Tokyo Electron and NVIDIA Expand Collaboration to Advance AI Automation in Wafer Fabs

en.Wedoany.com Reported - Tokyo Electron (TEL) and NVIDIA have expanded their collaboration to advance automation and intelligent equipment maintenance in semiconductor wafer fabs (Fab), aiming to enhance customer productivity and yield through the establishment of autonomous equipment operation systems.

Tokyo Electron announced on July 20 that it will comprehensively expand its collaboration with NVIDIA, a leader in artificial intelligence (AI) semiconductors, to strengthen the development of "Agentic AI" and robotics solutions.

Through this collaboration, Tokyo Electron will integrate NVIDIA's cutting-edge software technologies into its core digital transformation (DX) solution concept, "Epsira™." Specific technologies to be introduced include the "NVIDIA Agentic Toolkit," comprising NVIDIA NeMo, NIM, and the NemoClaw blueprint, as well as the open robotics development platform "NVIDIA Isaac."

Based on the NVIDIA Isaac platform, Tokyo Electron will build a 3D digital twin environment linked to the real-time computing platform "NVIDIA Omniverse" library. Equipment design data will be ported into this environment, enabling rapid and precise simulations in virtual space.

Semiconductor manufacturing equipment has complex structures, and robot maintenance must be precisely tailored to equipment configurations. Digital twin technology based on virtual simulation can significantly improve the speed and accuracy of robot work learning in a short time without damaging actual equipment. Additionally, through the agent-specific platforms NeMo and NemoClaw, the speed of AI model learning, tuning, and evaluation can be greatly enhanced, establishing a foundation for safe deployment and operation in real-world settings, thereby maximizing equipment uptime.

Tokyo Electron's Epsira is a digital transformation infrastructure that leverages data utilization to improve customer semiconductor manufacturing productivity. This solution incorporates technologies such as AI-based wet cleaning cycle extension (MTBWC) analysis, robot maintenance, and AI problem-solving analysis, minimizing scheduled maintenance and unexpected downtime, and enabling even inexperienced operators to easily achieve an optimal operating environment.

Keiichi Matsushima, Vice President of Corporate Development at Tokyo Electron, stated that deepening the partnership with NVIDIA, which leads the AI era, is highly significant. Based on this, Tokyo Electron will continue to enhance its robot maintenance and AI fault analysis capabilities, contributing to maximizing customer productivity.

Amit Goel, Senior Director of Robotics and Edge AI Ecosystem at NVIDIA, noted that semiconductor manufacturing is reaching a new turning point in automation, with robotics serving as a key axis for improving the uptime and resilience of advanced wafer fabs. Tokyo Electron is leveraging NVIDIA's agentic and physical AI frameworks to make fab operations more intelligent.