The company that makes glass for iPhones has seen its stock price surge 5x due to AI

On the afternoon of May 6, NVIDIA announced an investment. The amount wasn't particularly large, $500 million. But the contract stipulated that it could be expanded to $3.2 billion in the future. Corning's stock rose 14% that day.

More intriguing was the structure of the deal. Of the 18 million share certificates Corning issued to NVIDIA, 3 million shares had an exercise price of $0.0001. This meant those 3 million shares were essentially given to Corning. That same afternoon, at its investor conference in New York, Corning pushed its revenue growth target for 2030 up to $40 billion.

But this isn't the most unusual part of Corning's recent months. The quarterly report of this "iPhone screen glass supplier" stated that in the past few months, two other unnamed companies have each signed multi-year contracts worth $6 billion with Corning. The reason for saying "another" is because Corning had just signed a contract of the same scale with Meta.

Count them and you'll find that in the past 4 months, at least 4 major AI deals worth tens of billions of dollars have been concentrated on this 174-year-old glass company. Over the past 6 months, Corning's stock has risen 140%, and compared to two years ago, it has increased fivefold.

From Selling Phone Glass to AI Factory Darling

If you're reading this article on your phone, the screen covering it is most likely a piece of glass produced by Corning. Since Apple's first-generation iPhone in 2007, Corning's Gorilla Glass has almost become the default choice for high-end smartphone screens globally. But "phone glass supplier" is just one facet of Corning, and not even the most profitable one.

Gorilla Glass production line at a Corning factory, source: Apple

Founded in 1851, the company made the first glass bulb casing for Edison. In the 1970s, it invented low-loss optical fiber from scratch, pioneering the entire modern fiber optic industry. The iPhone glass in 2007 marked its third major business pivot. Today, Corning is undergoing its fourth transformation, with optical communications becoming the true engine of its business.

Corning's optical communications business has a history of over 50 years, but the customer structure of this business has completely flipped in the last two years.

For a long time, Corning's optical fiber was mainly sold to telecom operators like AT&T and Verizon. They used it for fiber-to-the-home and building 4G and 5G base stations. In 2009, Corning launched a data center cabling solution called EDGE, officially adding data center operators to its client list. Over the past decade-plus, fueled by the mobile internet boom, the proliferation of cloud services, and the explosive growth of remote work during the pandemic, Corning's optical communications business rose steadily but never became a major revenue driver.

In November 2022, OpenAI introduced ChatGPT to the public. From that moment, data centers worldwide began redesigning their physical infrastructure for the new computing task of AI training. The fiber optic density required for AI training is unprecedented.

The first sign appeared in August 2024. A U.S. telecom operator named Lumen booked 10% of Corning's global optical fiber production capacity in a single order for two consecutive years. This was the earliest public signal of Corning's transition towards the AI sector.

By early 2026, the aforementioned four $6 billion contracts came in a concentrated burst. Corning has partnered with data center operators for 15 years, but the shift from "secondary customer" to "absolute mainstay" has only happened in the past 24 months.

The direct effect of this customer shift is written in Corning's financial reports. Corning's full-year revenue in 2023 fell 11% year-on-year during an industry trough, but by 2025, full-year revenue surged to $15.6 billion, a 19% increase year-on-year. In the first quarter of this year, revenue grew another 18% year-on-year. The most explosive performance came from the optical communications business, which grew 35% for the full year. Optical communications' share of total revenue rose from 30% in 2020 to 37% in 2025. The absolute change is even more striking, growing from $2 billion five years ago to $6.3 billion in 2025—more than tripling.

This ascent from "secondary business" to "locomotive" is no accident. Behind it is a growth plan spearheaded by CEO Wendell Weeks. The plan has an internal codename: Springboard.

Two years ago, Corning was still described by Wall Street analysts as a "boring glass manufacturer," classified as a mature, low-growth dividend stock. But three years into the Springboard plan, Corning's stock price has risen from just over $30 in early 2024 to $162, a fivefold increase in two years, with a 140% surge in just the past 6 months. The glass factory has transformed into the "nervous system of the AI revolution."

Springboard was first announced in September 2024. Its starting point was the annualized revenue level of Q4 2023, approximately $13 billion. The initial goal was to increase annualized revenue by over $3 billion by the end of 2026 and achieve an overall operating margin of 20%.

But over the next year and a half, this target was raised three times, eventually reaching $6.5 billion, pushing the annualized revenue target for the end of 2026 to a $20 billion run rate. After NVIDIA's investment in Corning on May 6, the company directly raised its internal revenue target for 2030 to $40 billion. Meanwhile, Corning achieved its 20% profit margin target a year ahead of schedule in Q4 2025.

The key to the Springboard plan lies in "premiumization." The company's sales grew 18%, but earnings per share grew 46%—a profit growth rate 2.5 times that of sales. At the operational level, Corning focused on three specific things:

First, forced price increases on its legacy business. Corning's display glass is a mature business that hadn't grown for years. But at the end of 2024, Corning raised prices on this line by over 10% while locking in the yen exchange rate until 2030. The result is that even in a depreciating yen environment, this line consistently contributes $900 million to $950 million in net profit annually, maintaining a net profit margin of 25%.

Second, upgrading optical communication products. In full-year 2025, optical communications sales grew 35%, but net profit surged 71%. This means optical communications is not only selling more but also earning more per fiber optic cable.

Third, ramping up idle capacity. Instead of building massive new factories, Corning restarted capacity left idle during the previous cyclical downturn, boosting the company's overall gross margin from 33% in 2024 to 36% in 2025.

Of course, price increases are only possible if someone is willing to pay. Product upgrades generate more profit only if customers are willing to pay more for the upgraded products. The reason Springboard allowed Corning's profit growth to outpace revenue growth is essentially that its customer structure now includes a group of people willing to pay a premium.

Everyone Is Grabbing Fiber Optics

The AGI race and order demand have made every data center operator intensely anxious about time.

The core business of cloud giants has always been "renting IT to enterprises." Companies like Netflix, Airbnb, and Uber that emerged with the mobile internet generated mostly "north-south" traffic. A user opens an app from outside, the request is sent to a server in the cloud, and the server returns data. Servers communicate with each other occasionally, but the volume and frequency are not high. This network structure's demands on the underlying physical infrastructure were not stringent: Ethernet was sufficient, copper cables were sufficient, and ordinary fiber was sufficient. The cloud giants used this architecture themselves for over a decade—stable, reliable, and profitable.

Until ChatGPT was unveiled, the game began to change.

In the following years, almost all cloud giants started doing their own training. Microsoft is the largest computing provider for OpenAI, AWS is deeply tied to Anthropic, and Alibaba trains Tongyi Qianwen. The core business of cloud giants began shifting from "renting IT to enterprises" to "training AI for the world."

But the chain reaction this shift triggered at the physical infrastructure level defied the common sense accumulated over the past 20 years.

The traffic characteristic of AI training is "east-west." Training a large model might require tens of thousands of GPUs communicating with each other simultaneously, synchronizing the gradients calculated by each other. If any single line is slow, the entire training phase has to wait for it, causing tens of thousands of GPUs to become "cars stuck at an intersection." Therefore, the demands of east-west traffic on latency and bandwidth are dozens of times greater than past north-south traffic.

Before this, the vast majority of high-speed connections inside data centers used copper cables. Copper is cheap, easy to install, and performs stably, making it the default option for data centers. However, the geometric structure of AI training clusters is precisely what copper cables dislike most. Tens of thousands of GPUs are distributed across dozens of cabinets, with distances often exceeding ten meters—copper simply can't handle it. Fiber optics, on the other hand, have no distance limitations for this task.

Overnight, the previously adequate sparse networks became insufficient. Cloud giants needed to lay new fiber optics, denser than ever before.

The scale of this re-laying is already reflected in their capital expenditures. In 2026, the combined capital expenditure of the world's six largest cloud giants is expected to exceed $600 billion. The number of operational hyperscale data centers globally has reached 1,297, nearly three times the number at the beginning of 2018. In 2026 alone, the number of new data centers is expected to exceed 150, with corresponding AI infrastructure spending exceeding $400 billion.

Market research firms estimate that AI clusters require 10 to 100 times the total amount of fiber optics compared to traditional cloud services. This is the fundamental reason Corning can now sign four $6 billion contracts.

Between data centers, and between cabinets within a data center, all fiber optics must pass through something called conduit pipes. These are typically plastic or metal tubes with an inner diameter of 2 to 4 inches, buried underground or run on racks. These pipes have a characteristic: once laid, they are difficult to add to. Adding another pipe between cities means reapplying for right-of-way and digging up the road again, taking years. Adding a pipe inside an already operational data center means shutting it down for modification, taking months.

Conduit pipes about to be buried underground, source: Online

What Corning has done specifically for AI data centers in the past two years is to enable fitting more fiber optics into existing conduits without adding new ones.

Besides making the fiber itself thinner, Corning also changed the arrangement of the fibers from a loose "spaghetti" style into a rollable flat ribbon. When needed, it's unrolled; when not, it's rolled up and packed tightly into the cable. Originally, a 2-inch conduit could only hold a thousand-plus fibers. Corning's new design can fit over three thousand fibers, doubling the density. With a 4-inch conduit and six such cables running side-by-side, it can accommodate over twenty thousand fibers, more than six times the traditional design.

Corning's rollable fiber optic ribbon, source: Corning

It's not just about packing more in; termination is also much more labor-efficient. A 3,456-core cable, using traditional methods, would require over 200 man-hours to connect fiber by fiber. Corning's ribbon design can reduce this to under 40 hours, and cable preparation time is cut by 30%. Remember, the U.S. was already short on fiber optic engineers.

In the construction process of a large AI factory, every one-month delay means massive GPU depreciation and training task postponement, costing hundreds of millions on paper. A product that can cut months of time and millions of dollars in engineering fees is an incredible bargain, even with a 30% to 70% premium on the fiber itself.

Jensen Huang's "Unprecedented Scale"

On May 8, NVIDIA CEO Jensen Huang reiterated in an interview that next-generation AI infrastructure requires extensive optical connectivity, as copper wires can no longer meet the demand. He also stated that NVIDIA plans to scale the application of optical technology to an unprecedented degree.

The details of the investment deal with Corning a few days earlier indeed reflected this "unprecedented scale." Of the 18 million certificates, 3 million shares were essentially "given away." This structure is rare among NVIDIA's ecosystem investments over the past year, indicating that NVIDIA secured significant equity exposure in Corning without using cash immediately—it acted more like a signing bonus for a long-term partnership agreement.

And Corning is not the only piece NVIDIA has wagered on. Since last September, NVIDIA has entered a new phase of investment. First, the scale has increased. Second, the structures frequently involve financial instruments like "frameworks," "options," and "prepaid warrants" to lock in commitments first and realize them in stages. Besides the $100 billion investment framework for OpenAI, NVIDIA has also poured tens to hundreds of billions of dollars into Anthropic, Intel, CoreWeave, and other AI infrastructure players.

Most easily overlooked is its investment specifically in the optical communication line. Besides Corning, NVIDIA has also invested $2 billion each in Lumentum and Coherent, two of the largest global optical component companies. Counting Corning's initial $500 million plus the $3.2 billion option, NVIDIA has poured approximately $7.7 billion into the single niche of optical communications.

Laying out this investment portfolio on a single sheet reveals it perfectly matches an AI factory construction checklist: compute, networking, optics, power, cooling, software, customers, models—for every layer, NVIDIA has locked in at least one key supplier. At this year's GTC conference, NVIDIA integrated this entire stack into a publicly released design blueprint, launching a hardware reference architecture called Vera Rubin DSX and a digital twin solution called Omniverse DSX Blueprint. This whole package is essentially the "construction blueprint for an AI factory."

Building a GW-scale (enough power for 1 million households) AI factory takes 18 to 24 months from planning to production, involving coordinating over 100 suppliers. In the past, this was done by the data center operators themselves, each having to redo interface verification. But NVIDIA's Omniverse DSX systematizes this process. All partner products have already been verified in NVIDIA's digital twin, parameters have been aligned, and interfaces standardized. Cloud giants can simply buy according to NVIDIA's blueprint.

Jensen Huang unveiling the AI factory blueprint platform at the 2026 GTC conference, source: NVIDIA

This is a crucial step for NVIDIA transforming from a chip company into an "AI factory general contractor." Higher integration expands profit margins. Even if AMD or Broadcom produces a GPU with equivalent performance tomorrow, replicating this supply chain coordination capability—from chips to fiber optics to the power grid—would take them at least a few more years.

Therefore, the true meaning of NVIDIA's $3.2 billion option on Corning is locking in a key player for the "localized optical communication production capacity" slot in its AI factory blueprint. And, of course, the one capable of drawing up this blueprint right now is only NVIDIA.