LITE

LITE — Lumentum

TL;DR

Lumentum holds 50–60% of the global EML market and is still undershipping demand by ~30% even after adding 20% capacity. All EML capacity is locked under long-term agreements through calendar 2027; every incremental unit commands premium pricing. The OCS backlog went from $0to$400M+ in under a year; first scale-up CPO shipments are targeted for late 2027. Nvidia’s $2B equity investment (same day as Coherent, March 2026) anchors the CW laser/CPO demand side. The one structural risk: Coherent’s 6-inch InP platform is targeting half the cost-per-laser — if yields hold at scale, Lumentum faces a multi-year cost headwind.¹²

Business

Designs and manufactures InP-based photonic components: EML lasers (80–100G/lane, and 400G/lane DD-EML for next-gen), VCSELs, CW (continuous-wave) lasers, and optical components for transceivers and CPO. Also a leading supplier of MEMS-based Optical Circuit Switches (OCS) and coherent DCI transceivers. Operates across datacenter, telecom, and industrial verticals; datacenter photonics is the high-growth segment. Included in the S&P 500 as of March 2026. Revenue run-rate ~$665M/quarter(+65$805M guided for the next quarter.¹²

Thesis

• EML structural scarcity. Lumentum commands 50–60% global EML market share and cannot keep up with demand even after a 20% capacity addition. EML shipments grew >8× from FY2020 to FY2026E. All existing capacity is locked under multi-year LTAs through calendar 2027; incremental units sell at premium spot pricing. +50% additional EML capacity being added in CY2026 vs. CY2025. InP optical lane demand from AI datacenters projected at ~85% CAGR from CY2026 to CY2030.¹²
• 400G/lane DD-EML sets the 3.2T standard. Management argues silicon photonics lacks the bandwidth at 400G/lane and that EML is the de facto standard for 3.2T transceiver generation. Lumentum is the only high-power (350+ mW) EML vendor publicly disclosing linewidth specifications — a quality differentiation signal.²³
• **OCS — from zero to $400M+backloginoneyear.\ast \ast MEMS-basedOpticalCircuitSwitcheswitha300\times 300portcount;leadscompetitorsoninsertionloss.Backlogreached>$400M (~$0twelvemonthsprior).Threecustomers,fourusecases.Target:>$1B annual run rate in 2027. OCS unit CAGR CY2025–CY2028 guided at >150%. OCS SAM projected >$2.5B by 2029. Port-to-xPU ratio in scale-up architectures growing from 1.5:1 today toward 2:1–10:1.²
• Nvidia anchor investment — CPO/CW pull. Nvidia invested $2B in Lumentum equity (March 2, 2026 — same day as Coherent), accompanied by a multibillion-dollar supply agreement covering CW lasers for silicon-photonics CPO modules. First scale-up CPO shipments targeted for late calendar 2027, tied to the Rubin Ultra/Feynman roadmap. CW laser vertical integration (producing in-house rather than sourcing) expected CQ3 2026.²
• Operating margin inflection. Operating margin expanded from 7.5% to 25.2% in twelve months; 30%+ guided. FCF conversion improving as mix shifts toward higher-ASP EMLs and OCS.¹

Risks

• Coherent 6-inch InP cost pressure. Coherent’s industry-first 6-inch InP platform achieves 4× devices per wafer at approximately half the cost of 3-inch. Lumentum is believed to be on 3→4-inch InP. If Coherent’s 6-inch yields sustain at scale, Lumentum faces a compound cost headwind over 18–24 months — EML is a commodity at the margin if cost-per-lane collapses.¹²
• High valuation multiple. Stock at 60×+ forward earnings and +76% YTD as of mid-March 2026. The photonics cycle has pulled forward considerable EPS expectation; any supply normalization (Samsung/Broadcom InP ramp, Coherent 6-inch, SiPho gaining at 3.2T) would re-rate the multiple.²
• OCS concentration. The $400M+ OCS backlog comes from three customers. Loss of a single hyperscaler engagement or a technology shift (e.g. liquid-crystal OCS from Coherent displacing MEMS) would materially reset OCS trajectory.²
• CW laser sourcing risk before vertical integration completes. Until CQ3 2026, Lumentum sources CW lasers externally for CPO. Any supply chain disruption in the interim creates execution risk on the Nvidia agreement.
• Coherent VCSEL competition. Coherent demonstrated 100G/lane VCSEL at OFC 2026 vs. Lumentum’s 32G/lane NPO demo — if 100G/lane VCSELs achieve production yield, the within-rack short-reach market could shift away from Lumentum VCSEL.³

Recent catalysts

• 2026-03-23 — FundaAI OFC takeaway: CEO quote confirming OCS SAM doubled to $4B; CY2027 booking pace “filling very quickly”; earnings estimates revised higher.⁴
• 2026-03-18 — FundaAI Deep LITE: comprehensive initiation — EML market share, OCS backlog mechanics, InP lane demand CAGR, 1.6T transceiver volume ramp starting summer 2026.²
• 2026-03-02 — Chipstrat: Lumentum and the Laser Bottleneck — 30% demand undership despite capacity adds; LTAs locking all existing capacity; 60× forward earnings; OCS from $0\to$400M+ backlog; Nvidia $2B investment.¹
• 2026-03-02 — Nvidia $2B equity investment + multibillion CW laser supply agreement (received same day as Coherent).²
• 2026-03-27 — OFC 2026: Lumentum demonstrated 32G/lane VCSEL NPO; only vendor disclosing high-power EML linewidth specs.³

Second-order reads

• 2026-04-01 — Chipstrat, Coherent’s Vertical Integration Strategy — Coherent 6-inch InP cost advantage could compress LITE EML margins as new lower-cost supply enters the market → structural risk to LITE’s premium pricing on incremental units.
• 2026-03-27 — Irrational Analysis, OFC 2026 Recap — Nvidia CPO roadmap (Rubin Ultra inter-rack) is the CW laser demand event; both LITE and COHR have Nvidia supply agreements → positive pull-through for FN (transceiver manufacturing) and MTSI (CDR/TIA components in transceivers).
• 2026-03-18 — FundaAI, Deep LITE — EML structural scarcity into 2027 is a tailwind for any component that rides on EML volume → positive for FN (transceiver assembly) and the broader AAOI COHR supply chain.

Valuation & positioning

Stock ran +76% YTD to a 44.6B market cap by mid-March 2026 (S&P 500 inclusion March 23). Chipstrat pegs LITE at 60×+ forward earnings — a premium to COHR (~44×) despite comparable or deeper photonics franchise. FundaAI: InP optical lane demand from datacenter is ~85% CAGR CY2026–CY2030; OCS run rate target >1B in 2027. If both EML and OCS scale as modeled, LITE’s EPS trajectory is still steep enough to grow into the multiple — but the margin of safety is thin at current prices.¹²

Sources

Related

COHR — closest competitor; both received $2B Nvidia investment on same day; COHR has 6-inch InP cost advantage AVGO — competes on EMLs, CW lasers; AVGO bundles with DSP+switch, LITE DSP-agnostic AAOI — EML competitor; smaller, no OCS FN — transceiver contract manufacturer; downstream of LITE EML + 1.6T volume MTSI — CDR/TIA components in transceivers CIEN — DCI coherent overlap NVDA — primary CPO demand driver; Nvidia supply agreement anchor

Footnotes

1. Chipstrat — Lumentum and the Laser Bottleneck — 2026-03-02 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷

2. FundaAI — Deep LITE: Long-Term Upside Potential — 2026-03-18 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹²

3. Irrational Analysis — OFC 2026 Irrational Recap — 2026-03-27 ↩ ↩² ↩³

4. COHR OFC Takeaways, MU Crushes — 2026-03-23 ↩