San Francisco-based startup Niobia has developed AI-powered software that helps semiconductor and battery manufacturers detect production defects and investigate their causes. The system uses AI agents running directly on manufacturing equipment to identify small flaws, allowing engineers to analyze production problems through a chatbot rather than spending days investigating them manually.
Founded by Gaurav Jha and Harveen Kaur, Niobia was selected as a finalist in TechCrunch Disrupt 2026’s Startup Battlefield competition. The company aims to reduce manufacturing delays caused by defects that can cost manufacturers millions of dollars during research, development, and production.
How Niobia’s AI Detects Manufacturing Defects
Niobia’s software operates through self-contained graphics processing units (GPUs) installed directly on production equipment. These devices run AI models that monitor manufacturing processes and identify defects, including pinhole-sized scratches and other surface irregularities.
The system collects production information in a central platform, allowing engineers to investigate potential causes through natural language questions. According to Niobia’s technology overview, its AI agents can examine manufacturing records, process measurements, and equipment data to identify possible reasons for production failures.
Rather than replacing engineers, the software provides diagnostic information that helps them determine appropriate corrective actions. Niobia says its system can reduce the time required to investigate defects from days to minutes.
The company trains open-source AI models using specialized semiconductor and battery manufacturing research. This additional training allows the software to recognize defects and analyze production conditions associated with specific industrial processes.
Niobia Keeps Sensitive Manufacturing Data On-Site
Niobia designed its software to operate within customers’ manufacturing environments rather than sending proprietary information to public AI cloud services. Jha told TechCrunch that manufacturers are particularly concerned about exposing confidential production data, even when cloud providers offer security protections.
The company’s security framework allows customers to retain control of their encryption keys, preventing Niobia from independently decrypting sensitive information. The approach is intended to protect manufacturing processes and research data from unauthorized access.
Both founders hold doctorates in chemistry and previously worked in semiconductor and battery manufacturing. Their industry experience helped shape the company’s focus on production defects, which can disrupt manufacturing schedules and increase costs.
Niobia Targets Three Industrial Manufacturing Processes
Niobia is initially concentrating on film coating, electroplating, and fluidized bed reactions. These processes are commonly used in semiconductor fabrication, battery manufacturing, and materials production.
Film coating involves applying thin layers of material, while electroplating uses an electric current to deposit metal onto a surface. Fluidized bed reactions involve suspending solid particles in a moving gas or liquid to support chemical processing.
Jha said the company eventually wants its AI software to become more autonomous, allowing manufacturing equipment to identify and address problems with less human intervention. Its longer-term development plans include systems capable of supporting increasingly automated factory operations.
Featured image credits: Magnific.com
For more stories like it, click the +Follow button at the top of this page to follow us.