TAIPEI – United Microelectronics Corp. (UMC), Taiwan’s second-largest contract chipmaker, is partnering with a Harvard University spinoff to accelerate the production of advanced optical technology designed to improve data transmission speeds within artificial intelligence computing infrastructure. The collaboration, announced by a senior UMC executive, aims to address growing demands for faster processing in AI clusters and data centers.
The partnership focuses on mass producing a novel optical material and associated chips, according to the Nikkei Asia report. This technology is intended to overcome limitations in current data transfer methods, which are becoming bottlenecks as AI models grow in complexity and require increasingly rapid data access.
The Harvard connection centers on the Harvard Undergraduate Machine Intelligence Community (HUMIC) and potentially the Kempner Institute for the Study of Natural and Artificial Intelligence. The Kempner Institute recently announced a significant expansion of its AI supercomputing cluster, adding 424 NVIDIA H200 GPUs and 192 RTX PRO 6000 Blackwell GPUs, bringing the total to 1,144 GPUs. This expansion, completed on March 11, 2026, will provide a platform for research requiring exaFLOPS-level computing power, according to Elise Porter, the Institute’s Executive Director.
The new RTX PRO 6000 Blackwell GPUs are specifically designed to enable advanced optical and physics-based simulations, supporting research in robotics and neuroscience, areas where improved data transmission is critical. The Harvard Undergraduate Society for Artificial Intelligence also fosters research and development in the field.
UMC’s move reflects a broader industry trend toward advanced packaging technologies that combine multiple semiconductor chips to enhance computing performance. The demand for such technologies is being fueled by the expansion of AI from cloud-based systems to edge computing, requiring more efficient and powerful processing capabilities closer to the data source. The company’s investment in optical technology signals a strategic bet on the future of AI infrastructure.
The collaboration with the Harvard spinoff is expected to help UMC capitalize on the growing AI market, providing a competitive edge in the production of specialized chips and materials. The specifics of the manufacturing partnership, including financial details and production timelines, have not been disclosed.
The expanded Kempner Institute cluster will support a range of academic workflows, from large-scale AI training runs to smaller computational experiments. Max Shad, Kempner Institute Senior Director of AI/ML Research Engineering, and John Goodhue, Executive Director of the Massachusetts Green High Performance Computing Center, oversaw the network infrastructure supporting the new GPUs.
