Corning (GLW): A 47% Drawdown Into a Beat-and-Raise
Single-stock deep dive · Published Tuesday, July 28, 2026, pre-market (~7:30 a.m. ET) · @dailyanalysts · Last verified spot price: $143.36 (Mon Jul 27 close; Monday's intraday range $136.31–$148.69)
Highest-conviction finding: Corning just printed the single cleanest quarter of any AI-infrastructure supplier this season — core sales +17% to $4.738B, core EPS +30% to $0.78, core operating margin 20.9%, adjusted free cash flow $1.42B — and guided Q3 up to $4.9–5.0B / $0.85–0.89. The stock is 47% below its June 30 high of $271.78. The de-rating from ~82x to ~43x forward core earnings happened while the numbers accelerated. That is not a broken business; that is a broken multiple. I am flipping my stance from "don't chase the parabola" (my July 6 note at $196.79) to buy the de-rating.
1. What actually happened this morning (primary source)
I read the Q2 2026 earnings release in full, including the segment tables and both the Weeks and Schlesinger quoted commentary. Headline numbers:
| Core (non-GAAP) | Q2 2026 | Q1 2026 | Q2 2025 | Q/Q | Y/Y |
|---|
| Core sales | $4,738M | $4,345M | $4,045M | +9% | +17% |
| Core net income | $680M | $612M | $523M | +11% | +30% |
| Core EPS | $0.78 | $0.70 | $0.60 | +11% | +30% |
| Core gross margin | 39.6% | — | 38.4% | — | +120bp |
| Core operating margin | 20.9% | — | 19.0% | — | +190bp |
| Core ROIC | 14.9% | — | 13.1% | — | +180bp |
| Adjusted free cash flow | $1,420M | — | — | — | — |
Versus expectations: consensus was $0.7631 EPS on ~$4.654B revenue, and the company's own April guidance was ~$4.6B / $0.73–0.77. So Corning beat consensus by ~2% on EPS and ~1.8% on sales, and beat the top end of its own guide. Ninth consecutive quarter of year-over-year growth.
Segments — the entire story is one line item
| Segment | Q2 2026 sales | Y/Y | Q2 2026 segment net income | Y/Y | Segment margin |
|---|
| Optical Communications | $2,072M | +32% | $438M | +77% | 21.1% (vs 15.8%) |
| Glass Innovations | $1,463M | +1% | $354M | +9% | 24.2% |
| Automotive | $471M | +2% | $82M | +4% | 17.4% |
| Solar | $438M | +90% | −$7M | loss | −1.6% |
| Life Sciences & Emerging | $294M | −15% | −$21M | loss | −7.1% |
Inside Optical: Enterprise Networks +65% year over year, with "Gen AI product sales growing significantly faster." That is the AI data center line, and it is now 44% of company sales.
The number nobody will put in a headline: Optical Communications added $506M of year-over-year revenue and $191M of year-over-year segment net income. That is a 37.7% incremental margin on AI fiber and connectivity. Consequence #1: every incremental $1B of optical revenue from here drops roughly $375M to segment income — so the Meta, Amazon and NVIDIA volume ramps land on a structurally richer mix, not a diluted one. Consequence #2: this is why company core operating margin (20.9%) is already above the 20%-by-end-2026 Springboard target, one to two quarters early.
2. The guide is the whole trade — and it is the opposite of April
Management guided Q3 2026 core sales to $4.9–5.0B (+~16% y/y) and core EPS to $0.85–0.89 (+~28% y/y). Mid-point math: +4.5% sequential sales, +11.5% sequential EPS. CFO Ed Schlesinger also flagged Solar earnings turning from a drag to a contributor starting in Q3, "building toward a revenue stream of more than $3 billion."
This matters because the last two GLW prints were beats that the market sold on the guide:
- April 28 (Q1 2026): core sales $4.345B (+18%), core EPS $0.70 (+30%) — a clear beat — but Q2 was guided to only ~$4.6B / $0.73–0.77, dragged by an extra $30M of cost from an extended maintenance shutdown at the solar wafer facility (Q1 release, filed with the SEC). The stock fell ~7.7% that day.
- Today (Q2 2026): the shutdown is done, the drag reverses, and the guide accelerates in absolute dollars. The bear's April script — "beat on the print, miss on the guide" — has no page for today.
Also note Corning's own guidance conservatism: in April they guided ~$4.6B/$0.73–0.77 and delivered $4.738B/$0.78. Apply that same 2–3% cushion to the Q3 guide and the realistic Q3 outcome is ~$5.05B and $0.90. My estimate, not the company's.
The milestone hiding in the guide: Q3 mid-point of $4.95B annualizes to $19.8B. Springboard's "$20 billion annualized run rate by the end of 2026" is therefore effectively already achieved a quarter early — Q4 will clear it mechanically. That is the credibility bridge to the harder numbers: $30B annualized by end-2028 (a 22% CAGR from here) and $40B by end-2030, with a stated 19% sales CAGR Q4'26→Q4'30, earnings growing faster than sales. When a company hits the near milestone early, the market's discount on the far milestone is the opportunity.
3. Why the stock is down 47% — and which of those reasons survive contact with today's print
GLW closed at $143.36 Monday (−2.24%), after trading as low as $136.31 intraday. It set its 52-week high of $271.78 on June 30, 2026 and its record closing high near $255 in June. That is a 47% peak-to-trough drawdown, roughly 35% of it in the last month. Four causes, in order of how much they should actually matter:
| Cause | Real or sentiment? | Verdict after today |
|---|
| Valuation air (82x 2026 core EPS at the peak) | Real | Largely discharged — now ~43x 2026E, ~34x 2027E |
| Indiscriminate AI-capex de-rating (INTC beat-and-raised on Jul 23 and still fell 7.9%) | Real, but positioning-driven | Still live. This is the risk you are being paid to take |
| China DUV-lithography headlines hitting the whole "picks and shovels" complex Monday/Tuesday | Sentiment | Irrelevant to fiber. Nobody etches wafers with optical cable. Contagion, not fundamentals |
| Insider selling — ~15 open-market sells in six months, zero buys; CEO Wendell Weeks sold 100,000 shares at $186.46 on June 9 | Real | Insiders were right — they sold 30–45% above spot. But the signal has been paid out; watch for the first Form 4 buy |
On the market's mood: CNBC's Investing Club discussed GLW at Monday's morning meeting with the stock near $139, noting the "numbers should be pretty good on strong optical revenue driven by sales of hyperscalers" while simultaneously admitting "you could argue that we should just dump the rest" (CNBC, Jul 27). That is what capitulation into a good print sounds like.
4. Valuation: what you are actually paying
| Metric | At $143.36 (spot) | At $271.78 (Jun 30 peak) |
|---|
| 2026E core EPS ($3.30: $0.70 + $0.78 + $0.87 guide-mid + ~$0.95 est.) | 43x | 82x |
| 2027E core EPS (~$4.20, assuming 19% sales growth and EPS faster) | 34x | 65x |
| Adjusted FCF yield (Q2 $1.42B annualized on $125.5B cap) | ~4.5% | ~2.4% |
| Dividend ($0.28/qtr, declared Jun 24, payable Sep 29) | 0.78% | 0.41% |
| Reported trailing GAAP P/E | ~69x | ~131x |
2026 Q4 and 2027 EPS figures are my estimates and are labelled as such. Trailing GAAP P/E on $2.08 TTM EPS overstates the multiple you are underwriting because GAAP is depressed by hedge accounting, discrete tax items and restructuring — the release states this explicitly and the GAAP-to-core bridge is in the accompanying tables.
Sell-side targets set in the last four weeks — all struck when the stock was $180–250 — are BofA $243 (Jul 12), Citi $240 (Jul 13), Oppenheimer $230 (Jul 13), Truist $205 (Jul 1), with the published consensus clustering $195–215 and a $167–270 range (StockAnalysis). Opinion: I do not trade off consensus targets, and every one of these will be marked down if the tape stays hostile. I cite them only to show the gap between the sell-side's modelled fundamentals and the market's price — roughly 50%.
5. The contracted demand nobody is crediting
- Meta — up to $6B multiyear optical fiber and cable agreement announced January 27, 2026 (CNBC).
- NVIDIA — long-term partnership announced May 6, 2026: Corning to expand U.S. optical connectivity capacity 10x and U.S. fiber capacity by more than 50% (NVIDIA newsroom).
- Amazon — multiyear, multibillion-dollar fiber/cable/connectivity agreement announced June 8, 2026, with 1,000 new manufacturing jobs in North Carolina (Corning).
- Capacity being built: new optical connectivity plant in Stryków, Poland opening 2H 2026 (~2,500 jobs) for EMEA AI data centers.
Three of the five largest AI capex spenders on earth have signed multiyear supply agreements with this company in the last six months, and one of them (NVIDIA) is co-funding the capacity. Corning is not a company hoping AI capex shows up; it is a company whose customers pre-committed and whose cash conversion already reflects it ($1.72B GAAP operating cash flow in one quarter).
6. What I got right and wrong before this print
Scorecard callback. My July 6 GLW note (price then $196.79, TTM P/E 93.6, forward 106) flagged two things: the parabola and the insider selling — "~15 sells in 6 months, 0 buys; CEO Weeks sold 100k @ $186.46." That call has been correct: GLW is −27% since. My May 12 optical-complex note told readers not to chase AAOI at $185 and to wait for LITE at $850–900 and COHR below $320. Direction right, levels too shallow: LITE is $711.96 and COHR is $271.31 today — both blew through my "wait for" zones. Lesson I am applying here: in a de-rating tape, set entry zones on valuation floors, not on percentage retracements from the high. That is why my GLW entry band below is anchored to 40x 2026E core EPS, not to a Fibonacci level.
7. Three scenarios
Bull — 30%
Trigger: GLW holds a post-print gap and closes above $156.12 (the July 23 close) on above-average volume this week; FOMC holds Wednesday; MSFT/META (Wed) and AMZN (Thu) reaffirm or raise 2026-27 capex. Path: 40x my 2027E $4.20 = $168–175 inside 1–3 months, i.e. only a one-third retracement of the drawdown — this is not an heroic target. Second leg toward $195–210 if the sell-side re-anchors after the Q3 print in late October.
Base — 50%
Trigger: price holds Monday's $136.31 low but the gap gets sold; range trade $138–158 for two to four weeks while FOMC (Jul 29), core PCE and GDP (Jul 30) and mega-cap capex guidance resolve. Path: grind to $160–168 by the Q3 print. You get paid for patience, not for timing.
Bear — 20%
Trigger: the Fed hikes Wednesday (odds were repricing from ~11% to 32–38% over nine days on CME FedWatch into this meeting, with the 10-year at 4.68% — a cycle high) or a hyperscaler cuts capex guidance. Mechanism, quantified: at 43x 2026E, one turn of multiple is $3.30/share. A hike that re-rates high-multiple AI hardware from 43x to 36x takes GLW to $118–125 regardless of what Corning earns. That is a 15% further drawdown from a stock that has already halved, and it is why I am sizing this in two tranches rather than one.
8. The trades
1. LONG GLW HIGH CONVICTION
- Entry zone: $136–148 (first tranche, ~60% of intended size). Second tranche $126–133 only if the Fed-shock scenario delivers it — that is 38–40x 2026E core EPS.
- Target: T1 $172 (40x my 2027E core EPS of $4.20) in 1–3 months. T2 $196 (my July 6 reference price, ~46x 2027E) in 6–12 months.
- Invalidation (specific, measurable): two consecutive daily closes below $133 (i.e. decisively below Monday's $136.31 low). Fundamental invalidation: Optical Communications sales declining sequentially in any quarter, or a Q4 core sales guide below $5.0B — either would break the "$20B annualized run-rate already achieved" logic that underwrites this multiple.
- Why HIGH CONVICTION (two independent signals): (1) a beat on both lines plus a raised forward guide with 190bp of operating margin expansion and $1.42B of adjusted FCF — company-specific, verifiable this morning; (2) a 47% price de-rating to 43x/34x while three hyperscalers signed multiyear supply agreements — a valuation-versus-evidence divergence that does not require the AI narrative to re-inflate, only for it to stop deteriorating.
- Timeframe: 1–3 months (T1), 6–12 months (T2).
2. LONG GLW vs SHORT the high-multiple optics complex (COHR and/or LITE) SPECULATIVE
- Structure: long GLW at $136–148 against an equal-dollar short in COHR ($271.31) and/or LITE ($711.96).
- Thesis: GLW at ~43x 2026E core EPS with a 4.5% FCF yield, contracted multiyear hyperscaler agreements and 21% segment margins is the highest-quality vehicle in the optical complex. LITE (~52x forward, ~157x trailing as of my June note) and COHR (~189x trailing) carry the same AI-capex beta with more multiple to lose and, in COHR's case, direct exposure to the co-packaged-optics timing debate that has whipsawed the group twice this year.
- Target: 15 points of relative outperformance. Invalidation: GLW underperforms COHR/LITE by 8% from initiation, or either name posts optical revenue growth above GLW's next quarter.
- Timeframe: 1–3 months. Note LITE reports around Aug 11 — event risk in the short leg.
3. WATCH: the Fed-shock accumulation zone WATCH
If the FOMC hikes Wednesday, do not average into the knife the same afternoon. Zone $118–128 (36–39x 2026E) is where long-term money should be working, and the trigger to deploy is three consecutive daily closes above the post-shock five-day high — not a price alone. Invalidation for the whole franchise thesis is unchanged: sequential decline in Optical Communications.
9. What I want to hear on the 8:30 a.m. call — and the honest risks
- Capex, capex, capex. Management promised "substantially more free cash flow" alongside 10x connectivity capacity and +50% U.S. fiber capacity. Those two statements are in tension in 2027. If the call reveals a 2027 capex step-up that eats the $5.7B annualized FCF run rate, my 4.5% FCF-yield support argument weakens materially. This is the single most important thing on the call.
- Solar. $438M of sales, +90%, and still a $7M loss. Management expects improvement from Q3 and a >$3B revenue stream. Corning's own risk factors call out dependence on "government grants, tax credits or other government incentives" — a policy-sensitive earnings stream inside an AI-multiple stock. Treat any Solar profitability slippage as a guide-quality issue, not a rounding error.
- Life Sciences & Emerging: −15% y/y and a $21M loss. Small, but it is the one segment moving the wrong way in both directions.
- Concentration. Meta, Amazon and NVIDIA are now the growth engine. Moody's has flagged hyperscaler AI capex at ~$785B in 2026 with ~$460B of direct debt and $1.2T of off-balance-sheet lease commitments. Corning's revenue is only as durable as its customers' financing conditions — watch credit spreads, not just order books.
- No insider has bought. Zero open-market purchases in six months while ~15 sells cleared at $186–208. Directors' June 30 stock awards at $255.43 are grants, not conviction. The first open-market Form 4 buy under $150 would be a genuine third signal; I am not waiting for it, but I am noting its absence.
10. Bottom line
Corning is the rare AI-infrastructure name where the cash is already in the door: $1.42B of adjusted free cash flow in a single quarter, 20.9% core operating margin, 14.9% ROIC, and 37.7% incremental margins on the AI fiber that everyone claims to want exposure to. The market has spent four weeks pricing GLW as if the AI buildout were being cancelled, while Corning's own numbers accelerated and its guide moved up. When the market punishes a proven cash generator for the sins of unproven ones — as it did to Intel on July 23 — that is the trade. Buy $136–148, add $126–133 on a Fed shock, target $172 then $196, and get out on two closes below $133.
Sources (primary)
Nothing here is investment advice. Positions and views are the author's own and can change with the data. Prices referenced are the last verified levels as of Monday, July 27, 2026's close and pre-market Tuesday, July 28, 2026; the Q2 release was published pre-market Tuesday, July 28, 2026, and the earnings call at 8:30 a.m. ET had not yet occurred at publication.