T he AI boom runs on Nvidia chips, staggering amounts of electricity and water—and a lot of glass. Corning makes more of it than anyone else.

More precisely, the 175-year-old glass company makes the hair-thin optical fiber that hauls data between servers, buildings and data centers. Demand has become so intense that Corning has recently signed multibillion-dollar agreements with Amazon, Nvidia, Meta and Verizon. This week, AT&T announced a $3 billion deal with Corning to upgrade its home internet service. And Apple has committed another $2.5 billion for Corning’s Gorilla Glass, the tough, hard-to-scratch stuff used in iPhones and Apple Watches.

For one of the nation’s OG technology companies, founded in 1851 and responsible for everything from the glass in Thomas Edison’s light bulbs to the cookware known as Pyrex, AI infrastructure is a lively third act.

Corning’s stock has nearly doubled in the last 12 months, dramatically outperforming the S&P 500’s 16%. In 2025, the company posted net income of $1.6 billion on sales of $15.6 billion, up 19% over 2024. This year, analysts expect revenue to jump nearly 25% to $19.3 billion and net income to double. And Corning’s 67-year-old CEO, Wendell Weeks, aims to hit $40 billion in revenue by 2030.

“Fiber optics and all the products that are being used today [for generative AI] were originally invented for telecom,” says Weeks, a Harvard MBA who prefers tailored suits to the jeans and sneakers favored by West Coast tech tycoons. “We invented a lot of them decades ago, and we’re seeing a total reinvention of that product set.”

Yet amid Corning’s best year in a generation, there are reminders of its volatile and sometimes ugly past. Its stock surged nearly 200% during the first half of 2026 only to plummet 50% in July based on little more than analysts’ worries over its inflated valuation and the expanding AI bubble. For Weeks, who joined Corning in 1983 and took over its optical fiber business in 1996, it must have felt like déjà vu.

Many investors are worrying that the AI buildout is a bubble. Cadence Design Systems supplies the software, hardware and intellectual property that semiconductor companies use to design increasingly complex electronics. That business captures secular growth trends across AI infrastructure as well as other sectors including life sciences and autonomous vehicles. Cadence generated $1.58 billion in revenue for Q2 2026, up 24% year-over-year. Double-digit growth across all business segments led management to raise 2026 revenue guidance by 19% versus a year ago. Based on strong sales momentum, Cadence shares could reach $420 within 18 months.

Jon D. Markman is president of Markman Capital Insight.

During the telecom bubble of the late 1990s and early 2000s, when many thought the network buildout from the rapidly expanding internet would go on for decades, Corning borrowed heavily, spending $10 billion on acquisitions. After optical fiber orders from its best customers, including Nortel, suddenly dried up when the bubble burst in 2001, Corning’s stock nosedived from $113 to $1.10 in under two years.

“When I took over, we were at a spot that was relatively dark. We were coming off of the dot-com crash, which ended up devastating a lot of our customers,” says Weeks, who was named CEO in 2005. “For the first time in a long time, we had to do some pretty significant layoffs to get the company on positive ground. We had to think deeply about what our mission was. What did we want to do, and how do we go about building a better version of ourselves?”

Ultimately, what saved Corning was its treasure trove of innovations in materials science. First to the rescue were ultra-thin flat-panel LCD television displays (made with Corning glass), which consumers began buying en masse in the early 2000s. Then came Gorilla Glass, developed at Corning in the 1960s, which turned into a blockbuster after Steve Jobs persuaded Weeks during a late 2006 meeting to begin producing millions of tiny scratch-resistant screens for the iPhone’s launch in June 2007. Other scientific miracles made of glass and ceramics continued to replenish and fill Corning’s till, including pollution filters for cars and diesel trucks, specialty glass for the medical industry and ever thinner glass for TVs and laptops.

In a flashback to the late 1990s, optical fiber sales are once again leading the charge, contribu­ting 45% of Corning’s $8.6 billion revenue in the first half. Weeks thinks it’s going to keep growing for years but insists Corning won’t get caught in another tailspin because he’s seen that movie before and is obsessing over diversification.

One idea: convincing automakers to start making their windshields out of Gorilla Glass. They would be pricey, yes, but they would also be resistant to cracking and rock damage, and weigh significantly less. Supplying all the components needed to make solar panels—polysilicon wafers and ingots, solar cells and modules—at American factories (nearly all are made in Asia currently ) is another.

Research and development has always been a top priority for Corning. This year’s budget should total about $1 billion, or 7% of revenue. Because Corning prioritizes research—and gives them beautiful facilities, like the 2 million–square-foot Sullivan Park center, which sits on 400 acres in the hills of the Chemung River Valley, just outside Corning, New York—company scientists and engineers often stay for decades, even their entire careers, says chief technology officer Jaymin Amin.

“We’ve got folks who come over here and want to stay. It almost becomes like ‘Hotel California,’ ” quips Amin, a 29-year veteran and a holder of 50 patents, who joined Corning after getting his Ph.D. in optoelectronics from the University of Southampton in the U.K. Even Weeks, with a background in finance, has done research and is named on 47 patents.

The company is also smart in its partnerships. “Corning is very humble in strategic alliances and other cooperative technology ventures,” says Kathryn Harrigan, a professor at Columbia Business School.

That’s evident in its decades-long ongoing partnership with Apple, which resulted in Corning’s Harrodsburg, Kentucky, factory becoming the global supplier for all Gorilla Glass used in iPhones and Apple Watches, accounting for about $1.4 billion, or 9%, of Corning’s revenue in 2025. With Corning’s hyperscaler partners it’s developing denser, more advanced types of fiber that carry more data with lower latency. Eventually, those glass wires could replace copper ones in GPUs, helping to prevent overheating.

Given the current tariff turmoil, Corning’s expansion into polysilicon for solar panels and new solar module production—all made in the U.S.—couldn’t be better timed. Solar is Corning’s fastest-growing business line, with sales nearly doubling to $808 million in the first half of 2026 compared to a year ago.

“We had the capabilities to do it, so we built it. Ninety percent of the materials in a photovoltaic cell are materials that we’re expert in,” Weeks says. “We thought we could do in solar what we’ve done in industry after industry. We’ll become a low-cost producer, and then we’ll invent products and make them better.”

According to CTO Amin, that’s likely to include new types of solar panel glass that are more durable and resistant to damage from hail and windborne debris. It’s also researching ways to create solar cells that capture more of the spectrum, such as ultraviolet or infrared light, to increase the total amount of energy that solar panels generate.

“Are there ways to take bad photons and convert those into good ones? Take photons that come in the UV and convert them into visible light?” Amin asks rhetorically. “You can actually do that through very specific glass layers that you can put on the solar cells.”

Weeks also sees a big opportunity for Corning to modify ceramic filtration technology it created for vehicle exhaust treatment systems as a new, more affordable method for capturing industrial carbon emissions from refineries, chemical plants and steel mills.

“I’d always been one who believed there’s no way we would ever get ourselves in the position that we’d have to pull carbon directly out of the atmosphere. That just struck me as ridiculous because that takes energy,” Weeks says. “Now I’m starting to realize we just may be that silly.”

Proving yet again that you can teach this old glassmaker new tricks.