Key Highlights

  • GlobalFoundries and Marvell Technology expanded their multi-year agreement to increase production capacity for a semiconductor used in AI Data Center connectivity.
  • Shares of GlobalFoundries rose around 4%, while Marvell shares gained 6.3% following the announcement.
  • The expanded deal increases capacity for silicon germanium technology at GlobalFoundries’ Facility in Burlington, Vermont.
  • Rapid growth in AI data centers is driving Demand for faster and more energy-efficient optical networking components.

Expanded Capacity Agreement

GlobalFoundries (NASDAQ:GFS) and Marvell Technology (NASDAQ:MRVL) have expanded their agreement to increase production capacity for a type of semiconductor used in high-speed optical connections within data centers. Shares of GlobalFoundries rose around 4% in morning trading, while Marvell shares gained 6.3% following the announcement.

The multi-year agreement will increase capacity for GlobalFoundries’ silicon germanium, or SiGe, technology, which enables faster and more energy-efficient semiconductor devices, at its facility in Burlington, Vermont.

Demand Driven by AI Growth

The rapid growth of AI data centers is driving demand for SiGe technology used in optical networking, as higher data volumes require faster optical connections between servers and storage systems. "Our expanded collaboration with GF will help ensure we have the SiGe technology and Manufacturing capacity to support the significant growth we see ahead," said Robb Johnson, vice president of foundry technology at Marvell.

The companies said the increased manufacturing capacity would help Marvell meet demand for advanced optical networking products, including pluggable optical transceivers, near-packaged optics and co-packaged optics.

A Growing Niche of AI Infrastructure

The deal reflects a broader trend across the semiconductor industry, where companies are racing to expand manufacturing capacity for specialized components that support the AI infrastructure buildout, even in less visible parts of the Supply chain such as optical networking. As AI data centers scale to handle larger and more complex workloads, the connectivity infrastructure linking processors, memory and storage has become an increasingly critical bottleneck.

Outlook

Whether the expanded Partnership meaningfully eases capacity constraints for AI data center connectivity will likely become clearer as new manufacturing capacity comes online at the Vermont facility. Given the broader industry-wide scramble to secure specialized components for AI infrastructure, agreements like this one are likely to become more common as operators continue scaling their networks.