Coherent began construction of an expanded manufacturing facility in Sherman, Texas, to produce advanced optical components for AI systems. The site will feature the world’s first 6-inch indium phosphide fabrication line, essential for creating the optical backbone of modern AI infrastructure. NVIDIA founder and CEO Jensen Huang and Coherent CEO Jim Anderson attended the groundbreaking event, highlighting the partnership’s role in advancing U.S. semiconductor manufacturing. The facility will scale production of InP wafers that enable high-speed data transfer between chips, servers, and data centers. The project is supported by a $50 million CHIPS Act grant and builds on earlier state funding. Source: nvidia

The expanded facility will support more than 550 direct jobs and thousands of indirect jobs once fully operational. It will produce lasers, transceivers, and optical modules that enable data movement across NVIDIA networking systems. Coherent and NVIDIA have a long-standing partnership, with NVIDIA investing $2 billion in Coherent to support R&D, future capacity, and U.S.-based manufacturing. The collaboration includes a multibillion-dollar purchase commitment for advanced laser and optical networking products. The facility is part of a broader effort to bring advanced semiconductor manufacturing back to the U.S., with NVIDIA aiming to produce up to $500 billion of AI infrastructure through industry partnerships. Source: nvidia

The event showcased the growing importance of optical interconnects in AI systems, as copper wiring becomes inadequate for high-speed data transfer across large data centers. NVIDIA’s NVL576 system, which links eight NVLink racks of GPUs, requires silicon photonics to maintain efficiency. Optical solutions offer a one-time cost for converting electrical signals to light, but they provide nearly unlimited distance and power efficiency at scale. The 6-inch InP fab represents a significant leap from the 3- and 4-inch wafers used in most global production, increasing usable area and reducing costs. Source: nvidia