A supply-chain deal driven by AI scale
Meta Platforms has signed a multiyear agreement worth up to $6 billion with Corning to secure optical fiber, cable, and connectivity solutions for Meta’s U.S. data centers. The arrangement underscores how the AI buildout is reshaping the supply chains behind cloud computing: alongside chips and power, high-capacity networking gear is becoming a bottleneck that large buyers want to lock down years in advance.
Under the deal, Corning will supply advanced optical products designed for the density and data throughput requirements of modern AI facilities. Meta, which has been expanding its AI infrastructure ambitions, is positioning the agreement as part of building the most advanced data center network it can—one that supports training and running large models at scale.
North Carolina expansion and domestic manufacturing
Corning says it will expand manufacturing capacity in North Carolina, including at its optical cable facility in Hickory, where Meta will serve as the anchor customer. The companies have indicated that the expansion is expected to support employment growth in the state, reflecting the labor needs of higher output at large-scale fiber and cable sites.
The focus on U.S.-based production aligns with a broader trend among hyperscalers and major tech platforms: securing domestic manufacturing partners for critical components in an environment where geopolitics, shipping disruptions, and sudden demand spikes have made just-in-time sourcing riskier.
Why fiber matters for AI data centers
AI data centers require enormous east-west traffic inside facilities, moving data between GPU clusters, storage, and networking switches. High-quality fiber and cabling can become a gating factor when new halls or campuses need to be brought online quickly. A large, long-term commitment can help a supplier justify capital investment and give a buyer more predictable delivery schedules.
Investors responded strongly to the announcement, with Corning shares jumping on the day as traders priced in the significance of a large, multi-year anchor customer. The deal also puts a spotlight on the less glamorous side of AI infrastructure—components like glass, fiber, and connectors—where demand is rising alongside headline-grabbing spending on chips and data center real estate.