NVIDIA has launched a multiyear partnership with Corning that will dramatically expand US production of optical connectivity technology for AI data centers.
The partnership supports the AI sector’s shift to optical technologies that can move vast amounts of data between processors far faster than copper wiring. Today’s AI workloads rely on thousands of GPUs operating in parallel, creating enormous demands for bandwidth and power efficiency inside hyperscale data centers.
The agreement calls for Corning to build three advanced manufacturing facilities in North Carolina and Texas focused on optical networking products used in AI systems. The companies claim the expansion will create more than 3,000 jobs and increase Corning’s US-based optical connectivity manufacturing capacity by tenfold. Domestic fiber production capacity is expected to rise by more than 50%.
The Corning deal “is consistent with NVIDIA’s strategy of investing in the companies building the AI infrastructure,” Gil Luria, Head of Technology Research at D.A. Davidson, told Techstrong.ai. “By deepening these relationships not only is NVIDIA providing the capital for its partners to invest in growth, but also benefitting from their success as the value of their equity rises as a result.”
Faster and Less Expensive AI Data Transfer
Greater adoption of optical interconnects and co-packaged optics in AI infrastructure is a foundational trend in next-generation AI systems. Not only do optical fiber and photonics technologies allow data to travel faster, they require less power, a critical factor as energy-hungry AI clusters rapidly expand.
NVIDIA CEO Jensen Huang has repeatedly emphasized the importance of optical networking in the company’s future AI architecture roadmap. At the company’s GTC conference earlier this year, Huang described co-packaged optics as essential to scaling AI infrastructure.
The financial component of the partnership highlights NVIDIA’s long-term commitment to optical networking. NVIDIA secured warrants allowing it to purchase up to 15 million shares of Corning stock at an exercise price of $180 per share. The chipmaker also obtained a pre-funded warrant for up to 3 million shares through a $500 million investment.
Corning shares surged following the news, extending a remarkable rally that has seen the company’s stock climb sharply over the past year as demand for AI infrastructure components accelerates.
Though widely recognized for producing display glass used in Apple’s iPhone lineup, Corning’s optical communications division has become one of its fastest-growing businesses. The company pioneered low-loss optical fiber technology in 1970 and now supplies fiber and connectivity products used throughout data center networks.
The NVIDIA agreement builds on Corning’s broader effort to build its business in the AI infrastructure sector. The company has outlined aggressive expansion goals tied to what they describe as a generational increase in AI-related demand. Corning is targeting a $30 billion annualized sales run rate by 2028, driven in large part by growth in its photonics business.
The company’s strategy centers on becoming a key domestic supplier for AI data center construction at a time when the US government is placing greater emphasis on American manufacturing capacity and supply chain resilience.
The deal also demonstrates NVIDIA’s increasingly all-encompassing role across the AI ecosystem. In addition to dominating the GPU market used for AI training and inference, NVIDIA has invested heavily in networking, photonics and data center infrastructure companies. Earlier this year, the company invested billions in optical component vendors including Coherent and Lumentum.
Demand for high-speed connectivity has intensified as cloud providers and AI model developers invest in larger AI clusters. Technologies that the NVIDIA-Corning partnership will produce, ranging from silicon photonics to co-packaged optics, are now seen as crucial to support future AI scaling.
