At a glance
| Booth | P5720 |
| Country | US |
| Website | gausslabs.ai |
Company profile
Gauss Labs is an industrial AI startup founded in August 2020, with its business focused on applying artificial intelligence and machine learning to manufacturing. The company describes its team as machine-learning experts with deep knowledge of the manufacturing domain, combining AI expertise with manufacturing-process understanding. Gauss Labs provides an industrial AI solution called Panoptes and describes Panoptes as a software-defined metrology solution for manufacturing environments. Its anchor product, Panoptes Virtual Metrology (VM), uses existing fab data and AI technology to predict process outcomes and is intended to help customers manufacture products better, at lower cost, in greater volume, and faster through multiple use cases. The company’s broader manufacturing-data-intelligence approach uses analytics and insights derived from large volumes of machine-generated data that it describes as exceeding normal human capability and comprehension. Gauss Labs lists offices in Palo Alto in the United States and Seoul in South Korea, giving the industrial AI company operating locations in both countries.
Exhibits
Gauss Labs’ AI-powered manufacturing data-intelligence products analyze large volumes of machine-generated data and derive analytics and insights intended to support manufacturing decisions beyond what can be handled manually. Its models are data-driven and are designed to adapt to dynamic changes in data, allowing the software to respond as manufacturing data distributions and operating conditions change. The software automatically trains and manages models, reducing the need for users to have in-depth knowledge of machine-learning algorithms and making model creation and management part of the automated workflow. The solutions are built to scale to high-volume manufacturing and incorporate domain knowledge into models, supporting model operation at fab scale rather than only small experimental deployments. Gauss Labs deploys its solutions in the customer’s own environment to satisfy security-compliance requirements and avoid intellectual-property issues associated with moving manufacturing data outside that environment. Panoptes Virtual Metrology (VM) predicts process outcomes from process-state data and wafer-state data, and with minimal effort it supports model training and updating, fab-scale model operation, and adoption of domain knowledge. Panoptes VM sends all-wafer data to customer systems to support use cases such as process control, equipment maintenance, and yield management. The product is also described as predicting the process outcomes of all wafers in real time by leveraging fab data, with stated use-case effects including yield improvement, cost savings, and cycle-time reduction. Gauss Labs explains the need for virtual metrology by noting that traditional physical metrology can be constrained by equipment and sensor cost, limited fab space, time, and low sampling, which can leave manufacturers with only sampled data and delayed process visibility. Panoptes VM provides predicted measurement data for all wafers and uses existing fab data to predict process outcomes, with film thickness after a deposition process given as one example of a predicted outcome. Gauss Labs describes Panoptes VM as a Purpose-Specific Automatic Learning Machine (PSALM): it is specifically developed for virtual metrology in manufacturing with proprietary algorithms, includes automatic functions for model creation and management, and can operate hundreds of thousands of models in real time in high-volume manufacturing environments. The VM platform is presented around accuracy, usability, and scalability, with an innovative model architecture, fab-scale operation validated in real high-volume manufacturing fabs, and an interface intended for process engineers without coding or machine-learning knowledge. For process control, Panoptes VM can enhance the control system by helping it adjust process recipes at a single-wafer level in real time. For process monitoring, the platform helps engineers detect anomalies and identify root causes faster with the stated aim of preventing excursions. For metrology optimization, Panoptes VM analyzes process stability so that sampling can be reduced or expanded according to the observed stability of the process. For equipment maintenance, the platform supports engineers in detecting and predicting equipment anomalies and mismatching, extending virtual-metrology output into maintenance-related decisions. A customer reference for a #1 Global Memory Chip Manufacturer reports 5 sites in Korea and China, deployment in December 2022, publication to an APC system, and a 29% improvement in process variability. The same product page links an SK hynix deployment announcement stating that Gauss Labs’ AI-based virtual-metrology solution was deployed to predict wafer-manufacturing process outcomes. Panoptes Image Metrology (IM) is an end-to-end image-metrology solution with a flexible pipeline and customizable metrics, covering image processing, metrology, and analytics. Using AI-driven algorithms, Panoptes IM is described as producing precise, consistent, and repeatable results while requiring less training and repeated manual work, and it is intended to extract more information from metrology images with less operational effort. Panoptes IM is also intended to measure process outcomes more efficiently and unlock deeper insights from metrology images beyond the user’s current understanding. Technical resources listed for Panoptes VM include a 2025 white paper and publications from 2023 through 2026 in venues including the SEMI Advanced Semiconductor Manufacturing Conference and SPIE Advanced Lithography & Patterning.
Capabilities and products
- Panoptes Virtual Metrology AI system
- real-time process-outcome prediction for all fab wafers
- automatic AI model training and management
- fab-scale model operation for high-volume manufacturing
- real-time process recipe adjustment at single-wafer level
- process anomaly detection and root-cause identification
- metrology sampling optimization
- equipment anomaly and mismatch prediction
- Panoptes AI image metrology