معلومات أساسية
San Francisco-based startup Niobia is introducing AI software designed to identify production defects in battery and semiconductor manufacturing. By deploying AI agents directly onto manufacturing equipment, the company aims to help engineers diagnose failures in minutes rather than days, potentially saving millions of dollars in production costs.
The system utilizes AI agents mounted on self-contained GPUs to detect microscopic flaws, such as pinhole-sized scratches. These agents are powered by open-source models further trained on specialized research data. To address security concerns, Niobia allows customers to maintain their own encryption keys, ensuring proprietary data remains private and off the public cloud.
Co-founders Gaurav Jha and Harveen Kaur, who possess PhDs in chemistry and industry experience, are focusing the technology on three specific processes: film coating, electroplating, and fluidized bed reaction. Jha noted that while the market is niche, it is essential for powering modern technology.
Niobia is currently a finalist in the Startup Battlefield competition at TechCrunch Disrupt 2026. Looking ahead, the company intends to increase the autonomy of its technology, with the long-term goal of enabling fully self-driving factories for its manufacturing clients.
