GlobalFoundries will receive an expected $300 million U.S. CHIPS R&D award for silicon photonics research. The program will advance optical materials, wafer technologies, and advanced packaging.
As AI workloads grow, efficient connections between computing components have become increasingly important.
GlobalFoundries Secures $300 Million for AI Interconnect Research
On July 29, GlobalFoundries signed a letter of intent with the U.S. Department of Commerce. Under the agreement, the company expects $300 million from the CHIPS Research and Development Office.
The funding will support next-generation silicon photonics and related manufacturing technologies. Additionally, the program will advance near-packaged optics and co-packaged optics. Both approaches position optical technology closer to processors.
Alongside those efforts, the research will incorporate 3D hybrid bonding into advanced packaging processes. GlobalFoundries plans to develop the work across facilities in New York and Vermont.
The Growing Challenge of Moving Data Inside AI Systems
AI infrastructure increasingly requires high-bandwidth connections between processors and other components. Consequently, data movement has become a significant challenge as computing demands increase.
Traditional copper connections face practical limitations as bandwidth requirements rise. In contrast, silicon photonics uses light to transmit information between components. The approach can support higher bandwidth density while improving energy efficiency.
Meanwhile, GlobalFoundries has already developed technology for optical connectivity. In May, the company introduced SCALE, an optical module solution for co-packaged optics. The platform uses wavelength-division multiplexing to transmit data through optical fibers.
Silicon Photonics Becomes a Strategic AI Hardware Battleground
Beyond processor performance, advanced packaging increasingly influences how efficiently AI hardware operates. GlobalFoundries targets data transfer speeds of up to 400 gigabits per second. Furthermore, the company expects its approach to deliver improved energy efficiency.
Taiwan already has significant capabilities in advanced semiconductor packaging. For example, TSMC operates its 3DFabric platform, which includes CoWoS and SoIC technologies.
CoWoS integrates multiple system-on-chips with high-bandwidth memory for AI and high-performance computing. Because of this, GlobalFoundries is developing optical connectivity as part of its U.S. manufacturing strategy.
GlobalFoundries’ Path Toward Next-Generation AI Connectivity
In summary, the program will focus on research, development and manufacturing. At the same time, the company also aims to create a U.S.-based path from silicon wafers to tested optical modules. However, the award does not guarantee commercial success. GlobalFoundries must still translate research into reliable products and scalable manufacturing.
Meanwhile, other semiconductor companies continue developing advanced packaging and optical technologies. As a result, competition will extend beyond processor design.
Ultimately, AI hardware needs more than increasingly powerful processors. It also needs efficient connections capable of handling growing data demands. The GlobalFoundries investment shows how optical technology is becoming part of that effort.
