time by about 30%, according to etnews.com. The company also attributed roughly $7.5 million in annual cost savings to a single use case.
Seeq operates as an orchestrator over multiple AI agents. It is connected to about 70 systems across Intel's global sites and integrates 65-70 million time-series data tags. Around 1,000 Intel engineers use the software each week, the report said.
The rollout covers critical utility equipment, including ultrapure water, chillers, specialty gases and chemical supply systems. According to the figures, automatically generated work orders reduced technicians' workload by 15-20% and saved more than 120,000 labor hours per year. Final approval of drafted work orders remains with engineers.
Energy efficiency at the chilled water plant improved by about 15%.
From self-service analytics to agentic AI
About three years ago, Intel began looking for a self-service analytics platform to develop and deploy large-scale analytics, which started the collaboration with Seeq. As Intel expanded its AI adoption, the partnership grew toward agentic AI, according to the report.
Quick answers
How much did Intel save with Seeq?
About $7.5 million per year from a single use case, according to etnews.com.
How many Intel engineers use Seeq?
About 1,000 engineers use the software each week.
What does Seeq connect to at Intel?
About 70 systems across Intel's global sites, integrating 65-70 million time-series data tags.