Corning (GLW): A Genuine Beat That Fell 15% — Is the AI Optical Sell-Off a Gift?

Corning Incorporated (GLW) Q2 2026 — Corning (GLW), the 175-year-old materials-science company that now supplies the optical fiber and connectivity wiring AI data centers, reported Q2 2026 (quarter ended June 30, 2026) before the open on July 28 — and beat. Core EPS of $0.78 topped the $0.75 consensus and rose 30% YoY; core sales of $4.74B beat $4.63B and grew 17%, the ninth straight quarter of YoY growth. Quality improved too: core gross margin +120 bps to 39.6%, core operating margin +190 bps to 20.9%, core ROIC 14.9%. Optical Communications did the heavy lifting — sales +32% to $2.07B (44% of the company) with segment net income +77%, and Enterprise Networks (the data-center piece) up 65%. In the quarter Corning signed a long-term NVIDIA partnership to expand U.S. optical-connectivity capacity 10x and a multiyear, multibillion-dollar Amazon deal. Cash is real: $1.72B operating cash flow and $1.42B adjusted free cash flow, with a $0.28 quarterly dividend and ~$6.4B net debt against $1.7B of quarterly operating cash. And the stock still fell about 15% at the open, from $143.36 to roughly $122 — because Q3 guidance (core EPS $0.85-0.89, sales $4.9-5.0B) landed merely in line, optical is effectively sold out until new capacity ramps in 2027, and GLW was one of the most crowded AI-infrastructure trades on the tape after a ~190% run to a June high near $272. Our owner-earnings DCF on a ~$3.5B/yr base, blending a steady +11%/yr path and the +19%/yr Springboard path at a 9% discount rate, lands fair value near $155 versus ~$122 today — roughly 27% upside, with the market effectively paying only the steady case. Our call: BUY, 3/5 — the pullback is the gift, but it is still a premium multiple on a business whose near-term growth is capacity-gated. Wall Street rates GLW a Buy with an average target near $215 (~76% upside), so we AGREE on direction but are far more measured. Corning beat, and the stock fell 15%. That is the whole episode in one sentence. GLW is a 175-year-old materials-science company — specialty glass, ceramics, optical physics — that has quietly become one of the most strategic suppliers to the AI build-out, making the optical fiber, cable and connectivity that link GPUs and racks inside data centers. Q2 2026 (ended June 30): core EPS $0.78 vs $0.75 expected, up 30% YoY; core sales $4.74B vs $4.63B expected, up 17% — the ninth consecutive quarter of year-over-year growth. Core gross margin expanded 120 bps to 39.6% and core operating margin 190 bps to 20.9%, so this was operating leverage, not just volume. Optical Communications rose 32% to $2.07B — about 44% of sales — with segment net income up 77% and Enterprise Networks up 65%. And it is contracted, not hoped-for: NVIDIA signed a long-term partnership for Corning to expand U.S. optical-connectivity capacity tenfold, and Amazon signed a multiyear, multibillion-dollar agreement. The rest of the portfolio is ballast: Glass Innovations $1.46B (+1%), Automotive $0.47B (+2%), Solar/Hemlock $0.44B (+90%, still a small loss during a plant upgrade), Life Sciences and Emerging $0.29B (-15%). So why the 15% gap down? Three reasons. The Q3 guide — core EPS $0.85-0.89 against an $0.85 consensus, revenue $4.9-5.0B — was merely in line, and for a stock up ~190% on the year that reads as a disappointment. Positioning: GLW was among the most crowded AI-infrastructure trades, with JPMorgan flagging valuation and Morgan Stanley noting that sold-out optical caps near-term upside. And the tape: AI-infrastructure names had been selling off for weeks. The result is a genuinely accelerating business now trading roughly 55% below its June high near $272. On the numbers: operating cash flow $1.72B, adjusted free cash flow $1.42B even while capex ramps (partly funded by NVIDIA and Amazon customer prepayments), a $0.28 quarterly dividend (~0.9% yield), and about $2.5B cash against $8.9B debt — around $6.4B net debt, very manageable. The Springboard plan targets a $20B annualized sales run-rate exiting 2026, $30B by 2028 and $40B by 2030, a 19% sales CAGR with earnings growing faster. Our owner-earnings DCF starts from a ~$3.5B/yr normalized base and runs two paths — steady at +11%/yr and AI-optical at +19%/yr — at 8%, 9% and 10% discount rates. At 9%, the steady path is worth about $122 a share, which is almost exactly today's price; the AI-optical path is worth about $188. A fair 50/50 blend lands near $155, roughly 27% above the price — meaning the market is paying for the steady case alone and handing you the Springboard optionality nearly free. Our call: BUY, 3 out of 5. The pullback is the gift, but respect the risk — it is still a premium multiple on a business whose growth is capacity-gated until 2027 and whose legacy segments are flat to cyclical. We would accumulate here and add toward $110. Wall Street rates it a Buy with an average target near $215 (about 76% upside, high target Mizuho $270), so for once we AGREE on direction and simply sit more measured than the crowd. Prices referenced are from the July 28, 2026 session. Do your own research. Not financial advice. THE CALL: BUY (3/5, THE PULLBACK IS THE GIFT — A PREMIUM AI-OPTICAL COMPOUNDER FINALLY AT A FAIRER PRICE) — base-case value ~$155 vs ~$122 today. What to watch: evidence that the 10x U.S. optical capacity is coming online on schedule, optical re-accelerating into 2027, Solar/Hemlock swinging to profit, and management converting Springboard's run-rate targets into reported earnings — any of those re-rates the stock; the risk to respect is the mirror image — a genuine AI-capex digestion that pushes out the data-center build-out, an execution slip on the capacity ramp, or a broad de-rating of expensive growth, because at a mid-30s forward multiple momentum can overshoot on the way down exactly as it did on the way up Also on YouTube: @ChargedAlpha DISCLAIMER: For informational and educational purposes only. Not financial advice. Do your own research before any investment decision.