Corning (NYSE: GLW) was founded in 1851. It invented the glass for Edison's light bulb, the television tube, and the toughened glass in smartphone screens. And this quarter it posted numbers that look like those of a young AI infrastructure company.
About the Company: What Corning Actually Sells
Corning is a materials science company - it develops and manufactures glass, ceramics and optical materials. But that description misses the point. What the company actually does is make the physical component through which light travels.
Its main areas:
- Optical communications - optical fiber, cable and connectors. This is the largest segment, and it is the engine this quarter.
- Display technologies - glass for television and computer screens.
- Environmental technologies - emissions filters for vehicles.
- Life sciences - laboratory equipment.
- Specialty materials - including the Gorilla Glass in smartphone screens.
- Solar - polysilicon and wafer manufacturing in the United States.
Why Optical Fiber Suddenly Became a Growth Story
This is the point worth understanding, because it explains the entire report.
An artificial-intelligence data center is not a collection of independent servers. It is one large computer, in which tens of thousands of graphics processors need to talk to each other at enormous speed and continuously. Training a large language model means moving vast quantities of data between chips - over and over.
And here is the bottleneck: copper is limited. Beyond a certain distance and above a certain rate, a copper cable loses signal and generates heat. Light does not. Optical fiber carries more data, over a longer distance, using less energy.
The result is that every new generation of data centers replaces more copper with fiber - not only between buildings, but inside the facility, and gradually inside the rack itself. And that is exactly what shows up in the numbers.
The link to the photonics thesis
This is the same direction we covered in the piece on Tema's photonics fund - the idea that the move from electrons to photons is an entire infrastructure layer being built beneath artificial intelligence. Corning is the most mature and profitable example of that thesis: not a startup making promises, but a company with 175 years of history reporting 65% growth in the relevant segment.
What Was Reported
The main lines:
| Metric | Result | Change |
|---|---|---|
| Core sales | $4.74 billion | +17% |
| Core EPS | $0.78 | +30% |
| GAAP sales | $4.51 billion | +17% |
| GAAP EPS | $0.64 | +19% |
Note the gap between 17% sales growth and 30% earnings growth. That gap is the whole story - and it comes from the margins.
And margins expanded on every line:
| Metric | Current quarter | Improvement |
|---|---|---|
| Core gross margin | 39.6% | +120 basis points |
| Core operating margin | 20.9% | +190 basis points |
| Return on invested capital (ROIC) | 14.9% | +180 basis points |
The third row deserves an explanation: return on invested capital measures how much profit the company generates from every dollar it has put into the business. In a capital-intensive industrial company - with plants, furnaces and production lines - this is one of the most honest metrics there is. A rise of 180 basis points says the existing plants are working harder, not that new plants were built.
Cash flow. Operating cash flow of $1.72 billion, and adjusted free cash flow of $1.42 billion.
The Breakdown - Where the Story Is
Optical communications: $2.07 billion, a 32% jump.
And here is the important number inside the number: enterprise networks grew 65%. "Enterprise networks" in Corning's context means mainly data centers - and this is the component surging because of artificial intelligence. The company explicitly cited Gen AI products as growing faster than the average.
Solar: $438 million, a 90% jump - with a $7 million net loss.
And this point requires careful reading. The segment nearly doubled, but is still losing money. Corning is building a solar manufacturing chain in the United States - polysilicon, wafers and cells - and that investment is still in its ramp-up phase. The growth is impressive; the profitability has not arrived yet.
Guidance: Growth Is Set to Accelerate
| Metric | Third-quarter guidance | Implied growth |
|---|---|---|
| Core sales | $4.9-5.0 billion | about 16% |
| Core EPS | $0.85-0.89 | about 28% |
And the interesting point in this table: earnings are expected to grow at nearly twice the rate of sales. That is precisely the pattern of the current quarter, and management is signaling it continues.
So Why Is the Stock Falling Sharply?
This is the question anyone who saw the numbers asked, and it deserves a direct answer: the stock fell by a double-digit percentage in early trading - around 11% to 14% at different points - despite a strong report. And the reason is not the quarter that ended. It is the guidance.
First, guidance came in below consensus - but only marginally:
| Metric | Company guidance | Consensus |
|---|---|---|
| Third-quarter sales | $4.90-5.00 billion (midpoint: $4.95) | $5.00 billion |
| Core EPS | $0.85-0.89 | $0.86 |
A $50 million gap at the midpoint is not dramatic in itself. The drama is elsewhere.
Second - and this is the central point - the company signaled a possible slowdown in capital spending by major wireless carriers. That bears directly on the optical communications segment: precisely the segment that grew 32%, precisely the segment the thesis rests on, and precisely the segment for which the stock traded at a premium.
Third, the gap between the numbers. The quarter just reported showed optical communications surging 32% and enterprise networks surging 65%. Guidance for next quarter points to total growth of about 16%. The market did not reprice what happened - it repriced the pace ahead.
The distinction worth holding
A strong report and a falling stock are not a contradiction. A stock price does not reflect the quarter that ended - it reflects what the market thinks about the quarters ahead. When a stock trades at a valuation that embeds rapid growth, even an excellent report can disappoint if the guidance accompanying it signals a slowdown. That is exactly what happened here.
And what about Korea? This report was published on the day the KOSPI plunged 10.8% and knocked down markets worldwide. That backdrop certainly amplified the move - on a day of negative sentiment, every disappointment is priced more harshly. But a stock falling by a double-digit percentage on a day the indices are down low single digits has a company-specific reason, not a general one.
The simple read is also the accurate one: the market expected more. Not because the quarter was weak, but because the valuation already embedded the continuation - and management itself flagged that continuation as more cautious.
Why profit grows faster than sales
This is called operating leverage. An industrial company has large fixed costs - plants, equipment, workforce. When sales rise and fixed costs stay more or less in place, every additional dollar of sales contributes far more to the bottom line than the dollar before it. And when the product mix improves at the same time - meaning more high-margin products are sold, such as optical communications solutions - the effect strengthens.
The Bull Thesis
Whoever reads it positively will point to a rare combination: 17% sales growth with margin expansion on every line, earnings growing 30%, and guidance signaling a further acceleration to 28% earnings growth.
Beyond that: the optical communications segment rose 32% and enterprise networks 65% - a pace suggesting demand that is structural rather than cyclical. Free cash flow of $1.42 billion in a single quarter gives the company room to maneuver. And a return on invested capital rising 180 basis points says the growth is coming from existing assets, not from new investment yet to be proven.
And the strategic point: Corning is an infrastructure supplier, not a chip supplier. It does not need to guess which model will win or which AI company will survive - every data center built, by any player, needs fiber.
The Bear Thesis
Whoever reads it critically will point first to the warning management itself gave: a possible slowdown in capital spending by wireless carriers. That is a major customer of the optical communications segment, and it is the first time this season the company has flagged a headwind inside its main engine.
Second, concentration in a single story. A substantial part of the growth currently comes from data centers. If the pace of investment in AI infrastructure slows - an open question that has accompanied this entire season - this engine will slow with it.
Third, solar is losing money. A segment growing 90% but producing a net loss is for now a consumer of capital, not a contributor. The question is how long it takes to turn, and what it costs along the way.
Fourth, the traditional segments. Displays, environmental products and life sciences depend on the consumer cycle, on autos and on research budgets - and those do not enjoy the same tailwind.
And fifth, competition. Optical fiber is not an exclusive product; there are large manufacturers in East Asia, and price pressure is a permanent feature of the industry.
The debate in one line
The bulls see a veteran industrial company that has become an infrastructure supplier to artificial intelligence, with 65% growth in enterprise networks, margins expanding on every line and profit growing 30%. The bears see guidance below consensus, a warning about wireless-carrier spending, a solar segment still losing money, and a valuation that embedded a faster continuation. Both sides are reading the same report - and the market sided with the second one today.
My Angle
A personal opinion of Ilan Abramov - not advice, not a recommendation
What catches me in this report is who profits from artificial intelligence.
Most of the conversation around AI focuses on the top layer - the models - and on the layer beneath it, the chips. But between one chip and the next, something has to move the data. And that is exactly where a 175-year-old glass maker sits.
What I take from it: the difference between an infrastructure supplier and a solution supplier. Corning does not need a particular model to win, and does not need to guess which AI company survives the decade. Every data center built - by any player - needs fiber, connectors and optical infrastructure. That is broader exposure than a bet on a specific name.
And what sharpens it: the margins. 17% sales growth with 30% earnings growth says the company is not merely selling more - it is selling better products, to customers willing to pay. That is a sign of pricing power, not a volume race.
And the market reaction sharpens exactly that. The report was strong, and the stock fell by a double-digit percentage - because management flagged a possible slowdown in wireless-carrier spending. In other words: the market is not pricing the quarter, it is pricing the pace. That is a distinction worth remembering in every report, not just this one.
And what I will follow: the gap between the growth rate of enterprise networks and that of the segment as a whole. Right now it is 65% against 32% - meaning the "new" part is growing twice as fast as the old. As long as that gap holds, the thesis works. The day it narrows, it is worth asking whether that is because the old recovered - or because the new cooled.
Summary
Corning delivered a quarter with very few weak points in the numbers themselves: core sales of $4.74 billion (+17%), core EPS of $0.78 (+30%), margin expansion on every line - gross, operating and return on capital - and adjusted free cash flow of $1.42 billion. The main engine was optical communications, which jumped 32%, with a 65% surge in enterprise networks.
And yet the stock plunged by a double-digit percentage - because guidance for next quarter came in slightly below consensus, and more importantly, the company signaled a possible slowdown in spending by wireless carriers - in exactly the segment carrying the thesis. The other weak point is solar, which grew 90% but is still losing money.
And beyond the company, there are two data points here: the wave of investment in AI infrastructure keeps spilling outward from the chip layer to the physical connectivity layer - and at the same time, a stock price reflects expectations, not results. The question for the investor is not whether the demand exists - the report answers that - but how long it continues at this pace, and how much is already in the price.
Sources: Corning's official results announcement for the second quarter of 2026 (July 28, 2026), including core and GAAP sales and earnings, margins, return on invested capital, cash flow, segment breakdown and third-quarter guidance; third-quarter analyst consensus and coverage of the market reaction to the report as reported by Investing.com. Stock-move figures are accurate as of the time of writing and change during the session. The chart is shown in real time via TradingView. Nothing herein constitutes a forecast, recommendation or advice - see the full disclaimer at the bottom of the page.