US awards GlobalFoundries $300M for silicon photonics
The US Commerce Department has awarded GlobalFoundries $300 million to advance silicon photonics, a move aimed at boosting domestic AI infrastructure and reducing reliance on foreign chipmakers.
The U.S. Department of Commerce has signed a letter of intent to award chipmaker GlobalFoundries $300 million to accelerate the development of silicon photonics. This funding is designed to support next-generation wafer technologies, novel optical materials, and advanced packaging solutions, including the company's 3D hybrid bonding technology. Under a separate agreement, the federal government will also acquire an equity stake in GlobalFoundries representing approximately 1% ownership of the firm.
The investment will build directly on GlobalFoundries' SCALE (silicon photonics co-packaged advanced light engine) platform. This platform targets modular designs capable of achieving a 400 Gb/s data-transfer-rate performance. Crucially for energy-conscious operators, it aims to deliver a fivefold improvement in energy efficiency compared to current-generation implementations. This shift to optical connectivity replaces traditional copper with light-based data transfers to handle increasingly complex workloads.
This $300 million award follows a previous $1.5 billion in funding that GlobalFoundries received from the Department of Commerce in 2024 under the CHIPS and Science Act to expand its manufacturing facilities in Malta, New York, and Burlington, Vermont. The funding aligns with broader federal efforts to secure domestic supply chains, which currently rely heavily on overseas manufacturers like Taiwan Semiconductor Manufacturing Company and Israel-based Tower Semiconductor. Other recent federal initiatives include a proposed $150 million equity stake for next-generation semiconductor technology in January and a $2 billion commitment for quantum computing in May.
For AI infrastructure practitioners and data center architects, this domestic push for silicon photonics promises to alleviate bandwidth bottlenecks and power constraints. As GlobalFoundries CEO Tim Breen noted, the transition from copper to optical data transfer is no longer a future concept but an active reality. Practitioners can expect more reliable domestic sourcing for high-speed optical components, enabling them to scale up AI clusters with significantly lower thermal and energy overheads while maintaining high throughput.
This is our own summary of reporting by AI Business
