process-control
(7)
The Machines That Teach AI Chips to Behave
Onto Innovation does not design the AI chip. It builds the eyes, rulers and feedback loops that help factories make those chips without wasting a fortune.
Applied Materials
The company behind the machines that build almost every chip you own - and the atomic-scale materials engineering that keeps Moore's Law alive.
Emerson
Emerson spent a century making the machinery of modern life. Now it is assembling something harder to see: the digital nervous system for factories, power grids and process plants.
Honeywell Technologies
After spinning off aerospace and advanced materials, Honeywell is smaller, sharper and focused on one difficult job: teaching the physical world to run with less waste, fewer surprises and more autonomy.
PDF Solutions
PDF Solutions is a Santa Clara semiconductor analytics company that turns the firehose of chip-manufacturing and test data into decisions. Its Exensio platform connects fabs, foundries, OSATs and fabless designers, normalizing data across silos so customers can ramp yield faster, catch defects earlier and run AI/ML models on the factory floor. Founded in 1991 and public since 2001 (Nasdaq: PDFS), it serves 130+ semiconductor companies including TSMC, Intel, Qualcomm, Analog Devices and DENSO.
Relimetrics, Inc.
Relimetrics is an AI software company building ReliVision, a no-code visual inspection platform that lets manufacturers digitize quality assurance on the factory floor. Hardware-agnostic and self-installable, it puts model training, retraining and defect detection in the hands of plant engineers instead of data scientists - and its accuracy is unusual enough that reinsurer Munich Re will underwrite it. Founded in 2015 with offices in Silicon Valley and Berlin, the company counts HPE, Lockheed Martin, Foxconn and Siemens Gamesa among its users.
Gauss Labs
Gauss Labs is a Palo Alto-based industrial AI company building software that reads the firehose of data coming off semiconductor fabs and predicts what's about to go wrong. Spun out with $55M in seed capital from SK hynix in 2020, its Panoptes Virtual Metrology and Image Metrology products are deployed inside SK hynix's high-volume memory fabs, where they have measured tens of millions of wafers and cut process variability by double digits.