Conviction: High Status: Researching

---

Editorial Note

Lumentum makes the laser that makes AI possible — and there's no Plan B. LITE controls 50%+ of the global EML laser market, the component inside every 800G and 1.6T optical transceiver connecting AI data centers. NVIDIA committed $2B in non-cancelable orders through end of decade. The Tokyo factory expanded capacity 12x and it's STILL not enough — sold out through 2028. At $897, this is the most fundamentally justified monster in our universe: +1,650% 1Y, golden cross, RSI 65, and the CEO literally says "we are the bottleneck."

---

The Story Right Now

LITE is in the middle of the most violent re-rating in the optical networking space. Up 1,650% in one year, 164% in three months, $64B market cap — and the growth case is arguably STILL not priced in. The entire optical-supply-chain watchlist is in strong-up with golden crosses across the board. COHR +457% 1Y, CIEN +744% 1Y, AXTI +5,433% 1Y. This isn't a single-stock story — it's an industry inflection where AI compute demand is hitting a physical wall and optical interconnects are the only solution.

What makes LITE unique among its peers: monopoly positioning. Bank of America estimates 50%+ global market share in high-speed EML (electroabsorption modulated lasers). These are the lasers inside every 800G and 1.6T transceiver. There is no alternative supply chain that comes close. When NVIDIA needed to ensure its Rubin-era GPUs wouldn't be bottlenecked by optics, it put $2B in non-cancelable purchase agreements specifically with Lumentum — alongside $2B each to Coherent and Marvell ($6B total across the optical stack).

The supply constraint is real: LITE expanded its Tokyo factory capacity by 12x and CEO Alan Lowe says it's still not enough. Sold-out window extended from end-of-2027 to ALL of 2028. New capex guidance $100M-$250M, scouting additional factory sites in Japan. They also acquired an InP (indium phosphide) optical device factory — vertical integration into the scarce substrate material that makes these lasers work. Plus a new 240,000 sq ft facility in Greensboro, NC for US manufacturing.

The risk? At 27x PE, any demand deceleration gets punished hard. And the entire thesis depends on AI capex continuing to ramp. If hyperscalers cut spending, the sold-out window shrinks fast. But right now, every data point says demand is accelerating, not decelerating.

---

Quick Snapshot

---

Action Matrix

---

Price Data

Legend

• RSI 65 = Healthy momentum, not yet overbought (>70)
• Golden cross confirmed — SMA50 ($659) well above SMA200 ($321)
• -6.5% from 52-week high — near highs, strong-up trend intact
• 179.7% above SMA200 — extreme extension, pullbacks will be sharp

---

Company Overview

One-Liner

The monopoly laser maker powering AI optical interconnects — 50%+ global EML market share, sold out through 2028.

Business Model

Key Segments

Geographic Mix

---

Competitive Analysis

Industry Position

Competitive Moat

Moat Assessment

Moat Width: Wide Moat Trend: Widening — InP acquisition + new factories deepen manufacturing advantage

Summary:

LITE's moat is the rarest kind in tech: physics-based manufacturing monopoly. EML lasers require specialized epitaxial growth, packaging, and testing that takes years to develop. The InP factory acquisition means LITE now controls its own substrate supply. No competitor can replicate this integrated capability in under 3-5 years, and by then the technology cycle may have moved to CPO (where LITE is also investing).

---

Management Assessment

Leadership

Capital Allocation

Red Flags

• Excessive exec compensation — not flagged
• High turnover in key roles — not flagged
• Related party transactions — not flagged
• Aggressive accounting — not flagged

---

Valuation

Current Multiples

Fair Value Assessment

At 27x PE with sold-out capacity through 2028 and 40%+ revenue growth, LITE is not cheap but may be justified. The key question: can revenue growth sustain 30%+ for 3+ years? If AI capex continues ramping (and every hyperscaler says it will), the answer is likely yes.

Valuation risk: 179% above SMA200 means any negative surprise triggers a violent correction. The stock can be "right" on fundamentals and still drop 30-40% on a bad earnings print or guidance miss.

---

Bull Case

Why This Could Work

1. Monopoly bottleneck in AI's physical layer

   • Evidence: 50%+ EML share, no alternative supply chain, CEO confirms bottleneck status
   • Implication: Pricing power + demand visibility unlike any other AI play

2. NVDA $2B non-cancelable validates multi-year demand

   • Evidence: Through end of decade, covering Rubin-era GPU interconnects
   • Implication: Revenue floor is locked in, not speculative

3. Sold out through 2028 despite 12x capacity expansion

   • Evidence: Extended from 2027 to 2028, still scouting new factory sites
   • Implication: Demand growth is OUTPACING massive supply expansion

4. Four-wave technology roadmap gives longevity

   • Evidence: 800G (dominating) → LPO (ramping) → 1.6T (building) → CPO (investing)
   • Implication: Not a one-cycle play — each wave extends the growth runway

5. Vertical integration via InP acquisition

   • Evidence: Acquired indium phosphide optical device factory
   • Implication: Controls scarce substrate supply, deepens moat vs would-be competitors

Upside Scenario

---

Bear Case

What Could Go Wrong

1. AI capex cycle peaks

   • How it plays out: Hyperscalers cut spending, "sold out" becomes "overcapacity" in 12-18 months
   • Probability: Low (next 12 months) / Medium (2028+)

2. Valuation compression on growth deceleration

   • How it plays out: Revenue growth slows from 40%+ to 20%, multiple compresses from 27x to 15x = 45% drawdown
   • Probability: Medium (inevitable eventually, question is when)

3. Competition catches up

   • How it plays out: Broadcom or a Chinese player develops competitive EML, LITE loses pricing power
   • Probability: Low (3-5 year timeline for fab qualification)

4. Technology skip — direct to CPO

   • How it plays out: If CPO arrives faster than expected, 800G/1.6T discrete transceivers get bypassed
   • Probability: Low (CPO is 2028+ at earliest, LITE is investing in it too)

Thesis Killers

• NVDA or major hyperscaler cancels/reduces optical interconnect orders
• A competitor announces volume EML production (watch Broadcom, Chinese fabs)
• AI capex guidance cuts from 2+ hyperscalers in same quarter

Downside Scenario

---

Market-Moving News

Recent News

News Patterns

LITE moves on capacity announcements and customer commitments. The stock re-rates on every "sold out through [extended date]" headline. Pullbacks tend to come from broad tech selloffs, not LITE-specific bad news. The OFC conference (annual optical conference) is a reliable catalyst.

---

Related Markets & ETFs

Sector Context

The ENTIRE optical supply chain is in synchronized strong-up. This is not a LITE-specific move — it's an industry re-rating driven by AI infrastructure demand. 15 of 17 optical-supply-chain tickers have golden crosses. The only laggard is ALAB (death cross, -4.3% vs SMA200).

---

Cross-References (Where This Appears in Our System)

---

Related Tickers to Monitor

---

Catalysts & Timing

Upcoming Catalysts

---

Entry Strategy

Position Sizing

My conviction: High Target allocation: 5-7%

Entry Approach

Technical Levels

---

The Gold

Key Discoveries

Open Questions

• What is the actual PE on forward earnings? 27x trailing vs forward matters
• How much of the $2B NVDA deal is in the current run-rate vs backlog?
• What's LITE's CPO roadmap? If they lead in Wave 4 too, the story extends to 2030+
• Is the Japanese yen a tailwind or risk for Tokyo factory economics?
• What happens to EML demand if CPO adoption accelerates past current timeline?

Research Log

---

Sources

• Fundamentals figures: company-reported results (quarterly/annual filings) as available at the artifact date; predates the desk's EDGAR reconciliation gate — figures not re-verified after publication.
• Reference statements for re-verification: income-statements.