Nvidia and Corning Partner on Optical Tech to Scale AI Data Centers

The artificial intelligence boom is reshaping not just the frontiers of computing but the physical infrastructure that makes it possible. On Wednesday Nvidia and Corning announced a multiyear partnership to build three advanced manufacturing facilities in North Carolina and Texas that will be devoted entirely to optical technologies designed for the AI era. The deal, whose financial terms were not disclosed, will create at least 3,000 jobs and expand Corning’s U.S. optical manufacturing capacity tenfold. The news sent Corning shares up 14 percent and lifted Nvidia nearly 3 percent in early trading.

The collaboration marks a significant step in the race to overcome one of the central bottlenecks in scaling artificial intelligence. For years data centers have relied on copper cables to move information between servers and processors. As AI models grow larger and more complex the limitations of copper in speed power consumption and heat have become acute. Nvidia is now turning to optical fiber and glass-based interconnects an approach known as co-packaged optics to transmit data at the speeds and efficiencies required for the next generation of AI systems. At Nvidia’s GTC conference last year CEO Jensen Huang described co-packaged optics as essential to the entire AI buildout calling the shift from copper a foundational change in how data centers are designed.

Corning chairman and CEO Wendell Weeks framed the partnership in explicitly national terms. “What Nvidia is doing is nothing short of extraordinary not just for the future of AI but for the American advanced manufacturing workforce,” he said in a joint statement. The 175-year-old glassmaker has undergone a striking transformation. Its stock is up more than 250 percent over the past year as it pivots from traditional markets into the infrastructure of the new economy. That pivot received a major boost in January when Meta pledged up to $6 billion to help expand Corning’s optical cable plant in Hickory North Carolina. The new Nvidia facilities will sit alongside that momentum effectively turning parts of the American South and Southwest into specialized hubs for the hardware that powers large language models and other frontier AI systems.

The announcement arrives at a moment when demand for AI-related hardware is not simply strong but accelerating in ways that continue to surprise even optimistic forecasts. A day earlier AMD reported first-quarter results that exceeded Wall Street expectations on nearly every metric. Revenue reached $10.25 billion up 38 percent from a year ago and well ahead of the $9.89 billion consensus estimate. Data center revenue the segment most closely tied to AI workloads jumped 57 percent to $5.8 billion. Net income nearly doubled. The company’s guidance for the current quarter also topped forecasts with expected revenue of roughly $11.2 billion.

AMD’s stock soared as much as 20 percent in premarket trading reflecting investor conviction that the company is successfully carving out a larger share of the AI accelerator market long dominated by Nvidia. CEO Lisa Su described the data center business as now the “primary driver” of the company’s revenue and earnings growth. She expressed “strong and increasing confidence” that AMD can achieve tens of billions of dollars in data center AI revenue next year and surpass its long-term target of more than 80 percent annual growth in that segment. The results reinforced a broader market narrative that the AI buildout is moving from hype to heavy industrial reality with multiple winners across the supply chain.

Super Micro Computer and other companies that build the physical servers and racks for AI training clusters also rose sharply in premarket trading on Wednesday. Together the moves illustrate how the appetite for computational power is rippling outward from chip designers to materials scientists to factory floors. Optical interconnects may sound like a narrow technical detail but they are becoming as strategically important as the GPUs themselves. Without faster ways to shuttle data between thousands of processors the clusters required to train and run the largest models risk hitting physical limits on bandwidth and energy use.

The developments also highlight a quiet renaissance in American manufacturing. Facilities that once produced fiber for telecommunications or specialty glass for consumer products are being reoriented toward the insatiable needs of AI. Policymakers have spent years talking about bringing advanced manufacturing back to the United States; the combination of private capital from Nvidia Meta and others with strategic investments in places like North Carolina and Texas is achieving some of that goal in real time. At least 3,000 new jobs will come directly from the Nvidia-Corning project with thousands more likely in the surrounding supply chains.

Yet the speed of this transformation also raises longer-term questions about concentration and sustainability. A handful of companies now drive extraordinary amounts of capital investment and energy demand. The optical fibers and advanced chips being manufactured in these new plants will power data centers that already consume electricity at the scale of small cities. As the infrastructure buildout accelerates the economic benefits are clear but so are the stakes for everything from national energy policy to the geographic distribution of high-tech employment.

For now the market’s verdict is unambiguous. Investors are pricing in years of continued expansion. The partnership between a chipmaker that has become the most valuable company in the world and a legacy American manufacturer signals that the AI boom is entering a new phase one defined less by flashy model releases and more by the gritty work of building the physical backbone that makes those models possible. The factories rising in North Carolina and Texas may ultimately prove as important to the AI future as the algorithms running inside them.