Free Cash Flow Tripled and Enterprise Optical Grew 65%, Yet the Stock Sits 46% Off Its High: Upgrading Corning

Results vs. Consensus

Versus the Company's Own Guidance

This is the cleanest print of the SpringBoard era on the metrics that matter, and it came in above both the Street and a guide the company set only three months ago while absorbing the full $30M of solar-wafer shutdown cost it had flagged. A note on comparability before the tables: effective April 1, 2026, Corning replaced its constant-currency core adjustment with an adjustment for hedged exposures and did not recast prior periods, so the core-basis year-over-year comparison is not strictly like-for-like. The distortion is immaterial to the growth rate. GAAP sales grew 16.7% against core's 17.1%, so the basis change moves the top-line growth figure by well under half a point.

Year-over-Year Comparison

Sequential Comparison

Quality of Beat

• Revenue: high quality, but narrower than the headline. Optical added $506M of year-over-year sales and Solar added $207M. Everything else, Glass Innovations plus Automotive plus Life Sciences, went backwards by a combined $20M. So the entire 17% consolidated growth rate, and then some, came from two segments. That is not a criticism of the print (those are the two segments the thesis is about) but it does mean the company's growth rate is now effectively a levered bet on AI optical demand and U.S. solar policy, with a flat-to-shrinking $2.2B remainder attached.
• Margins: real operating leverage, plus a bigger addback stack. The genuine leverage is visible in operating expense, where core SG&A and RD&E together grew 13.2% against 17.1% sales growth, taking combined opex from 19.3% to 18.7% of core sales. But GAAP operating margin expanded only 70 bps against core's 190 bps, and much of that gap is a heavier set of non-currency addbacks: restructuring, impairment and other charges added back went from $1M of pretax impact a year ago to $71M this quarter, with pension mark-to-market up as well. Read core operating margin as directionally right and roughly half as flattering as it looks.
• EPS: operationally driven; ignore the GAAP tax rate. Core operating income rose $219M and the core tax rate improved 100 bps to 18.5%, against a 1.2% increase in the diluted share count to 875M. That accounts for essentially all of the $0.18 of core EPS growth. The GAAP figure of $0.64 was flattered by a 6.2% effective tax rate (against 14.4% a year ago) on $79M of discrete tax benefits, which the core reconciliation correctly strips out. Non-controlling interest, largely Hemlock's minority, rose to $50M from $31M.

Segment Performance

Corning reports five segments following the Q1 2026 restructure that carved Solar out on its own and merged Display with Specialty Materials into Glass Innovations. Segment sales sum exactly to consolidated core sales, so the mix below is complete rather than indicative.

Optical Communications

This is the whole story, and it was better than the segment line suggests. Sales of $2.072B grew 32% and net income grew 77% to $438M, but the composition is what matters: enterprise grew 65% year over year to $1.27B, adding roughly $300M sequentially, while the AI-data-center portion of enterprise nearly doubled. Segment net income margin reached a record 21.1%. Since SpringBoard began, enterprise sales have more than tripled.

The single most useful sentence on the call was management's reminder of what is not in these numbers.

"In quarter 2, we grew sales 65% year-over-year to $1.27 billion. And our Gen AI product sales nearly doubled. Keep in mind, this is all just scale out. We're not yet seeing scale up or photonics in our results." — Wendell Weeks, Chairman & CEO

Assessment: A 65% growth rate with a record margin, entirely from the mature half of the opportunity, is a stronger result than the 32% headline implies. It also reframes the bear case: the risk is no longer whether the enterprise ramp is real, it is whether the scale-up and photonics layers that management has now sized arrive on the assumed schedule. Those are different risks, and the second one is priced far more cheaply than the first was.

Carrier (within Optical)

Carrier is the quarter's genuine disappointment. Having grown 36% in Q1, it grew 1% in Q2, leaving it at roughly $0.8B and down sequentially. Management attributed the swing entirely to project timing and pointed to the half-year figure rather than the quarter.

"If you look at the first half sort of this year of 2026 versus the first half of 2025, sales are up about 17% in Carrier, about $0.25 billion." — Wendell Weeks, Chairman & CEO

The longer-term framing was also trimmed. Where the Q1 call described a fiber-to-the-home recovery running well ahead of trend, the guidance now offered is mid-single-digit growth driven by fiber-to-the-home and data center interconnect.

Assessment: The half-year math checks out and carrier is a lumpy, project-timed business, so the explanation is credible. But a swing from +36% to +1% in one quarter is not noise, and the mid-single-digit long-term frame is a step down from April's tone. We treat carrier as a roughly flat contributor for modelling purposes and give the growth credit entirely to enterprise. The good news is that this is exactly how the guide already behaves.

Glass Innovations

Sales of $1.463B grew 1% while net income grew 9% to $354M, lifting segment net-income margin to 24.2% from 22.8% in Q1. Higher display glass sales carried the quarter, helped by an unusual second-order effect: as memory costs rise, TV brands and panel makers are shifting toward larger, higher-priced sets, which favors Corning's Gen 10.5 position. The same memory dynamic works against the company in handsets, where management now expects full-year units down a mid-teens percentage.

"For the full year, we expect memory prices to impact the handheld market with units to be down a mid-teens percentage. Despite these headwinds, we expect Gorilla Glass sales to outperform the end market driven by strong demand for our innovations and our position in the premium segment of the market." — Edward Schlesinger, CFO

Assessment: A 24% net-income margin on a flat top line is precisely the ballast this equity's growth investors ignore, and the display/handset offset means the memory shock is closer to neutral for the segment than the headlines imply. The advanced-optics call option on semiconductor and memory capex is the underappreciated piece here. We model Glass Innovations flat on revenue with slowly rising margin, and treat it as the funding source rather than a growth vector.

Solar

Sales grew 90% to $438M and the segment posted a $7M net loss, a modest deterioration from Q1's $7M profit that reflects the fully absorbed $30M of incremental shutdown cost management had guided to in April. The important development is that the shutdown is finished: the wafer facility has moved to permanent power and its production equipment has been repaired and upgraded.

"In Q2, as expected, we experienced an additional $30 million of expense versus Q1 as our solar wafer factory underwent an extended maintenance shutdown transitioning to a permanent power system while repairing and upgrading production equipment. Customer demand is strong across the map and we expect sales and profit to improve beginning in the third quarter." — Edward Schlesinger, CFO

Management reiterated that Corning remains the only U.S.-based polysilicon and wafer manufacturer and that it is "firmly on track to build our solar business into a $3 billion revenue stream with profitability above the corporate average."

Assessment: The execution crack we flagged in April as widening has been closed on schedule and at the cost management quoted. That matters more than the $7M loss: it converts a "management said it would get better and could not say when" item into a "management said it would get better by Q3 and did the work" item. Solar goes from the reason to worry about execution to a Q3 tailwind, and the $3B target with above-corporate margin is now the thing to hold them to.

Automotive and Life Sciences

Automotive grew 2% to $471M with net income up 4% to $82M, outperforming a global vehicle market down 2% on the strength of the More Corning content strategy; diesel grew 3% year over year and 13% sequentially on improving North American Class 8 orders. Life Sciences and Emerging Growth Businesses was the one segment in outright decline, with sales down 15% to $294M and a $21M net loss against $6M of income a year ago. Management addressed it in a single sentence and framed it sequentially, noting sales up 8% and net income up 13% quarter over quarter on strength in the life sciences research business.

Assessment: Automotive is doing its job as a stable, content-driven share gainer against a shrinking market, and the Class 8 sequential improvement is a small positive. Life Sciences is the segment nobody is asking about, which is roughly correct at 6% of sales, but a 15% decline and a widening loss deserve more than one sentence framed on the flattering axis. It is not thesis-relevant today; it would become so if the loss keeps widening.

Key Topics & Management Commentary

Overall Management Tone: Management was confident and, for the first time in the SpringBoard era, largely in recap mode, with roughly half the prepared remarks re-presenting the May investor day rather than breaking new ground. On the quarter itself the posture was assured and the numbers carried the argument, but on the two questions the market actually had, whether the Q3 guide implies deceleration and whether scale-up timing has slipped, the answers were framework-level rather than quantified. The tone was steadier than April's mix of excitement about optical and apology about solar, and conspicuously unbothered by a stock that had already given back nearly half its value.

1. The Q3 Guide, and Why a Beat Lost 12%

The proximate cause of the sell-off is arithmetic. Corning guided Q3 core sales to $4.9B to $5.0B, about 16% growth, against a Street bar sitting at roughly $5.0B. A midpoint roughly 1% below consensus was enough to take 12% out of the stock. Two separate analysts pressed the point, one asking why Q4 implies so little growth off a $5B run rate, the other stating flatly that the midpoint implies a point of deceleration.

"Our guide for Q3 is not intended to imply any deceleration in growth. I think it's similar -- it's intended to imply similar year-over-year growth as we had in Q2. And then as we've shared longer term, we expect that growth rate to actually accelerate. So that's sort of how you should think about it in total. And for sure, enterprise growth is a significant component of that as you go into Q3." — Edward Schlesinger, CFO

The more revealing answer came on why the plan was not simply raised again. The Q3 guide midpoint of $4.95B annualizes to $19.8B, and the high end annualizes to exactly the $20B run-rate target management set for year-end 2026. In other words, the guide the market punished puts Corning at its own year-end goal a quarter early.

"in May, we provided that we get to the $20 billion run rate by quarter 4. And our growth rates are just going better. And so it looks like we could hit it a quarter early, and we didn't want to upgrade our springboard plan again after just doing it in May. So when we get around to updating and providing quarter 4 guide, we'll do that. Meanwhile, there's no implied message." — Wendell Weeks, Chairman & CEO

Assessment: A company that hits its annual run-rate target one quarter early and declines to re-raise a plan it upgraded eleven weeks ago is being conservative, not decelerating. The guide is a floor, and management said so about as plainly as a CFO can. The 12% reaction is a statement about who owns the stock, not about what the business is doing.

2. What Is Not Yet in the Numbers

The most important framing on the call was about absence. Every dollar of the 65% enterprise growth, the record 21.1% optical margin, and the nearly-doubled Gen AI product sales comes from scale-out, the mature portion of the data-center opportunity. Scale-up optics and the entire Photonics platform, the two things the May investor day was built around and the two things carrying the 2028 to 2030 half of the plan, contribute nothing to the reported figures yet.

Assessment: This cuts both ways and investors are currently only pricing one of them. It means the near-term numbers are not dependent on an unproven technology transition, which de-risks 2026 and much of 2027. It also means that the part of the story generating the multiple has produced no revenue and can therefore only disappoint on timing. The correct response is to underwrite the scale-out business on its own merits and treat scale-up as a free option, which is roughly what the current price does.

3. Free Cash Flow Inflects and the Customer Funding Appears

Adjusted free cash flow was $1.423B against $451M a year ago, and $1.611B for the half against $450M. The mechanism is visible on the face of the cash flow statement: customer deposits and government incentives contributed $880M in the quarter, against $59M in the year-ago quarter. Operating cash flow was $1.717B against $708M. Capital expenditure of $422M was comfortably covered several times over.

"Typically, when we invest organically, we invest significant amounts of capital upfront which means we take risk before the revenue and free cash flow shows up. ... So the results will be attractive and we expect free cash flow to grow even as we invest to capture the higher sales." — Edward Schlesinger, CFO

Assessment: This is the quarter's most important disclosure and it received almost no attention in the reaction. For two quarters we have written that the risk-sharing claim was credible but unproven in reported cash flow. It is now proven, in a single line item, at a scale ($880M) that is nearly half the entire year's capital budget. A growth business whose customers prepay for capacity deserves a different multiple than one that funds capacity off its own balance sheet, and the market spent the day trading the revenue guide instead.

4. Capex Rises to $2 Billion

Management raised 2026 capital expenditure to approximately $2B from the roughly $1.7B guided in April, with the investment run rate stepping up through Q3 and Q4. First-half capex was $754M, so the back half carries roughly $1.25B.

"For CapEx, we expect to increase our investment run rate into Q3 and Q4, and to invest approximately $2 billion for the year to support the compelling growth plan Wendell just described in optical communications." — Edward Schlesinger, CFO

Assessment: An 18% capex increase would be a straightforward negative in isolation, and in April we would have treated it as one. Read against the $880M of customer deposits in the same quarter, it is the opposite: capacity is being added because customers are contracting and paying for it, and the incremental spend is being pre-funded rather than levered. The line to watch is whether the deposit inflow keeps pace with the capex step-up in Q3 and Q4, not the capex number by itself.

5. Scale-Up, Quantified: From 16 Fibers Per GPU to 160

The technical core of the call, carried over from the May event, was a bracketing exercise on how much optical content a scale-up transition adds. Today the scale-up network is entirely copper and Corning's content is 16 fibers per GPU, all of it in scale-out. Using the announced 14.4 terabits per second of scale-up bandwidth and 1.6 terabits per second of scale-out bandwidth against a 200G SerDes, a fully optical scale-up system would require 144 fibers for scale-up plus 16 for scale-out.

"When we combine these demands, we get a total fiber content of 160 fibers per GPU, which is 10x the amount of fibers of the current scale out network." — Wendell Weeks, Chairman & CEO

Management was careful to say the real answer is neither bookend, since the recently announced Vera Rubin Ultra configuration is a hybrid: 576 GPUs across eight racks, interconnected with copper within the rack and extended rack-to-rack optically. Combining cluster-size growth (clusters above 130,000 GPUs require a third optical layer, worth 50% more content), bandwidth (assumed neutral in the plan), and scale-up, management sized optical content per GPU rising 1.3x to 1.5x by 2028.

Assessment: This is the quantification we said in April was the swing factor, and it is more rigorous than we expected: an explicit bookend, an explicit hybrid caveat, and a plan built on the conservative end of bandwidth. The 1.3x to 1.5x content figure through 2028 is a modelable number that assumes only partial scale-up penetration. The genuine unknowns, the optical port percentage and the adoption slope, are correctly identified as unknowable rather than papered over.

6. Photonics as a $10 Billion Platform, and a Pushback on the Street's Math

The Photonics platform, which sells passive optical components inside the box for co-packaged and near-package optics, is targeted at a $10B revenue stream by 2030. Corning has historically had zero inside-the-box content, so this is a genuinely new market rather than a share-shift.

"Based on our assumptions and our discussions and agreements with customers, we believe we have the opportunity for a $10 billion market access platform by 2030." — Wendell Weeks, Chairman & CEO

What was notable was the pushback on the Street's near-term interpolation. Asked whether Photonics could still be a billion-dollar business next year, management declined the premise directly: "I don't think we guided anything for next year on Photonics, just to be clear."

Assessment: Correcting an over-eager Street model downward on a day the stock is already falling is the behaviour of a management team more interested in being right than in being liked, and it is a point in their favour. It also removes a 2027 disappointment that was being built into models by people other than the company. The absence of any annual waypoint between here and $10B in 2030 remains the largest single gap in the disclosure.

7. Two Plans: 20-30-40 Internal, 27/35 High Confidence

Management now maintains two published trajectories. The internal SpringBoard plan targets an annualized sales run rate of $20B by end-2026, $30B by end-2028 and $40B by end-2030, built bottom-up by each market access platform at a stated 70% confidence interval and used to set objectives and compensation. The corporate high-confidence plan, risk-adjusted for macro, policy and adoption timing, targets $27B by end-2028 and $35B by end-2030.

"Our high confidence plan is to grow sales to an annualized run rate of $27 billion by the end of 2028 and $35 billion by the end of 2030. Interestingly, in either case, we expect to double the size of the company by the end of 2030." — Edward Schlesinger, CFO

Management was explicit about where most of the risk adjustment goes: "One of the most significant areas we are adjusting for is the timing on Scale-Up of the network. This impacts both enterprise and photonics."

Assessment: Publishing both a compensation-linked plan and a deliberately haircut investor plan, and naming the specific variable that separates them, is unusually honest disclosure. It also hands investors the bear case in the company's own words: the $5B gap in 2030 is scale-up timing. Note that both plans double the company by 2030, differing only in the starting line. We model to the high-confidence plan and treat the internal plan as upside.

8. NVIDIA and Amazon, and Long-Term Agreements Becoming the Default

The two unnamed hyperscaler agreements from the Q1 call now have names attached. In May, NVIDIA entered a multiyear commercial and technology partnership to expand U.S.-based manufacturing of advanced optical connectivity, targeting a tenfold increase in U.S. optical connectivity capacity and more than a 50% increase in U.S. fiber production. In June, Amazon announced a multiyear, multibillion-dollar agreement covering fiber, cable and connectivity for its U.S. data-center build. These sit alongside the up-to-$6B Meta agreement from Q1 and the Apple cover-glass commitment.

Asked what share of optical now sits under long-term agreements, management said such deals will be "the lion share" of the optical business, because every major capacity expansion is now underpinned by one.

"We create value by delivering innovations in fiber cable connectivity and to dramatically reduce our customers' cost to increase their speed of deployment or increase the reliability of their networks. And then when we create value for our customers with our innovations, we get to keep a portion of that value of the innovation. And that is what's driving this really significant increase in profitability that you see. I would expect that to continue." — Wendell Weeks, Chairman & CEO

Assessment: Two of the three watch items from January and April are now closed: the customers are named and the funding is visible in cash flow. The margin explanation is also the right one. A 77% increase in optical net income on a 32% increase in sales is not a bare-fiber pricing story, it is a mix shift into proprietary high-density products, which is far more durable than a cyclical price spike. The remaining "trust us" is deal-level economics, which stays undisclosed.

9. Supply Remains the Constraint

On pricing and capacity, management described a demand environment that continues to exceed what Corning can produce, particularly in the newest products.

"We continue to have the enviable situation of if we could make more, we could sell more. And especially if we could make more of our most innovative products, we see nothing but growing appetite for these new high-density product sets that we started on so long ago that are definitely gaining and accelerating in their adoption rate." — Wendell Weeks, Chairman & CEO

Assessment: A supply-constrained business with contracted, pre-funded expansions and accelerating orders is about as good a setup as an industrial gets. It also explains why the capex increase and the deposit inflow arrived in the same quarter. The risk in this framing is that "if we could make more, we could sell more" is precisely what every supplier says at the top of a capex cycle, which is why the customer prepayments matter so much: they convert the claim into an obligation on the customer's side.

10. The Operating Margin Target, Deferred a Third Time

Corning has now run at or above its 20% operating margin target for three consecutive quarters and printed 20.9% this quarter, and management again declined to replace the target it cleared a year early.

"You can expect us to continue to run at or above 20% operating margin even as we continue to invest to capture all of the growth. We'll come back later this year to give you an update on how we're thinking about operating margin." — Edward Schlesinger, CFO

The reasoning offered was that solar profitability is still ramping and the Photonics business has yet to be built, and management wants more experience with both before resetting the anchor. ROIC guidance was given more freely: approximately 15% today, expected to reach the high teens through the planning cycle.

Assessment: Three deferrals is a pattern, and it is the one place management is consistently less forthcoming than the business justifies. The reasoning is defensible and the ROIC path partially substitutes, but a company promising earnings growth faster than a 19% sales CAGR through 2030 is implicitly promising margin expansion, and refusing to quantify it leaves the most valuable part of the model to investor inference. This is now our top disclosure ask.

11. What the Plan Assumes

Management disclosed the key inputs behind the 2027 to 2030 internal plan, which is useful for stress-testing. The plan assumes a forward rate of ¥150 per U.S. dollar; flat TV, IT and smartphone end markets; the impact of higher memory prices; declining internal-combustion demand offset by rising Corning auto content; an upgraded solar outlook; and accelerating fiber-to-the-home and data center interconnect in carrier. In aggregate, consumer electronics, solar, carrier, automotive and life sciences are planned at a mid-single-digit combined CAGR.

Assessment: These are conservative assumptions, and that is the point worth extracting. Flat consumer end markets and a weak yen are the kind of inputs that produce a plan you can beat rather than defend. It also confirms the arithmetic of the story: if the non-optical two-thirds of the company compounds at mid-single digits and the whole compounds at 19%, optical and photonics are carrying essentially all of the incremental growth. That is the bet, stated plainly.

Guidance & Outlook

The Longer-Dated Plan

Implied ramp: The Q3 guide midpoint of $4.95B annualizes to $19.8B and the high end to $20.0B, which means Corning reaches its end-2026 run-rate target in Q3 rather than Q4. Against the Q4 2023 SpringBoard starting point of $13B, Q2 2026 represents a 45% increase in sales, a 460 bps improvement in operating margin, a doubling of EPS and a 610 bps expansion in ROIC. The first SpringBoard phase, through Q4 2026, delivers a 15% sales CAGR; the plan calls for that to step up 400 bps to 19% for the 2027 to 2030 phase.

Street at: Full-year 2026 consensus core EPS sat at $3.18 into the print. That figure now looks slightly stale: H1 core EPS is $1.47 and the Q3 guide midpoint is $0.87, which would require a Q4 of roughly $0.84 to land on $3.18, below Q3 despite Q4 seasonality and improving solar profitability. We use $3.20 to $3.25 for 2026 and $4.22 for 2027.

Guidance style: Conservative, and demonstrably so. Corning beat the high end of both its Q2 sales and EPS guides while absorbing the full solar cost it had flagged, and it is now hitting its annual run-rate target a quarter early while explicitly declining to raise the plan again. Treat the Q3 guide as a floor.

Analyst Q&A Highlights

Why the fourth quarter looks flat off a $5 billion run rate

The call opened on the shape of the second half. The questioner noted that optical capacity is ramping and the solar migration is complete, then asked why the implied fourth quarter shows so little growth off the third-quarter exit rate, and whether something in the other end markets explained it. Management's answer reframed the guide as an early achievement rather than a ceiling, and pointed to the acceleration built into the plan.

Q: "Why should 4Q, assuming an exit run rate of $5 billion a year, why is there just very limited 4Q growth? Is there something in the other markets that we should be thinking about?"
— Asiya Merchant, Citi

A: "In the third quarter, our guide implies that we might actually get to that $20 billion that we put out for the end of this year, a quarter or so early, and then we expect to continue to grow from there. And I think most importantly, we've been growing at a mid-teens growth rate year-over-year, call it, 15% or so. We're starting to see that growth rate accelerate, and we actually expect our growth rate from the end of this year through the end of the planning cycle to also accelerate up to closer to 20%, our CAGR is 19%."
— Edward Schlesinger, CFO

Assessment: The most consequential exchange of the call, and the market ignored the answer. Management did not defend the guide, it explained that the guide already delivers the annual target early and that the plan calls for acceleration from there. Anyone who reads "no implied message" as evasion should note that the same team upgraded the plan eleven weeks ago and declined to do so again on principle.

Whether anything changed since the May investor event

Given the volume of industry chatter about supply-chain readiness and potential delays in optical scale-up, the most direct question of the call asked whether management would characterize anything differently than it had two months earlier, and whether the near-term Photonics visibility the Street had modeled still held. Management reaffirmed its fundamental view without qualification, then explicitly declined to endorse the near-term Photonics number the questioner attributed to it.

Q: "To be crystal clear here, I mean, has anything changed now versus 2 months ago? And specifically thinking about photonics and Scale-Up just -- there's been a lot of industry chatter around supply chain readiness for this and potential delays. I mean it seems like you're reiterating that Photonics could be $1 billion-plus business potentially next year. do you still have your visibility to that?"
— Joshua Spector, UBS

A: "So what you're hearing from us is we see no changes to our fundamental belief that we just shared with you in May. Yes, there has been much industry speculation on the exact timing of when everything begins and that's fair enough. ... We're doing the steady work that it takes to build a $10 billion new photonics platform and bringing a significant new technical node to overall AI, which is scale up, which is where the biggest opportunity is for us. And it's just inside the ecosystem, there's just way less change than from what sort of gets speculated on."
— Wendell Weeks, Chairman & CEO

Assessment: Two things happened here and only one was noticed. The plan was reaffirmed in full against a wall of industry speculation, and a Street model that had run ahead of company guidance was corrected downward. The second is the sort of discipline that reduces future disappointment risk, and it cost management something to do on a down day.

How much of optical sits under long-term agreements, and what drives pricing

A line of questioning probed the share of optical revenue covered by long-term agreements and whether Corning is raising prices for customers outside them. Management said such agreements will become the overwhelming majority of the optical business because every major capacity expansion is underpinned by one, then rejected the framing that pricing drives the margin improvement.

Q: "I was just curious about the mixture of the optical business that's covered under long-term agreements. I'm wondering what that looks like right now. I'm wondering what that looks like maybe a year from now. And then I guess I assume you guys are going to be more aggressive on pricing in the optical business for customers that are not under LTAs."
— George Notter, Wolfe Research

A: "We create value by delivering innovations in fiber cable connectivity and to dramatically reduce our customers' cost to increase their speed of deployment or increase the reliability of their networks. And then when we create value for our customers with our innovations, we get to keep a portion of that value of the innovation. And that is what's driving this really significant increase in profitability that you see. I would expect that to continue."
— Wendell Weeks, Chairman & CEO

Assessment: The distinction being drawn matters for how the margin should be capitalized. Cyclical fiber pricing mean-reverts; value capture on proprietary high-density products does not, at least not on the same clock. A 77% increase in segment net income on a 32% increase in sales is more consistent with the second explanation than the first, and the answer is a commitment management can be held to when fiber pricing eventually normalizes.

Whether the third-quarter midpoint implies deceleration

The deceleration question came back in more pointed form, paired with a question about why a roughly $300M sequential increase in enterprise sales did not produce more visible operating leverage in the segment. Management rejected the deceleration reading outright and attributed the quarter's segment margin behaviour to ongoing mix and investment.

Q: "At the consolidated level, at the midpoint, you have a point of deceleration implied in your Q3 guide? And I hear all the positives around your longer-term story around reacceleration. Q2 was obviously underpinned by very strong optical enterprise strength. So the question is really, do you expect that strength to continue?"
— Wamsi Mohan, Bank of America

A: "So our guide for Q3 is not intended to imply any deceleration in growth. I think it's similar -- it's intended to imply similar year-over-year growth as we had in Q2. And then as we've shared longer term, we expect that growth rate to actually accelerate. So that's sort of how you should think about it in total. And for sure, enterprise growth is a significant component of that as you go into Q3."
— Edward Schlesinger, CFO

Assessment: As direct a denial as a CFO gives. The leverage half of the question is the more interesting one and was answered less well: optical net-income margin rose only 10 bps sequentially on a 12% sales increase, which is worth watching if it repeats. One quarter of flat incremental margin at a record absolute level is not a trend.

The implied margin step-up into the third quarter

Working backwards from the Q3 revenue and EPS guides, one questioner noted that the implied gross and operating margins step up meaningfully and asked what drives it and whether it is sustainable, given management is still nominally anchored to a 20%-plus target. Management confirmed the arithmetic and named solar as the driver, then explained the third consecutive deferral of a new margin target.

Q: "When I do the rough back of the envelope in half on the 3Q guide, it implies a pretty nice uptick in both gross and operating margins and a nice improvement in the associated incrementals there. First, is that fair directionally? And then second, if so, what are the moving pieces driving the Step-Up into 3Q? And how should we think about the sustainability there?"
— Joseph Cardoso, JPMorgan

A: "I think your assessment is correct. And one thing I would call out is we expect our solar business to improve profit and profitability from Q2 to Q3 as we've gotten past this cycle of improving our manufacturing performance there. ... We haven't set a new target. We'll come back later this year. ... we're still ramping that solar business, so that's not fully done. So profitability will continue to improve. We want to see that. We want to get a little bit more of that behind us. And then we have to build a very significant photonics business."
— Edward Schlesinger, CFO

Assessment: The confirmation that Q3 margins step up is a bigger disclosure than it sounds, because it means the guide's roughly 28% EPS growth against roughly 16% sales growth is structural rather than a tax or share-count artifact. The margin-target deferral is the frustrating half: the reasons are legitimate, but the company has now cleared its old target for three straight quarters without replacing it.

The carrier deceleration

Carrier's fall from 36% growth to 1% drew a question about whether it reflected difficult comparisons against customers ramping a year earlier, a reallocation toward enterprise, or something structural, given management's stated expectation of mid-single-digit growth. The answer was project timing, supported by half-year rather than quarterly math.

Q: "Maybe just on the carrier business for a second. Is that a reflection of just challenging comps in terms of some of the newer customers that were ramping last year. Is that reallocation of some of the enterprise share? Just trying to get a little bit of insight into this quarter's results versus kind of your indications that, that would still be mid-single-digit growth."
— Meta Marshall, Morgan Stanley

A: "If you look at the first half sort of this year of 2026 versus the first half of 2025, sales are up about 17% in Carrier, about $0.25 billion. So just the way Carriers work, what project is happening when in our timing of what we're doing with that particular customer. But so we're not seeing anything in Carrier where we're not -- where we don't see a growing set of demand, both in DCI, as you mentioned, but also in fiber to the home."
— Wendell Weeks, Chairman & CEO

Assessment: The half-year figure is verifiable and supports the timing explanation, so the answer is credible rather than deflective. What went unaddressed is that a reallocation of scarce capacity toward higher-value enterprise demand would produce exactly this pattern, and would be a good outcome the company has no incentive to describe. Either way, carrier should be modeled flat to mid-single-digit, not extrapolated from either quarter.

How much scale-up content is actually in the plan

The closing exchange sought to pin down how much of the scale-up and fiber-array-unit opportunity is already embedded in the SpringBoard numbers rather than sitting as upside. Management clarified the boundary between the two maps, fiber to the faceplate counting in enterprise and everything inside the faceplate counting in Photonics, and confirmed that the $10B Photonics figure is an assumption-driven baseline whose slope nobody can yet call.

Q: "So the $10 billion baseline assumption does include some of the opportunities highlighting the yellow color on Slide 26. But there could be upside or this is just based on the assumption, we don't know the slope of the adoption?"
— Mehdi Hosseini, Susquehanna

A: "Yes, a lot that depends -- the biggest driver here is going to be, if you're on the right question, which is what percent of the ports are optical, right? And then how successful is optical Scale-Up in AI factories. And that dynamic you're just seeing you're going to have 2 parts to Scale-Up, which is why we pay such close attention to it, which is both the dramatic increase in our enterprise products, as well as the dramatic increase in our inside-the-box or photonics products."
— Wendell Weeks, Chairman & CEO

Assessment: The clean answer to the most important modelling question. Scale-up shows up twice, once as more fiber in enterprise and once as new content in Photonics, which is why the two plans diverge on a single variable. It also confirms the $10B figure is a probability-weighted assumption rather than a bookend, meaning the distribution around it is genuinely two-sided rather than a stretch target dressed as a plan.

What They're NOT Saying

1. Any new operating-margin target: Deferred for a third consecutive quarter, now to "later this year," after three straight quarters at or above the old 20% target and a 20.9% print. A company promising earnings growth faster than a 19% sales CAGR is implicitly promising margin expansion it will not quantify.
2. Fourth-quarter or full-year guidance: Asked directly why Q4 implies little growth off the Q3 run rate, management said there is "no implied message." Corning still guides one quarter at a time while asking investors to underwrite a plan that runs to 2030.
3. What percentage of scale-up ports are optical: Explicitly withheld as confidential. This is the single variable that decides whether enterprise content is 16 fibers per GPU or 160, and management confirmed it knows the answer and will not share it.
4. Any annual Photonics waypoint: Management declined to validate the roughly $1B-next-year figure the Street had built off the May event, and has provided no revenue marker at all between zero today and $10B in 2030. The correction is welcome; the vacuum it leaves is not.
5. The $500 million of warrant proceeds: Second-quarter financing activities show $500M of proceeds from warrants with no counterpart in the prior-year period, alongside additional paid-in capital up roughly $1.1B over six months. Neither the release narrative nor the call addressed what these instruments are or who holds them.
6. Why buybacks stopped: Corning repurchased no common stock in the first half of 2026, against $133M in the first half of 2025, while generating $1.6B of adjusted free cash flow and paying $495M of dividends. In April, buybacks were described as the primary vehicle for returning excess cash. The change in posture went unmentioned.
7. Deal-level economics on NVIDIA and Amazon: Both partners are now named, which closes a genuine gap. Neither dollar value, duration, volume commitment, nor the split between prepayment, take-or-pay and guaranteed revenue is disclosed for either.
8. Life Sciences and Emerging Growth: Sales down 15% year over year with a $21M net loss against $6M of income. It received one sentence, framed on the sequential axis where the numbers are positive rather than the annual axis where they are not.

Market Reaction

• Pre-print setup: GLW closed at $143.36 on July 27, up 63.7% year to date and up 158.7% over the trailing twelve months, but down 35.1% over the trailing 30 days and already 43.9% below its June 29 closing peak of $255.69. The stock entered the print having given back roughly five months of gains.
• Reaction day (before-open print, same session): shares gapped down 15.1% to open at $121.75, traded as low as $114.50 (down 20.1%), and closed at $126.01, down 12.1% or $17.35. Volume was 46.2M shares, 3.1x the 30-day average. The close was 10.1% above the intraday low. The S&P 500 rose 0.2%, so the move was entirely stock-specific.
• Following sessions: down 1.6% to $124.05 on July 29, then up 9.0% to $135.22 on July 30 and up 2.2% to $138.25 on July 31. Three sessions after the print the stock had recovered 9.7% off the reaction close and sat 3.6% below the pre-print level.
• Read-through: the guide dragged the wider optical complex with it, with several peers in fiber, optics and photonics components falling more than 10% on the day, and photonics-focused funds selling off alongside.

The trigger was 1% of one quarter's revenue guide. That is the whole of it: $4.95B at the midpoint against a bar at roughly $5.0B, on a print that beat consensus on both lines, beat the company's own guide on both lines, tripled free cash flow and set a record optical margin. What made a 1% shortfall worth 12% was everything that had already happened before July 28.

Entering the print, GLW had fallen 35% in thirty days and 44% from its June closing peak, having compounded at 159% over twelve months. That is a positioning unwind already in progress, and the guide arrived as confirmation rather than as news. The forward multiple had already compressed from roughly 69x at the June peak to the low 50s by mid-July. What the print did was hand a de-risking market a reason, and the reason it chose was the one number that was not obviously excellent.

Two features of the tape are worth more than the headline decline. The stock closed 10% above its intraday low on 3.1x normal volume, which is the signature of buyers arriving into a forced-seller flush rather than of distribution. And it recovered 9.7% over the following three sessions to close July at $138.25, within 4% of where it traded before the print. The 12% headline number is now the least accurate description of what happened. This is the fourth consecutive quarter in which Corning has sold off on a beat, but it is the first in which the sell-off was met with immediate demand, and the first in which the valuation being defended was not a demanding one.

Street Perspective

Debate: Is the third-quarter guide conservatism or the first sign of deceleration?

Bull view: The guide reaches the company's own end-2026 run-rate target a quarter early, management stated flatly that no deceleration is implied, and the same team declined to re-raise a plan it upgraded eleven weeks earlier out of discipline rather than caution. Corning beat the high end of both its Q2 guides; treat this one as a floor.

Bear view: A 16% growth guide against 17% delivered is not acceleration, and the 19% CAGR narrative requires a step-up that has not appeared in any guided quarter yet. Carrier just decelerated from 36% to 1%, which is what the beginning of a normalization looks like from the inside.

Our take: The bulls have this one on the evidence. The plan explicitly places the acceleration in the 2027-to-2030 phase, not in Q3 2026, so demanding it now is arguing with the calendar. Carrier is the legitimate concern, and the honest response is to stop giving it growth credit rather than to reject the guide.

Debate: Does optical scale-up arrive on the timeline the plan assumes?

Bull view: Corning has quantified the opportunity rigorously, sized content per GPU rising 1.3x to 1.5x by 2028 on conservative bandwidth assumptions, and secured named, funded, multiyear partnerships with the three parties that matter most to the outcome. The near-term numbers do not depend on it at all.

Bear view: Scale-up is 100% copper today, adoption curves for new interconnect technologies are notoriously back-loaded and prone to slipping years, and management itself says the timing is the largest single item in its risk adjustment. A $5B gap between the internal and high-confidence 2030 plans is management admitting it cannot call this.

Our take: Both descriptions are correct, and the disagreement is really about what the price already pays for. At roughly 33x 2027 estimates, the market is underwriting the scale-out business plus a modest photonics contribution. Scale-up on schedule is upside; scale-up two years late costs the 2030 target and not much of the 2027 number. That asymmetry is the reason to own it.

Debate: What multiple does a contracted 19% compounder deserve?

Bull view: Growth funded by customer prepayments, secured by long-term agreements, with 20%-plus operating margin, ROIC heading to the high teens and earnings compounding faster than a 19% sales CAGR is a materially higher-quality growth stream than a typical industrial. It should not trade at an industrial multiple.

Bear view: It traded at 69x forward five weeks ago and is still north of 40x on this year's numbers. The AI-infrastructure complex is de-rating broadly, the marginal buyer has left, and there is no reason a stock that has halved cannot halve again if the capex cycle turns.

Our take: This is where we have changed our mind. In April we argued a full valuation capped the upside, and we were right for one quarter. The de-rating has now done its work: roughly 43x this year's earnings and 33x next year's, for a business whose earnings are growing 30% and whose capacity is prepaid, is a different proposition from 48x on an unproven cash-flow story. The bears are describing a risk that has largely been paid.

Model Update Needed

Valuation impact: At the July 31 close of $138.25, GLW trades at roughly 43x our 2026 core EPS range of $3.20 to $3.25 and roughly 33x the $4.22 Street figure for 2027. Five weeks ago, at the $255.69 closing peak, the same business carried a forward multiple near 69x. Nothing about the plan has weakened in the interval; the plan was upgraded in May and reaffirmed in full in July, and the cash flow underneath it improved dramatically. We frame twelve-month fair value at $165 to $185, which is roughly 39x to 44x the 2027 number and implies 19% to 34% upside from that $138.25 close. That is a lower multiple than the stock carried at any point in the first half of 2026, applied to a larger and better-funded plan. The framework requires the 2027 acceleration to become visible in guidance during the next two or three quarters; if Q4 and Q1 guides continue to imply mid-teens growth rather than a step toward 19%, the multiple deserves to stay where it is and we would revert to Hold.

Thesis Scorecard Post-Earnings

Commitments delivered from April's watch list: Quantify and extend SpringBoard through 2030, delivered (20-30-40, plus a $10B Photonics platform and a sized scale-up opportunity). Set a new operating-margin target, missed for a third time (deferred to later in 2026). Bring the solar wafer plant back online and start reducing the drag, delivered (shutdown completed at the quoted cost; profit inflects in Q3). Land Q2 at or above the guide, delivered (above the high end on both lines). Disclose the unnamed hyperscaler customers, delivered (NVIDIA and Amazon). Show customer funding in reported free cash flow, delivered emphatically ($880M of customer deposits; adjusted FCF tripled). Five of six delivered, one missed for the third consecutive quarter.

Overall: The thesis strengthened on every operating axis while the valuation risk that has dominated the last two quarters was realized in full and, in the process, removed. Both bull pillars that carry the story were confirmed and the third was confirmed with the specific evidence we said we required. The stock is 10% cheaper than it was after the Q1 print despite two more quarters of delivery, a plan upgrade, two named hyperscaler partners and a cash-flow inflection.

Action: Add. We wrote in April that we would revisit Outperform on a meaningful de-rate or a quantified scale-up and Photonics step-up, and that we would move to Underperform if execution slipped again. Both upgrade conditions were met simultaneously and the execution question was answered on schedule. Upgrading to Outperform with conviction 7 rather than higher, because the acceleration to 19% has been described but not yet guided to, and because the AI-infrastructure de-rating that took 46% out of this stock is not obviously finished. We would move back to Hold if the next two guides imply mid-teens growth rather than a step toward the plan, or if the customer-deposit inflow fails to keep pace with the capex step-up in the second half.