Corning
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
A 175-year materials-science franchise whose glass, fiber and ceramics processes are hard to copy, now with hyperscalers co-funding its fiber capacity on multi-year agreements; the moat is real but sits in physical products that still trade on price, and a growing share of sales rides AI data-center capex.
Corning's durability is process know-how plus customer-funded capacity — less a fortress than a supplier its customers cannot easily replace:
- Hyperscalers Are Paying for Its Capacity: Meta signed an agreement worth up to $6B for data-center fiber, anchoring a Hickory, North Carolina expansion; NVIDIA is investing $300M to fund three new US fiber plants; AT&T signed a $3B-plus fiber deal; and two more hyperscalers entered large long-term agreements in Q1 2026. Customers pre-committing capital to one supplier is the strongest evidence of switching cost a materials company can show.
- Process IP Across Several Glass Markets: Corning invented low-loss optical fiber and the fusion-draw process behind display glass and Gorilla Glass, and runs a large R&D organisation on glass and ceramic chemistry. That know-how lets it take the same science into new markets — AI optical connectivity, solar wafers, semiconductor glass — faster than single-product rivals.
- Still Competitive, Still Cyclical: Prysmian, YOFC, Fujikura and Sumitomo make fiber and cable; AGC and NEG compete in display glass. Display is a mature, low-growth business and solar is new. Much of the 2026 acceleration comes from AI data centers, so a capex pause would slow the largest growth engine.
Moat Verdict
Corning's moat is AI-resilient and physical: hyperscalers co-funding its fiber capacity is the strongest pillar, with intact process IP, talent, scale, bundling and the Gorilla Glass brand behind it. Nothing here is threatened by AI; the risk is cyclical, because AI data-center fiber is now the main growth engine.
73.1 resilient · 65.0 vulnerable · 80/20 = 71.5 · = 72
Open a moat to read its note.
Corning sells materials and components to manufacturers; there is no user interface to learn.
A physical manufacturer: glass and fiber know-how is rated under proprietaryData and scaleEconomics, not as software logic.
Corning does not derive advantage from controlling access to public data.
Glass and ceramic scientists with decades of process experience are rare, and Corning's Sullivan Park research center is one of the few places they concentrate. AGC, NEG and the fiber makers have their own teams, so it is a deep bench rather than a unique one.
Corning sells fiber, cable, connectivity and hardware as end-to-end data-center systems, which simplifies hyperscaler installation; each component still has alternative suppliers, so the bundle wins share without locking customers in.
Decades of glass-composition and process data, plus inventions such as low-loss optical fiber and fusion-draw glass, give Corning a head start in each new glass application. Patents expire and rivals have caught up in fiber and display glass, so this is an advantage, not an exclusive dataset.
No licence or certification regime ties customers to Corning's products.
Glass and fiber have no network effect; more customers do not make the product more valuable to each other.
Customers are funding Corning's capacity on multi-year agreements: Meta's fiber agreement is worth up to $6B, NVIDIA is investing $300M in three new US fiber plants, AT&T signed a $3B-plus deal, and two more hyperscalers signed large long-term agreements in Q1 2026. A customer that co-funds a supplier's plant does not switch on price.
Corning makes physical components; nothing downstream defers to it as a source of truth.
Corning is among the largest optical fiber and cable makers and runs the fusion-draw process at global scale, but Prysmian, YOFC and Fujikura in fiber and AGC and NEG in display glass also operate at scale, so the cost position is competitive rather than a structural lead.
Gorilla Glass is a rare ingredient brand that phone makers name on spec sheets, which helps Corning hold premium cover-glass pricing; the rest of the business sells on specification, not brand.
Combined average of Moat (AI Resilience), Growth, and Valuation scores.
Moat Score
A 175-year materials-science franchise whose glass, fiber and ceramics processes are hard to copy, now with hyperscalers co-funding its fiber capacity on multi-year agreements; the moat is real but sits in physical products that still trade on price, and a growing share of sales rides AI data-center capex.
Growth Score
Q2 2026 core sales grew 17% to $4.74B and core EPS 30% to $0.78, after Q1's $4.35B (+18%) and $0.70 (+30%). Optical Communications grew 32% to $2.07B, with Enterprise Networks (the data-center business) up 65% and Gen AI products growing faster; Solar grew about 90%. Core operating margin reached 20.9% (+190bp YoY) and ROIC 14.9%, up from 8.8% in Q4 2023. Q3 is guided to $4.9–5.0B of core sales (+16%) and $0.85–0.89 of core EPS (+28%). In May Corning upgraded and extended its Springboard plan to a $20B annualised sales run-rate by end-2026, $30B by end-2028 and $40B by end-2030, with a $27B 'high-confidence' 2028 floor and a 19% sales CAGR from Q4 2026 to Q4 2030. The stock ran from about $77 at the start of 2026 to a peak above $271 in late July, then fell more than 40% after the softer Q3 guide.
Valuation Score
At $153.76 (September 30 close, ~$135B market cap), Corning sits between the $105 bear and $165 base, about 81% of the way from bear to base, for a valuation score of 70. That is roughly 47× a ~$3.30 2026 core EPS and ~37× a ~$4.15 2027 — a growth multiple on a company that traded at 15–20× through most of the last decade.
The Glass Nervous System
Corning's durability is process know-how plus customer-funded capacity — less a fortress than a supplier its customers cannot easily replace:
- Hyperscalers Are Paying for Its Capacity: Meta signed an agreement worth up to $6B for data-center fiber, anchoring a Hickory, North Carolina expansion; NVIDIA is investing $300M to fund three new US fiber plants; AT&T signed a $3B-plus fiber deal; and two more hyperscalers entered large long-term agreements in Q1 2026. Customers pre-committing capital to one supplier is the strongest evidence of switching cost a materials company can show.
- Process IP Across Several Glass Markets: Corning invented low-loss optical fiber and the fusion-draw process behind display glass and Gorilla Glass, and runs a large R&D organisation on glass and ceramic chemistry. That know-how lets it take the same science into new markets — AI optical connectivity, solar wafers, semiconductor glass — faster than single-product rivals.
- Still Competitive, Still Cyclical: Prysmian, YOFC, Fujikura and Sumitomo make fiber and cable; AGC and NEG compete in display glass. Display is a mature, low-growth business and solar is new. Much of the 2026 acceleration comes from AI data centers, so a capex pause would slow the largest growth engine.
Moat Verdict
Corning's moat is AI-resilient and physical: hyperscalers co-funding its fiber capacity is the strongest pillar, with intact process IP, talent, scale, bundling and the Gorilla Glass brand behind it. Nothing here is threatened by AI; the risk is cyclical, because AI data-center fiber is now the main growth engine.
73.1 resilient · 65.0 vulnerable · 80/20 = 71.5 · = 72
Open a moat to read its note.
Corning sells materials and components to manufacturers; there is no user interface to learn.
A physical manufacturer: glass and fiber know-how is rated under proprietaryData and scaleEconomics, not as software logic.
Corning does not derive advantage from controlling access to public data.
Glass and ceramic scientists with decades of process experience are rare, and Corning's Sullivan Park research center is one of the few places they concentrate. AGC, NEG and the fiber makers have their own teams, so it is a deep bench rather than a unique one.
Corning sells fiber, cable, connectivity and hardware as end-to-end data-center systems, which simplifies hyperscaler installation; each component still has alternative suppliers, so the bundle wins share without locking customers in.
Decades of glass-composition and process data, plus inventions such as low-loss optical fiber and fusion-draw glass, give Corning a head start in each new glass application. Patents expire and rivals have caught up in fiber and display glass, so this is an advantage, not an exclusive dataset.
No licence or certification regime ties customers to Corning's products.
Glass and fiber have no network effect; more customers do not make the product more valuable to each other.
Customers are funding Corning's capacity on multi-year agreements: Meta's fiber agreement is worth up to $6B, NVIDIA is investing $300M in three new US fiber plants, AT&T signed a $3B-plus deal, and two more hyperscalers signed large long-term agreements in Q1 2026. A customer that co-funds a supplier's plant does not switch on price.
Corning makes physical components; nothing downstream defers to it as a source of truth.
Corning is among the largest optical fiber and cable makers and runs the fusion-draw process at global scale, but Prysmian, YOFC and Fujikura in fiber and AGC and NEG in display glass also operate at scale, so the cost position is competitive rather than a structural lead.
Gorilla Glass is a rare ingredient brand that phone makers name on spec sheets, which helps Corning hold premium cover-glass pricing; the rest of the business sells on specification, not brand.
Growth Analysis
Growth Drivers
Key Risk
Enterprise optical fiber for AI data centers is now Corning's largest growth engine, so a 2027–28 hyperscaler capex pause would slow the Springboard path just as new capacity comes online. Falsifiable: a 2028 annualised sales run-rate below the $27B 'high-confidence' floor, or core operating margin falling back below 19%, would break the thesis. Secondary: solar demand depends on US policy, and display pricing is set by an AGC/NEG oligopoly.
Score Derivation
80.3 base + 2.7 trajectory + 4 margin − 5 risk = 82
Base 80.3 (13–18% CAGR, midpoint 15.5%) + 2.7 trajectory (Optical Communications and Solar accelerating; Display and Specialty Materials stable) + 4 margin (expanding: core operating margin 20.9%, +190bp YoY, and ROIC 14.9% vs 8.8% in Q4 2023) − 5 moderate risk (AI fiber demand is the main growth engine and a capex pause would slow it, but long-term agreements and a diversified base limit the downside versus pure AI hardware names) = 82
Price Scenarios (12–24 Months)
Valuation Analysis
2026 core EPS is about $3.30 if Q4 lands near $0.95 after Q1 $0.70, Q2 $0.78 and a Q3 guide of $0.85–0.89. The base of $165 prices ~33× a ~$5.00 2028 EPS, which needs Springboard to keep growing earnings faster than sales. The bear of $105 is ~25× a ~$4.20 2028 if AI fiber digests. Bernstein initiated at $140 on September 30; the Street mean is ~$176.50. Add on weakness toward the $120s rather than at the current price. ~$165.
Valuation Multiples
| P/E (2026E core) | ~47× |
| Forward P/E (2027E) | ~37× |
| PEG Ratio | ~2.4× |
| Price / Sales (run-rate) | ~7× |
After a 40%-plus fall from the July peak Corning is no longer priced for perfection, but at ~37× next year's earnings it still prices most of the Springboard plan. The valuation only works if earnings keep compounding faster than sales into 2028.
Approximate figures as of September 30, 2026.
Where We Are vs Targets
Loading live price…
AI fiber digests in 2027 as hyperscalers absorb capacity, Springboard slips toward the high-confidence floor, and the multiple reverts toward Corning's history.
- Enterprise Networks growth falls below 15% YoY as hyperscaler capex pauses
- Core operating margin stalls near 20% as new fiber plants run under-utilised
- 2028 EPS near $4.20 at ~25×, giving about $105
Corning hits the $20B run-rate exiting 2026 and tracks between the $27B floor and the $30B plan for 2028, with earnings growing faster than sales.
- Q4 2026 exits near a $20B annualised sales run-rate
- Meta, NVIDIA-funded and other hyperscale fiber capacity ramps on schedule
- 2028 EPS near $5.00 at ~33×, giving about $165
Co-packaged optics pull fiber deeper into the rack, Springboard beats the $30B 2028 plan, and Corning is valued as AI infrastructure rather than as a glass maker.
- Corning fiber and connectivity designed into co-packaged optics systems with Lumentum and others
- A further hyperscaler signs a multi-billion-dollar fiber agreement
- 2028 EPS near $5.50 at ~44×, giving about $240