Corning Lifts Expectations Amid Fiber Demand in AI Data Centers
Corning increased its second quarter sales expectations by around $200 million due in large part to greater-than-expected demand for fiber connections inside of data centers running AI applications.
This article originally appeared on Light Reading.
Fiber vendor Corning said it's seeing better-than-expected sales of its fiber products among data center operators.
Specifically, the company said it now expects second-quarter sales of around $3.6 billion, which is up from its prior guidance of $3.4 billion. The company also forecast a return to year-over-year growth in its earnings.
"The outperformance was primarily driven by the strong adoption of our new optical connectivity products for generative AI," Corning's CEO Wendell Weeks said in a release. He said Corning now expects to grow by more than $3 billion in annualized sales over the next three years as "cyclical factors and secular trends combine."
Corning is scheduled to report its full second quarter results on July 30.
Corning attributed its more optimistic outlook to a variety of factors, but primary among those was sales of fiber for graphical processing unit (GPU) interconnections inside data centers. GPUs – including those from Nvidia – are widely viewed as essential to powering advanced AI services.
For example, Corning said GPU-based parallel computing for AI requires 10x as much fiber as traditional cloud networks. "Blackwell GPU [from Nvidia] rack houses 72 GPUs, each connected by 8 fibers, for a total of 576 fibers, or 18X as many fibers as a legacy CPU rack," Corning wrote.
Corning's stock bounced on its revised second quarter guidance, rising from around $38 per share prior to the company's announcement Monday to around $42 per share after.
AI's Influence
Corning manufactures much of the physical cabling used in US fiber networks. As a result, the company is often viewed as a bellwether in the US telecom industry.
But fiber is also a key networking component inside data centers. And AI providers like Microsoft, Google and Amazon have been investing heavily into data centers in order to install the Nvidia GPUs that power most AI services like ChatGPT.
Indeed, Synergy Research Group recently reported that the number of large data centers operated by hyperscale providers globally reached 992 at the end of 2023. The firm said that figure ought to increase by 120-130 every year, "with generative AI technology being a prime reason for that increased scale."
The upswing in data center investments coincides with a dramatic reduction in spending among telecom network operators globally. That spending slowdown has cut into the finances of vendors ranging from Ericsson to Crown Castle to Corning.
Indeed, the slowdown has pushed a number of telecom vendors to seek opportunities elsewhere. Nokia, for example, cited the data center opportunity as a prime reason behind its $2.3 billion proposal to purchase Infinera.
"One of the key attractions of this acquisition is that it significantly increases our exposure to data centers," Nokia CEO Pekka Lundmark explained of the deal.
Nokia isn't alone. For example, Cisco too has high AI expectations.
But suppliers including Corning have suggested that network operators will eventually resume spending.
"We anticipate optical communications sales will spring back because we believe and our carrier customers have confirmed that they purchased excess inventory during the pandemic and that they've been utilizing this inventory to continue deploying their networks," Corning CEO Weeks said in January. "We believe these carriers will soon deplete their inventory and execute on the increased broadband deployment plans they've communicated to us over the last several months."
In a document linked to its latest release, Corning reiterated those expectations, pointing specifically to the US government's $42.5 billion Broadband Equity Access and Deployment (BEAD) program. The company said it expects operators to begin spending on BEAD-related fiber projects starting next year.
About the Authors
You May Also Like