Public companyResearch currentFresh

Lumentum Holdings Inc

By Scott McLeodPublished by PXS Research Updated October 26, 2026

Connectivity & Positioning · Physical AI

SignalHOLDReviewed 26 Oct 2026

Current thesis

Connectivity & Positioning exposure in Physical AI.

Lumentum occupies one of the tightest supply bottlenecks in the AI infrastructure buildout: the indium phosphide (InP) lasers that drive 800G and 1.6T optical interconnects between GPU clusters. These lasers are the physical layer that converts electrical signals to light and back — and Lumentum's lasers are **sold out through 2027, potentially 2029**. The CEO has stated the company is "significantly under-shipping demand." When a critical component is on allocation and capacity cannot be added quickly (InP wafer fabs take 2-3 years to build and…

Catalysts

Q1 FY27 revenue/OM guide already issued; 1.6T ramp and ultra-high-power CPO lasers; first ELS module order and NPO engagement conversion; customer LTA renewals.

Invalidation

CPO deployed broadly before 2028; LITE loses key hyperscaler LTA; Chinese laser module competition floods market

Latest change

30 Sept 2026 · Conviction moved from 1 to 2.

Research history

Earlier changes

See all LITE changes

30 Sept 2026

Target
$1,108.90 → $1,167.01

Capital in context

Mapped funds and investors

Direct relationships supported by published evidence.

Reported holdings as-of

No sourced relationships mapped yet.

Disclosed investment history

No sourced relationships mapped yet.

Showing up to four relationships in each group. ETF weights describe the fund’s portfolio, not ownership of this company. Private investment history does not establish current ownership.

Institutional positioning

ins-$14.1M · 13F 20+/5-×0.0 · short↑0.35

Current snapshot

Social sentiment reference

X: bullish — InP laser supply experts confirm LITE/AXTI/AAOI sold out 27-30% undersupply; structural multi-year constraint

Lower-confidence snapshot

Valuation reference

Target price: $1208.35

Dated research target

Research rigor

Moat & position

Evidence Grade: Insufficient
Moat Durability: Capacity cycle
Toll Road Tier: Strong proxy

Timestamped research update

Deep research update

8,954 words · Updated 21 Oct 2026 · 18 sources

Business

Lumentum supplies lasers and photonic components for two massive markets: (1) AI data center optical interconnects (InP-based EML, CW, and coherent lasers for 800G/1.6T transceivers), and (2) 3D sensing and LiDAR (VCSEL arrays for face ID, depth sensing, and automotive/industrial LiDAR). The AI datacom business is the explosive growth engine — revenue has tripled in ~5 quarters under CEO Hurlston. Lasers are sold out through 2027, potentially 2029. The 3D sensing/VCSEL business is a steady contributor but not the growth story.

Lumentum is the leading independent InP laser supplier for AI data-center optical interconnect. It sells (1) Components — EML (electro-absorption modulated) and CW lasers on InP substrates that go into 800G/1.6T transceivers, plus ultra-high-power CPO (co-packaged optics) lasers and ELS (external laser source) modules; and (2) Systems — ROADMs, optical circuit switches (OCS), and transport modules. FY2026 revenue $3.014B (+83.2% YoY), split ~66.5% Components / ~33.5% Systems ($2,005.6M / $1,008.4M). The legacy VCSEL/3D-sensing and LiDAR laser business (the original "Perception & Sensing" label) is now a minority sidecar.

Q4 FY26: rev $1,006.3M (+24.5% QoQ / +109.3% YoY); GAAP GM 47.4%, non-GAAP GM 50.4%; GAAP OM 27.8%, non-GAAP OM 36.6%; non-GAAP EPS $3.23.
FY26: rev $3,014.0M (+83.2% YoY); non-GAAP GM 46.0%; non-GAAP OM 29.8%; non-GAAP NI $782.3M ($8.67/sh).
Q1 FY27 guide: rev $1.225-1.275B; non-GAAP OM 39.5-40.5%; non-GAAP EPS $4.05-4.35.
Balance sheet: $2.7B cash + ST investments.

Lumentum makes optical and photonic components and systems for cloud/AI interconnect, telecom, industrial lasers, and 3D sensing. It reports as one segment and splits revenue into Components (laser chips/assemblies that go into someone else's module) and Systems (stand-alone modules, optical circuit switches, industrial lasers). The 10-K is explicit that 3D-sensing VCSELs are a consumer-electronics sidecar, not the growth engine.

FY26 net revenue $3,014.0M vs $1,645.0M (+83.2%). Components $2,005.6M (66.5%, +79.7%); Systems $1,008.4M (33.5%, +90.7%).
GAAP GM 41.7% (vs 28.0%); GAAP operating income $524.8M (17.4% OM) vs −$180.1M.
GAAP net loss $6,935.1M (−$92.96/sh) on $7,756.6M debt-extinguishment loss. Non-GAAP NI $782.3M / $8.67; non-GAAP GM 46.0%; non-GAAP OM 29.8%.
Q4: $1,006.3M (+109.3% YoY / +24.5% QoQ); GAAP GM 47.4% / OM 27.8%; non-GAAP GM 50.4% / OM 36.6%; non-GAAP EPS $3.23.
Q1 FY27 guide: revenue $1.225–$1.275B; non-GAAP OM 39.5–40.5%; non-GAAP EPS $4.05–$4.35.
Cash + ST investments $2,738.4M ($2,043.5M cash + $694.9M ST). FY26 CFO $751.4M; capex $451.3M; Greensboro fab acquired 2026-03-17 for $38.0M cash.
International ship-to revenue 79.2%. Employees 13,757 (11,916 manufacturing). Thailand PPE $450.6M vs U.S. $153.0M / Japan $232.1M — manufacturing is Asia-weighted.

monetization model: One-time hardware sales of optical/photonic components, modules, subsystems, and lasers; limited NRE/services. (FY2024 10-K, Item 1 Business)
revenue mix: Q3 FY2025: Cloud & Networking ~89% ($379.1M/$425.2M); Industrial Tech ~11% ($46.1M). FY2024 10-K legacy segments: Optical Communications ~87%, Commercial Lasers ~13%. (Q3 FY2025 10-Q segment note; FY2024 10-K Note 15)
who pays: Hyperscalers/cloud providers, network-equipment OEMs, telecom carriers, industrial OEMs. (FY2024 10-K Item 1; Risk Factors — customer concentration)
margin_value_capture: Q3 FY2025 GAAP gross margin 26.2%; non-GAAP gross margin 32.4%; value accrues in III-V photonics and qualified high-speed optical components. (Q3 FY2025 10-Q MD&A)

customer classes: Hyperscalers, Data Center Operators, Industrial Automation - OEMs, Semiconductor Fabs - IDM Buyers

Sources: FY2024 10-K (accession 0001633978-24-000054), sec.gov ; Q3 FY2025 10-Q (accession 0001633978-25-000024), sec.gov .

Lumentum Holdings Inc. is a designer and manufacturer of optical and photonic products. It operates through two reportable segments: Cloud & Networking and Industrial Tech.

According to the Lumentum Holdings Inc. Form 10-K for fiscal year 2024, filed 2024-08-21 with the SEC, the company reported:

Net revenue: $1,359.2 million
Cloud & Networking net revenue: $1,063.9 million
Industrial Tech net revenue: $295.3 million
GAAP gross margin: 19.0%
GAAP operating loss: $(394.7) million
GAAP net loss: $(551.9) million

Source: SEC EDGAR, Lumentum Holdings Inc., Form 10-K, filed 2024-08-21, fiscal year ended 2024-06-29, accessed 2026-04-26. URL and document identifier in Sources.

The total revenue figure represents a decline from the fiscal 2023 total of $1,767.3 million, also disclosed in the same 10-K. The company’s own filings attribute the decline primarily to soft telecom end-market demand, inventory digestion at customers, and weakness in industrial end-markets—partially offset by demand in datacom and AI-related optical connectivity.

Using the segment figures above:

Segment	FY2024 revenue	Share of total
Cloud & Networking	$1,063.9M	78.3%
Industrial Tech	$295.3M	21.7%
Total	$1,359.2M	100.0%

Source: same Lumentum FY2024 Form 10-K.

The company is therefore predominantly a networking-optical business. The “Cloud & Networking” segment includes telecom optical components/modules, datacom transceivers, tunable lasers, receivers, modulators, ROADMs, amplifiers, and pump lasers. “Industrial Tech” includes diode lasers, direct-diode systems, ultrafast lasers, and VCSEL arrays for industrial, consumer electronics, automotive, and life sciences.

Lumentum monetizes by selling hardware components and modules to OEMs, not by selling directly to end users. The customers are network equipment manufacturers, datacom equipment manufacturers, and industrial/consumer electronics manufacturers. Revenue is recognized on product shipment when control transfers. The fiscal 2024 Form 10-K describes that the company’s products are embedded in customers’ systems and often require design wins, qualification cycles, and multi-year supply arrangements.

Source: Lumentum FY2024 Form 10-K, filed 2024-08-21.

From the same filing:

FY2024 GAAP gross margin: 19.0%
FY2024 GAAP operating loss: $(394.7) million
FY2024 GAAP net loss: $(551.9) million

The operating loss includes goodwill impairment and other charges. The large GAAP-to-non-GAAP adjustment is a recurring feature of Lumentum’s filings; management’s non-GAAP presentation removes stock compensation, amortization of acquired intangibles, restructuring, and impairment charges. That gap is not a sign of fabricated accounting, but it does mean GAAP profitability is weak while the company argues its adjusted economics are better.

Source: Lumentum FY2024 Form 10-K, filed 2024-08-21.

The company’s 10-K describes customers as OEMs in telecommunications, data communications, industrial, and consumer markets. The filings do not name a single largest customer by legal name in the customer-concentration discussion; they disclose concentration by percentage only. That is important: a widely cited Apple relationship is not named in the FY2024 10-K. The 10-K does state that customer concentration is material and that a small number of customers account for a large share of revenue. For the exact percentage, see the Form 10-K customer concentration footnote; I will not quote a specific percentage without a directly verified primary-source number.

Source: Lumentum FY2024 Form 10-K, filed 2024-08-21.

Within the filed document, top customers are not identified by name in the segment/customer-concentration section. The filing names customer categories rather than named end customers. If a customer name appears in a 10-K risk factor or contract disclosure, it is incidental. Therefore, under the rule “named from filings only,” I do not infer a named top customer such as Apple, Cisco, Ciena, or Infinera. needs primary source validation if a specific named top-customer claim is required.

Lumentum sells to:

Telecommunications network equipment OEMs
Datacom equipment makers and, indirectly, hyperscale data center operators
Cable/MSO equipment suppliers
Industrial manufacturing equipment OEMs
Consumer electronics OEMs for 3D-sensing/lidar-related VCSEL arrays
Automotive lidar/industrial end-users

The FY2024 Form 10-K does not disclose the legal names of its largest end customers. It describes them by category. Any named-customer assertion based on prior reporting or industry lore is needs primary source validation unless confirmed in the filed document. Do not treat an Apple relationship as a primary-source fact from the FY2024 filing.

The filing does state that customer concentration exists: the top customers account for a meaningful share of revenue. The company’s risk factors describe reliance on a limited number of OEM customers and their qualification cycles. If you need the exact percentage concentration, you must pull the customer concentration note directly from the 10-K; I am not inserting an unverified number.

Source: Lumentum FY2024 Form 10-K, filed 2024-08-21.

Lumentum Holdings Inc. is a designer and manufacturer of optical and photonic products. Its SEC filings describe the company as operating in two broad areas:

Communication and commercial optical products, including lasers, modulators, receivers, wavelength-selective switches, optical transceivers, and components used in telecom and datacom networks.
Commercial and industrial lasers, including high-power diode lasers and fiber lasers used in materials processing, semiconductor manufacturing, and other precision applications.

These descriptions are drawn from Lumentum’s public SEC filings, particularly the Business section and segment footnote. The exact current segment names and classifications should be read from the latest Form 10-K. Source: Lumentum Holdings Inc. SEC EDGAR filing index — sec.gov — accessed 2026-04-26. Exact accession number: needs primary source validation.

Lumentum has changed its segment reporting over time. Historically, the company reported:

Optical Communications — telecom transport products, datacom transceivers, 3D sensing laser chips, and related components.
Commercial Lasers — kilowatt-class fiber lasers, diode lasers, and laser subsystems for industrial and commercial markets.

After recent acquisitions and business-line reorganization, management has discussed a segment structure oriented around:

Cloud & Networking
Industrial Tech

However, the exact segment names, percentages, and dollar splits must be verified against the FY2024 Form 10-K segment note and management’s discussion and analysis (MD&A). The revenue mix by segment as a percentage of total revenue is:

Segment	Approximate revenue mix	Verification status
Cloud & Networking / Optical Communications	Majority of revenue	needs primary source validation
Industrial Tech / Commercial Lasers	Smaller, but material	needs primary source validation

Specific percentages are omitted because I could not verify them against a primary filing in this session.

Lumentum monetizes primarily through product sales, not recurring software or subscription revenue. The company manufactures and sells optical components, modules, and laser systems. Revenue is generally recognized when control of the product transfers to the customer, typically at shipment or delivery, depending on contractual terms.

The qualitative revenue-recognition model is standard for a hardware supplier. Source: Lumentum Form 10-K, Revenue Recognition note, available via SEC EDGAR index above; document accession number: needs primary source validation.

The paying customers are typically:

Telecom equipment manufacturers that incorporate Lumentum components into optical transport systems.
Hyperscale data center operators and their direct or contract-manufacturer supply chains for optical transceivers.
Consumer electronics OEMs that use Lumentum’s laser chips for 3D sensing, facial recognition, and LiDAR-related applications.
Industrial equipment OEMs and integrators that use Lumentum’s high-power lasers for cutting, welding, marking, and other materials-processing tasks.

This list is based on the qualitative business description in Lumentum’s SEC filings. It is not an assertion that any specific customer is current; customer names must be verified from the latest 10-K customer-concentration note.

I am not providing specific gross margin, operating margin, or net income figures in this draft because I could not verify them directly against a filed 10-K or 10-Q.

Required margin fields:

Gross margin, FY2024: needs primary source validation.
Operating margin, FY2024: needs primary source validation.
Net loss or income, FY2024: needs primary source validation.
Non-GAAP gross margin, if disclosed: needs primary source validation.
Segment profit or loss: needs primary source validation.

The reason for this restraint is source-tier discipline: any margin number would be a Tier 1 financial claim. Without direct EDGAR retrieval, providing a number risks fabrication.

Lumentum’s annual report typically includes a risk factor and/or revenue-concentration disclosure about customers that represent more than 10% of revenue.

In prior years, Lumentum’s filings have identified Apple Inc. as a significant customer for laser chips used in consumer-electronics 3D sensing. However, the current percentage, whether Apple is still named, and whether additional customers exceed 10% in the latest fiscal year must be re-verified. Therefore:

Customer	Current status	Verification
Apple Inc.	Historically named as a significant customer in prior filings	needs primary source validation for latest filing
Other 10% customers	May exist in telecom and datacom	needs primary source validation

No other customer name is asserted here.

From the qualitative SEC business description, Lumentum sells to the following customer classes:

Customer class	Buyer archetype	Verification
Telecom equipment makers	Component/design engineers and supply-chain managers at optical transport system vendors	needs primary source validation for current named customers
Hyperscale data center operators	Network architects and procurement teams at large cloud providers, or their contract manufacturers	needs primary source validation
Consumer electronics OEMs	Hardware, camera, and depth-sensing module teams	needs primary source validation for current named customer
Industrial equipment OEMs	Laser system integrators, machine builders, and factory automation suppliers	needs primary source validation

Because the requirement is to name customers from filings only, I will not invent any. The only safe statement is:

In prior SEC filings, Lumentum has disclosed Apple Inc. as a significant customer. The latest fiscal-year customer-concentration note, including whether Apple remains above 10% and whether any other customer is named, must be re-read from the primary 10-K. Current percentage: needs primary source validation.

Potential customers such as Cisco, Ciena, Infinera, Nokia, or hyperscale cloud operators may be disclosed in Lumentum’s revenue-concentration table or risk factors. I do not assert them as confirmed named customers without a direct pull.

Lumentum’s 10-K likely includes a risk factor stating that a small number of customers account for a significant portion of revenue. Customer concentration creates:

Revenue volatility if a large customer changes suppliers.
Pressure on pricing and margins.
Risk that a large customer vertically integrates or multi-sources critical components.
Risk that a design change reduces demand for Lumentum’s product.

This is a qualitative risk disclosure and does not require a number for the argument that customer concentration is material. Source: Lumentum Form 10-K Risk Factors — SEC EDGAR index; accession number: needs primary source validation.

Revenue and scale

Annualized revenue exceeding $2B with strong double-digit YoY growth. Q2 calendar 2026: ~$426M (beat). Full-year guidance raised to $1.75-1.8B. Q4 FY2026 guidance $960M-$1.01B (reporting Aug 11). Non-GAAP operating margins 32-35% in strong periods. Demand significantly outstrips supply — CEO says "significantly under-shipping demand."

Thesis connection

Physical AI training clusters cannot scale without optical interconnect. LITE sells the laser that sits in almost every high-speed module and the OCS that reroutes around failed GPUs. NVIDIA underwrote the capacity bet with a $2.0B preferred-stock check.

Key evidence

Parameter	Value	Source Confidence
Q4 FY26 net revenue	$1,006.3M (+109.3% YoY)	measured (8-K EX-99.1)
Q4 FY26 non-GAAP GM / OM	50.4% / 36.6%	measured
FY2026 net revenue	$3,014.0M (+83.2% YoY)	measured
Q1 FY27 revenue guide	$1.225-1.275B	measured (guidance)
Q1 FY27 non-GAAP OM guide	39.5-40.5%	measured (guidance)
Components % of Q4 revenue	64.5% ($649.4M)	measured
Cash + ST investments	$2.7B	measured
Convertible equitization loss	$7.8B non-cash	measured
200G lanes per 1.6T port	8	inferred (industry standard)
InP substrate supply (AXT)	thru 2031, $87M prepay	inferred (prior note, needs 10-K)
"Sold out through"	2027, potentially 2029	inferred (CEO statement)

Parameter	Value	Source Confidence
FY26 net revenue	$3,014.0M (+83.2% YoY)	measured (10-K)
FY26 Components / Systems	$2,005.6M / $1,008.4M (66.5% / 33.5%)	measured (10-K)
FY26 GAAP GM / OM	41.7% / 17.4%	measured (10-K)
FY26 GAAP OI	$524.8M	measured (10-K)
FY26 non-GAAP GM / OM	46.0% / 29.8%	measured (8-K EX-99.1)
Q4 FY26 revenue	$1,006.3M (+109.3% YoY)	measured (8-K EX-99.1)
Q4 non-GAAP GM / OM	50.4% / 36.6%	measured (8-K EX-99.1)
Q1 FY27 revenue guide	$1.225–$1.275B	measured (guidance)
Customer A / B FY26	26.6% / 15.0%	measured (10-K)
Customer 1 AR	30.4%	measured (10-K)
International revenue	79.2%	measured (10-K)
Cash + ST investments	$2,738.4M	measured (10-K)
FY26 CFO / capex / FCF	$751.4M / $451.3M / $300.1M	measured / derived (10-K)
NVIDIA preferred	2.9M sh @ $695.31 = $2.0B (2026-03-02)	measured (10-K)
Common shares 2026-08-14	89.7M	measured (10-K cover)
FY26 basic WAS	74.6M	measured (10-K)
AXT InP reservation	thru 2031-12-31; $43.5M + $43.5M deposits	measured (AXT 8-K 2026-07-29)
Greensboro purchase	$38.0M cash, 2026-03-17	measured (10-K)
Patents	~1,000 US + ~900 foreign	measured (10-K)
800G / 1.6T copper reach	~10 m / ~2–3 m	inferred (CEO, DB 2026-08-27)
200G EML ASP vs 100G	~2x; >25% of EML revenue	inferred (CEO, DB 2026-08-27)
OCS next-quarter run-rate	>$100M	inferred (CEO, DB 2026-08-27)
Cloud Light qtr run-rate	>$200M	inferred (CEO, DB 2026-08-27)
NVIDIA qual cycle	~12 months	inferred (CEO, DB 2026-08-27)

\text{Cloud \& Networking share} = \frac{1,063.9}{1,359.2} \times 100 = 78.3\%

\text{Industrial Tech share} = \frac{295.3}{1,359.2} \times 100 = 21.7\%

\text{Gross profit} = 1,359.2M \times 0.190 = \$258.2M

\text{Total YoY growth} = \left(\frac{425.2}{366.5} - 1\right) \times 100 = 16.0\%

\text{Cloud \& Networking YoY growth} = \left(\frac{338.8}{306.7} - 1\right) \times 100 = 10.5\%

\text{Industrial Tech YoY growth} = \left(\frac{86.4}{59.8} - 1\right) \times 100 = 44.5\%

Parameter	Value	Source (URL + ID + date)	Tier	Measured/derived/assumed	Cross-verified
FY2024 net revenue	$1,359.2M	Lumentum Form 10-K, filed 2024-08-21	Tier 1	Measured	Single-source
FY2024 Cloud & Networking revenue	$1,063.9M	Lumentum Form 10-K, filed 2024-08-21	Tier 1	Measured	Single-source
FY2024 Industrial Tech revenue	$295.3M	Lumentum Form 10-K, filed 2024-08-21	Tier 1	Measured	Single-source
FY2024 GAAP gross margin	19.0%	Lumentum Form 10-K, filed 2024-08-21	Tier 1	Measured	Single-source
FY2024 GAAP operating loss	$(394.7)M	Lumentum Form 10-K, filed 2024-08-21	Tier 1	Measured	Single-source
Q3 FY2025 total revenue	$425.2M	Lumentum Form 10-Q, filed 2025-05-06	Tier 1	Measured	Single-source
Q3 FY2025 Cloud & Networking revenue	$338.8M	Lumentum Form 10-Q, filed 2025-05-06	Tier 1	Measured	Single-source
Q3 FY2025 Industrial Tech revenue	$86.4M	Lumentum Form 10-Q, filed 2025-05-06	Tier 1	Measured	Single-source
Q3 FY2024 total revenue (comparative)	$366.5M	Lumentum Form 10-Q, filed 2025-05-06	Tier 1	Measured	Single-source
Cloud & Networking FY2024 mix	78.3%	Derived from Lumentum Form 10-K	Tier 1	Derived	Single-source
Industrial Tech FY2024 mix	21.7%	Derived from Lumentum Form 10-K	Tier 1	Derived	Single-source
Coherent Corp. FY2024 net sales	~$4,708M	Coherent Corp. Form 10-K, filed 2024-08-16	Tier 1	Measured	Single-source

The cross-verified column is “Single-source” because I do not use two independent originating sources for each issuer-derived figure. This is the correct treatment under the source-tier rule: verified primary filings may remain single-source, but corroboration should not be invented.

\text{Switching cost per unit} = \frac{5,000,000}{400,000} = \$12.50 \text{ per unit}

NRE cost	Annual volume	Qualification life	Switching cost/unit
$2,000,000	500,000	2 years	$2.00
$5,000,000	200,000	2 years	$12.50
$5,000,000	100,000	2 years	$25.00
$10,000,000	100,000	2 years	$50.00
$10,000,000	50,000	2 years	$100.00

The result suggests that for a high-quality optical component with moderate NRE and moderate volume, switching cost can be meaningful but not overwhelming relative to the likely module price. If a module ASP is, for example, $1,000, a $12.50 switching cost is only about 1.25% of ASP. But if NRE rises to $10 million and volume falls to 50,000 units per year, switching cost reaches $100 per unit, or 10% of a $1,000 ASP.

The protective effect is stronger when:

Qualification requires customer-specific testing.
Manufacturing yield is hard to replicate.
Reliability failures would be catastrophic for the customer.
The product is designed into a platform with a multi-year life.

The protective effect is weaker when:

The customer uses standard multisource agreements.
The product is a commodity transceiver.
The customer can second-source from multiple vendors.
Cloud customers use internally developed or silicon-photonic alternatives.

This framework is not sourced from Lumentum’s financials. It is a derived analytical exercise. All numeric inputs are assumptions and are not evidence.

Bottleneck and moat

VCSEL manufacturing: LITE and Coherent dominate high-volume, high-reliability VCSEL production for consumer/auto
InP laser supply: multi-year sold-out demand, long-term InP wafer agreements (e.g., AXT through 2031, $87M prepayment)
Capacity-constrained — allocating lasers to customers, pricing power
Manufacturing scale and qualification cycles create barriers to entry

Every 1.6T optical link requires 8×200G lanes, and every 200G/lane link requires either an InP EML or an InP CW laser. Lumentum is one of only two independent high-volume suppliers (with Coherent) of the 200G/lane InP laser that makes the 800G→1.6T transition work. The physical constraint is InP wafer fabrication capacity + laser yield at 200G/lane chirp/bandwidth specs — which is why revenue grew +109% YoY while the company is still "significantly under-shipping demand."

InP EML/CW laser process: 200G/lane (and emerging 400G/lane) requires InP epitaxy + regrowth + high-speed modulation that only Lumentum, Coherent, and captive suppliers (Broadcom, Mitsubishi for internal use) produce at volume and reliability.
Multi-year qualification: lasers are qualified into each transceiver maker's platform (Innolight, Eoptolink, Fabrinet, etc.). Once a laser is designed into an 800G/1.6T module, the switching cost is a full re-qualification cycle (~12-24 months).
Sold-out capacity: lasers "sold out through 2027, potentially 2029"; Lumentum is allocating supply to customers.
Long-term InP substrate supply: AXT (AXTI) agreement through 2031 with an $87M prepayment (per prior vault note — needs 10-K confirmation).
New moat frontier: ultra-high-power CPO lasers, ELS modules (initial order), and NPO (near-package optics) engagements — CEO cites these as "optics starting to penetrate in-rack connectivity, significantly upping our optical TAM."

Every 800G/1.6T optical lane still needs an InP light source. At 1.6T, copper's reliable reach collapses from ~10 m (800G) to ~2–3 m, so optics moves from scale-out cables into the rack. Lumentum is one of two high-volume 200G/lane EML suppliers (with Broadcom) and the only merchant OCS vendor at scale. The binding constraint is time — fab tools, InP substrates, and ~12-month customer qualification — not capital.

200G EML process + yield: CEO (Deutsche Bank, 2026-08-27): Broadcom and Lumentum are the two large 200G EML suppliers; 200G ASP is ~2x 100G; laser quality sets transceiver yield, which is why the premium holds.
Qualification clock: NVIDIA qualification took ~1 year of reliability testing (same conference). Once a laser is designed into a platform, switching is a full re-qual, not a second-source PO.
Substrate reservation: AXT Capacity Reservation Agreement dated 2026-07-26 — six years through 2026-12-31 → 2031-12-31; two $43.5M deposits ($87M total) applied as shipment credits; Lumentum pays shortfall vs annual commitment. CEO still said Japanese + AXT deals "might not be enough."
IP: ~1,000 U.S. patents + ~900 foreign, expirations through 2046; ~670 applications pending (FY2026 10-K).
New frontier: ultra-high-power CPO lasers, ELS modules, NPO engines, and merchant OCS. NPO timing pulled from late-2028/early-2029 to late-2027/early-2028.

Copper-wall physics (CEO, DB 2026-08-27). 800G copper reach ~10 m; 1.6T copper reach ~2–3 m. Reach compression = 10 / 2.5 = 4.0x shorter. Any in-rack or in-cluster link longer than ~3 m that used to be copper must go optical. That is the TAM expansion, not "more transceivers of the same type."
Capacity ramp is still the P&L gate. Q3 FY26 $808.4M → Q4 $1,006.3M = +$197.9M (+24.5% QoQ). Q1 FY27 guide midpoint $1,250M = +$243.7M (+24.2% QoQ). Sequential add is still ~$200–244M per quarter of physical capacity, not demand. Annualizing the guided quarter = $5.00B vs FY26 $3.014B = +65.9% if the run-rate holds.
Operating leverage on the next $244M. Q4 non-GAAP OM 36.6% → $368.3M. Q1 midpoint 40.0% → $500.0M. Incremental operating income $131.7M on +$243.7M revenue = 54.0% incremental margin. Scarcity pricing, not mix luck.
Customer-A concentration (10-K). Customer A = 26.6% of FY26 revenue = $801.7M; Customer B = 15.0% = $452.1M; combined 41.6% = $1,253.8M. Customer A jumped from 15.4% in FY25 (+11.2 pp, +$337.6M of mix). AR is even tighter: Customer 1 = 30.4% of receivables (vs 13.2% prior year).
NVIDIA underwrite vs cash. 2.9M Series A preferred @ $695.31 on 2026-03-02 = $2.0B cash (10-K). Cash + ST investments $2,738.4M at year-end; NVIDIA proceeds = 73% of the cash pile. Preferred is 2.9 / 89.7 = 3.23% of the 2026-08-14 common share count — small equity, large cash. This is a customer-financed fab, not a passive stake.
AXT deposit vs revenue. $87M total deposits = 2.89% of FY26 revenue; first $43.5M = 1.59% of year-end cash. Tiny vs the $451.3M FY26 capex, but it is a revealed-preference bid for a wafer Lumentum cannot grow itself. Cost-of-moat vs moat-defended: $87M buys a reserved InP queue defending $2,005.6M of Components revenue (4.3% of Components; 2.9% of total).
Dilution from the equitization. FY26 basic WAS 74.6M shares; 89.7M common outstanding as of 2026-08-14 = +20.2% vs the FY average. GAAP net loss $6,935.1M is almost entirely the $7,756.6M non-cash loss on convertible equitization. Non-GAAP FY NI $782.3M ($8.67/sh) is the operating number. Q4 non-GAAP diluted shares implied by $326.3M / $3.23 = ~101.0M.
FCF after the build. Operating cash $751.4M − PP&E $451.3M = $300.1M FCF (9.96% of revenue). Capex/revenue = 15.0%. The company is FCF-positive while still under-shipping.
OCS materiality (conference, not 10-K). OCS expected to exceed $100M next quarter vs $1,250M Q1 midpoint = 8.0% of guided revenue — no longer a rounding error, still not the P&L. Cloud Light (Systems) now "well over $200M a quarter" (conference) vs Systems Q4 $356.9M — modules are the other half of Systems, and management admits module GM still lags InnoLight/Eoptolink.
200G mix (conference). 200G EMLs >25% of EML revenue at ~2x 100G ASP. Let u = 200G unit share among EML units and ASP_200 = 2 × ASP_100. Then revenue share = 2u / (u+1). Set equal to 0.25 ⇒ 2u = 0.25u + 0.25 ⇒ 1.75u = 0.25 ⇒ u ≈ 14.3% of EML units already producing 25% of EML revenue. Mix still early.

product_service: Optical transceivers/components, ROADMs, lasers, and photonic subsystems for cloud, telecom, and industrial lasers.
technical function: Electro-optical conversion, light amplification/switching/routing, and laser pumping using InP photonics, EMLs/DFBs, tunable lasers, integrated optical modules.
why it is hard: Requires III-V wafer fab, photonic integration, high-reliability packaging, and multi-year hyperscaler/OEM qualification; scale and customer concentration matter.
substitutes: Copper for short reach; silicon photonics; other optical component vendors; vertical integration by network OEMs/hyperscalers.
pricing power: MED — differentiated high-speed optics carry value, but competition, customer concentration, and price erosion in telecom/datacom limit sustained pricing.

Pricing power is mixed. In commoditized pluggable transceivers, price erosion is structural and rapid. In differentiated coherent components—tunable lasers, high-performance modulators, receivers—Lumentum historically had better pricing power because there are fewer qualified suppliers and the qualification cycles are long.

However, the fiscal 2024 telecom downturn showed weak pricing power when carrier inventories were oversupplied. If Lumentum tried to hold price, it likely lost volume. If it chased volume, gross margin compressed. This is a key tension: Lumentum can raise prices only when capacity is scarce and design wins are sticky; it cannot raise prices broadly across commodity modules.

Source: Lumentum Form 10-K fiscal 2024 management discussion, SEC EDGAR company index, CIK 0001633978 — exact accession number to be confirmed; access date 2026-04-26.

# ## Who they sell to

Lumentum sells mostly to other manufacturers, not to end users.

1. Network equipment manufacturers

These buyers integrate Lumentum components into coherent line cards, transport systems, amplifiers, and switch platforms. Example archetypes include system vendors in the optical transport market. The filing generally does not name them.

2. Data center operators and hyperscalers

Lumentum sells transceivers and transceiver components for data center interconnect and intra-data-center links. The fiscal 2024 Cloud & Networking segment is the principal exposure to this buyer class. Exact customer names are needs primary source validation.

3. Telecom service providers

Service providers buy indirectly through network equipment manufacturers, but in some cases may specify Lumentum components in their builds. The filing does not generally disclose direct service-provider customers.

4. Industrial laser and advanced manufacturing customers

These buyers use Lumentum’s fiber and ultrafast lasers in automotive, electronics, medical device, and general industrial cutting/welding applications.

5. Consumer electronics OEMs

The 3D-sensing laser franchise sells into consumer devices. The filing has historically disclosed one customer above 10% of revenue. The identity is not always explicitly named in the 10-K, but the 3D-sensing application is strongly associated with one large consumer-electronics customer. For fiscal 2024, exact concentration is needs primary source validation.

Source: Lumentum Form 10-K fiscal 2024, customer concentration and segment disclosures, SEC EDGAR company index, CIK 0001633978 — exact accession number to be confirmed; access date 2026-04-26.

The phrase “customer classes” is more robust than “customer names.” Lumentum filings describe customers as OEMs, network equipment providers, and industrial manufacturers. The 10-K risk factors emphasize:

Customer concentration
Long sales cycles
Qualification cycles
Inventory corrections in telecom
Exposure to U.S.–China trade restrictions

These are all buyer-related risks that shape the revenue mix.

# ## Worked calculation: incremental gross profit from Cloud & Networking recovery

This is an illustrative sensitivity, not a forecast. It uses explicit assumptions. Replace the assumed segment figures with filed numbers before relying on the result.

Assume fiscal 2024 total net revenue is approximately:

Total revenue: $1,359 million — needs primary source validation

Assume segment split:

Cloud & Networking: $1,090 million
Industrial Tech: $269 million

These are arithmetic assumptions consistent with the approximate 80/20 mix above — needs primary source validation.

Assume segment gross margins:

Cloud & Networking gross margin: 18%
Industrial Tech gross margin: 35%

These are assumed; they are not confirmed in this environment — needs primary source validation.

Cloud & Networking gross profit = $1,090M × 18% = $196.2M

Industrial Tech gross profit = $269M × 35% = $94.2M

Modeled total gross profit = $290.4M

Modeled total gross margin = $290.4M / $1,359M = 21.4%

If the actual overall GAAP gross margin was lower, then the assumption set is too optimistic or the segment mix/gross margins differ.

Assume total revenue recovers by 10%, or $135.9M. Assume all recovery is in Cloud & Networking.

Incremental revenue = $135.9M

Assumed incremental gross margin = 40%

Incremental gross profit = $135.9M × 40% = $54.4M

Assumed incremental opex = 8% of incremental revenue = $10.9M

Incremental operating income = $54.4M - $10.9M = $43.5M

Incremental gross margin assumption	Incremental gross profit	Incremental opex	Incremental operating income
30%	$40.8M	$10.9M	$29.9M
40%	$54.4M	$10.9M	$43.5M
50%	$68.0M	$10.9M	$57.1M

Lumentum’s operating income is highly sensitive to the incremental gross margin on recovery. If the recovery is led by commoditized transceivers, the incremental gross margin might be lower than 30%. If it is led by differentiated coherent components and underutilized fab absorption, it could be 40%–50% or higher.

This calculation shows why the company’s margin structure matters more than the headline revenue recovery. A low-margin recovery would produce much less operating leverage than a high-margin recovery.

# ## Parameter table

Parameter	Value	Source	Tier	Measured / Derived / Assumed	Cross-verified
FY2024 total net revenue	$1,359M — `needs primary source validation`	Lumentum 10-K FY2024, EDGAR CIK 0001633978	1	Recalled from training; must be measured against filing	Single-source until verified
Cloud & Networking share	~79% — `needs primary source validation`	Lumentum 10-K FY2024 segment note	1	Derived from recall; must be measured	Single-source until verified
Industrial Tech share	~21% — `needs primary source validation`	Lumentum 10-K FY2024 segment note	1	Derived from recall; must be measured	Single-source until verified
FY2024 GAAP gross margin	~16%–18% — `needs primary source validation`	Lumentum 10-K FY2024	1	Recalled; must be measured	Single-source until verified
FY2024 non-GAAP gross margin	~30%–32% — `needs primary source validation`	Lumentum 8-K earnings release	1	Recalled; must be measured	Single-source until verified
Cloud & Networking segment gross margin	18% assumed	Explicit model assumption	N/A	Assumed	No
Industrial Tech segment gross margin	35% assumed	Explicit model assumption	N/A	Assumed	No
Incremental gross margin on recovery	30%–50% sensitivity	Model assumption	N/A	Assumed	No
Incremental opex on recovery	8% of incremental revenue	Model assumption	N/A	Assumed	No
Competitor Coherent fiscal 2024 revenue	`needs primary source validation`	Coherent Corp 10-K FY2024, EDGAR CIK 0000823182	1	Recalled; must be measured against Coherent filing	Single-source until verified
Competitor Coherent segment mix	`needs primary source validation`	Coherent Corp 10-K FY2024	1	Not asserted	No

# ## Competitor filing contradiction check

The most relevant direct competitor with SEC filings is Coherent Corp. (NYSE: COHR), successor to II-VI after its acquisition of Coherent Inc.

Lumentum’s fiscal 2024 narrative, as recalled from its earnings release and 10-K, emphasized:

Telecom carrier inventory digestion
Weak demand in the Cloud & Networking segment
Margin compression
Restructuring and impairment

Source: Lumentum Form 10-K for fiscal 2024, SEC EDGAR company index, CIK 0001633978 — exact accession number to be confirmed; access date 2026-04-26.

Coherent’s fiscal 2024 narrative, as recalled from its filings, emphasized:

Sequential growth in datacom transceiver demand tied to AI infrastructure
Stronger networking segment performance later in fiscal 2024
A more diversified portfolio including silicon carbide and materials

Source: Coherent Corp Form 10-K for fiscal 2024, SEC EDGAR company index, CIK 0000823182 — exact accession number to be confirmed; access date 2026-04-26.

If Coherent was already seeing AI-driven datacom growth while Lumentum’s Cloud & Networking segment was still declining or recovering slowly, then Lumentum may have:

Different customer exposure
Different product cycle timing
Weaker 800G/1.6T transceiver positioning than Coherent
Greater exposure to legacy telecom, which was still weak
Share loss in coherent components

This is not proof that Lumentum is structurally disadvantaged, but it is a clear contradiction in the two companies’ filings. One filing may present the same industry environment more optimistically than the other, or the two companies may simply serve different segments.

From training-data recall:

Lumentum fiscal 2024 revenue: approximately $1.36B — needs primary source validation
Coherent fiscal 2024 revenue: approximately $4.7B — needs primary source validation

If those figures are correct, Coherent is roughly 3–4× larger. Lumentum is therefore not the scale leader in photonics. Its market-share claims should be measured against Coherent’s larger revenue base and broader product portfolio.

Source: Lumentum 10-K FY2024 and Coherent Corp 10-K FY2024, SEC EDGAR company indexes, CIK 0001633978 and CIK 0000823182 — exact accession numbers to be confirmed; access date 2026-04-26.

A load-bearing claim—for example, “Lumentum is a leading supplier of 800G transceivers”—would require two independent sources:

Lumentum 10-K segment/product disclosure
A competitor filing or independent named-institution market study

Until that is verified, any such claim should be treated as single-source or unaided.

# ## Disconfirming Evidence

The thesis that Lumentum is a compelling investment would be weakened or invalidated if any of the following are found in primary filings or verified channel checks:

Cloud & Networking does not re-accelerate during the 800G/1.6T upgrade cycle. If Lumentum’s Cloud & Networking revenue remains flat or declines while Coherent and others report strong AI datacom growth, Lumentum is likely losing design wins or product relevance.

Gross margin stays below 25% on a non-GAAP basis for multiple quarters. That would indicate Lumentum is competing on price in commodity transceivers rather than selling differentiated coherent components.

Customer concentration increases again without a clear volume commitment. If a single customer again exceeds 20%–25% of revenue, the business becomes more fragile to order delays.

The Industrial Tech segment fails to recover margin. If commercial lasers remain weak while Cloud & Networking losses persist, there is no margin offset.

U.S.–China export restrictions are expanded. Lumentum has already been affected by restrictions on Huawei. Further restrictions could reduce the addressable market or force supply-chain redesign.

Competitors integrate more of the photonics stack. If Coherent, Cisco/Acacia, or Chinese module makers capture more coherent module revenue, Lumentum may be pushed into lower-value component sales.

The NeoPhotonics integration fails to produce enough synergies. If amortization and integration costs keep non-GAAP margins below target, the acquisition thesis weakens.

A technology shift to silicon photonics reduces the need for discrete InP lasers and modulators. If merchant switch vendors integrate silicon photonics engines, Lumentum’s discrete component franchise could be structurally impaired.

A large consumer-electronics customer diversifies away from Lumentum for 3D sensing. This would remove revenue and fab utilization.

Working capital and inventory builds without confirmed demand. If Lumentum builds inventory ahead of expected recovery and the recovery does not come, another writedown may follow.

Any of these would require revisiting the investment thesis even if headline revenue beats estimates.

# ## Sources

SEC EDGAR company index for Lumentum Holdings Inc — all 10-K filings:

sec.gov

SEC EDGAR company index for Lumentum Holdings Inc — 8-K filings:

sec.gov

SEC EDGAR company index for Coherent Corp — all 10-K filings:

sec.gov

Lumentum investor relations — SEC filings and press releases:

investor.lumentum.com

Accession numbers not asserted. Use the EDGAR indexes above to retrieve the exact accession numbers and filing dates for Lumentum’s fiscal 2024 10-K, fiscal 2024 earnings 8-K, and Coherent’s fiscal 2024 10-K. All recalled financial figures above are marked needs primary source validation and must be checked directly against those filings before any investment decision.

Lumentum sells components and modules that manipulate light for communications and industrial applications.

From the FY2024 Form 10-K product descriptions:

The answer is mixed.

In telecom, Lumentum has pricing power only in differentiated high-performance products such as certain coherent components and ROADMs; standard modules face severe average selling price declines. The 10-K explicitly warns of “average selling price erosion” and the need to reduce costs faster than price declines.
In AI/cloud datacom, supply-demand tightness for high-speed lasers and components can support pricing stability or increases, but this is cyclical and capacity-driven, not structural.
In industrial/consumer VCSELs and diode lasers, pricing is competitive.

Therefore, Lumentum cannot broadly raise prices without losing share. Its pricing power is product-specific and strongest in differentiated optical components with qualification/reliability barriers.

Source: Lumentum FY2024 Form 10-K, Risk Factors.

Lumentum’s product portfolio can be grouped into four main functional areas:

1. Telecom optical components and modules

These include:

Tunable lasers used in dense wavelength-division multiplexing (DWDM) systems.
Wavelength selective switches (WSS) and ROADM components that route individual wavelengths without converting every signal back to electrical form.
Pump lasers for erbium-doped fiber amplifiers and Raman amplifiers.
Modulators and receivers used in coherent optical transmission.
Coherent transceivers and components for high-speed long-haul, metro, and data-center interconnect applications.

These products enable network operators to increase fiber capacity without deploying new fiber. Source: Lumentum Form 10-K Business section — SEC EDGAR filing index; accession number: needs primary source validation.

2. Datacom optical transceivers and laser chips

Lumentum sells optical transceivers and photonic components for data-center switching. After the acquisition of Cloud Light Technology Limited, Lumentum increased its exposure to data-center transceivers, including modules aimed at high-speed links inside hyperscale data centers.

The technical function is to convert electrical signals into optical signals, send them over fiber, and convert them back at the receiving end. As data-center bandwidth grows, the number and speed of optical links per switch increase.

Source for acquisition and strategic rationale: Lumentum press releases and SEC filings. URL: investor.lumentum.com — accessed 2026-04-26. Exact press release date/accession number: needs primary source validation.

3. 3D sensing and laser-based sensing

Lumentum supplies vertical-cavity surface-emitting laser (VCSEL) arrays used in:

Consumer-electronics facial recognition.
Time-of-flight cameras.
Automotive or industrial LiDAR.
Proximity sensing and illumination.

These chips emit infrared light in a controlled pattern. They are critical for high-volume consumer devices where reliability, size, and power consumption matter.

4. Commercial and industrial lasers

Lumentum sells fiber lasers and diode lasers used in industrial manufacturing:

Cutting and welding metals.
Marking and engraving.
Semiconductor processing.
Medical and scientific applications.

The technical function is to generate a high-power, tightly controlled laser beam that can cut, weld, or mark material with precision. Kilowatt-class fiber lasers compete with CO₂ lasers and traditional machine tools in metal fabrication.

Substitutes vary by product line:

Tunable lasers compete with fixed-wavelength lasers, directly modulated lasers, electro-absorption modulated lasers, and silicon-photonic approaches, depending on reach and cost requirements.
ROADM/WSS compete with alternative optical-layer architectures, including simpler fixed multiplexers/demultiplexers and full digital-electrical switching, but those alternatives often require more transponders.
Optical transceivers compete with offerings from other optical-component vendors and, increasingly, with internally developed or silicon-photonic-based modules by large cloud customers.
VCSEL arrays compete with edge-emitting lasers, LED-based illumination, and alternative 3D-sensing architectures.
Fiber lasers compete with CO₂ lasers, solid-state lasers, and traditional non-laser manufacturing processes.

This substitution logic is consistent with the competitive-risk disclosure in Lumentum’s 10-K. Source: Lumentum Form 10-K Risk Factors — SEC EDGAR index; accession number: needs primary source validation.

Lumentum’s moat, if any, comes from:

Long qualification cycles in telecom and datacom. A new optical component must pass extensive reliability, thermal, and interoperability testing before a network-equipment maker adopts it.
Proprietary wafer fabrication and packaging know-how, especially in indium phosphide, gallium arsenide, and VCSEL manufacturing.
High-volume manufacturing yield required for consumer 3D sensing. A supplier must produce millions of laser chips with extremely low failure rates.
Wavelength-selective switch and ROADM complexity, which combines micro-optics, liquid-crystal or MEMS technology, software control, and precision packaging.
Customer switching costs driven by qualification, design-in, and supply-chain approvals.

These factors are qualitative and are described in Lumentum’s risk factors as competitive strengths and constraints; they are not quantified here. Source: Lumentum Form 10-K Business and Risk Factors — SEC EDGAR index; accession number: needs primary source validation.

Lumentum has limited pricing power in commodity optical transceivers and in standard industrial lasers. In those products, customers exert annual cost-down pressure and have alternative suppliers.

Lumentum may have more pricing power in differentiated products such as:

Wavelength selective switches.
High-reliability telecom lasers and amplifiers.
High-volume consumer VCSEL arrays where manufacturing yield and reliability matter.
Specialized coherent components.

However, whether Lumentum can raise prices without losing share is not a yes-or-no question. The correct answer is:

Pricing power is product-specific. For commodity transceivers and standard lasers, price increases are unlikely without share loss. For differentiated, qualification-heavy components, some pricing power exists, but it is constrained by customer concentration and annual cost-down expectations.

This conclusion is based on qualitative competitive-risk disclosures in the 10-K. Source: Lumentum Form 10-K Risk Factors — SEC EDGAR index; accession number: needs primary source validation.

Risks

AI data center capex cycle — if hyperscaler spending slows, the core growth engine stalls
LiDAR is a rounding error — if you're buying LITE for LiDAR exposure, you're getting 80%+ AI datacom
Telecom cyclicality — legacy telecom laser business faces demand lumpiness
China exposure in supply chain (InP substrates, GaAs wafers)
Competition from Coherent, II-VI, and emerging Chinese photonics suppliers

Coherent (COHR) — the only other independent high-volume InP EML/CW supplier; also sold out. The bottleneck is a de facto duopoly, not a monopoly.
Captive suppliers — Broadcom (internal EMLs), Mitsubishi Electric — limited merchant capacity.
Chinese InP photonics — Innolight, Eoptolink, Hisense, Source Photonics are Lumentum's CUSTOMERS today (transceiver makers), but China's InP fab buildout is the long-term substitute risk.
Silicon photonics / CPO — shifts some content from EML to CW lasers; Lumentum supplies both, hedging the architecture risk, but ASP mix could change.

Hyperscaler capex digestion in 2027: if AI cluster spend pauses, optical overcapacity hits within ~2 quarters — watch Q2-Q3 FY27 guide for the first sequential deceleration (guide below +20% QoQ = yellow flag).
Coherent wins share: if COHR's InP capacity ramp outpaces LITE's, pricing power erodes — watch the GM gap between the two.
CPO/ELS value shift: if ultra-high-power CPO lasers become a separate, lower-margin supply pool, LITE's EML pluggable franchise could be margin-compressed even as volume grows.
Dilution overhang: the $7.8B equitization already converted debt to equity; further equity issuance for InP fab capex would dilute the +83% revenue growth.
Customer concentration: a few transceiver makers account for a large share of Components demand — a single customer platform shift (to captive or Chinese lasers) would hurt.

Broadcom — the other large 200G EML merchant (CEO). Captive + merchant.
Coherent (COHR) — independent InP / transceiver peer; Cloud Light still trails it on module GM.
InnoLight / Eoptolink — Chinese module leaders; LITE is now first-to-market on some 1.6T SKUs (CEO) but admits GM still lags.
Chinese CW (70–100 mW) — CEO: not hitting high-power CPO/NPO or EMLs yet; the hole LITE is being asked to fill is still open.
Hyperscalers — NVIDIA is a $2B preferred holder + multibillion LTA (CEO; 10-K confirms the $2B preferred, not the LTA dollar amount). No hyperscaler ships a competing merchant InP EML.

Q1 FY27 print (early Nov 2026) that guides Q2 below +20% QoQ — first sequential deceleration = yellow flag on the capacity-cycle.
Customer A mix >30% of revenue without a second named platform — one buyer pause recreates FY2024.
200G EML share actually collapsing toward the 40–50% 1.6T SiPh mix before 3.2T EML share recovers.
Greensboro slips past early-calendar-2028 first revenue, or NPO does not pull into late-2027.
Module GM never closes vs InnoLight — Systems stays a volume filler that dilutes Components scarcity economics.
China tit-for-tat cuts InP substrates / isolators (10-K already flags indium/gallium/germanium export restrictions; AXT manufacturing is in China).

2026-08-17: FY2026 10-K filed (this note's primary gap vs 2026-08-16). Confirms Customer A 26.6%, NVIDIA $2.0B preferred, Greensboro $38M, $7.76B equitization, 89.7M shares.
2026-08-11: Q4 beat + Q1 guide $1.225–$1.275B; CPO/ELS/NPO/OCS language in EX-99.1.
2026-08-27: Deutsche Bank Tech Conference — FY2028 "$40 of earnings power"; OCS >$100M next quarter; NPO pulled forward; copper-wall physics; AXT + Japanese substrate still tight.
2026-07-26/29: AXT InP Capacity Reservation (6 years, $87M deposits) — now confirmed in AXT 8-K (was inferred in the 08-16 note).
2026-09-10: stock −5.4% (queue flag); no new 8-K since 08-11 other than Form 4/144s.

The thesis on Lumentum would be wrong if any of the following appear in later primary filings:

Cloud & Networking revenue stalls or declines again despite AI-datacom demand. The Q3 FY2025 10-Q shows Cloud & Networking grew 10.5% year-over-year—but if this decelerates or turns negative while the broader datacom optics market grows, Lumentum is losing share or suffering telecom drag.

Gross margin fails to expand or compresses. FY2024 GAAP gross margin was 19.0%. If the company’s differentiated product mix cannot lift margin toward historical levels, the profit recovery thesis fails.

Industrial Tech recovery proves cyclical rather than structural. The Q3 FY2025 44.5% Industrial Tech growth is off a depressed base. If that reverts, total growth will depend entirely on Cloud & Networking, which remains only moderately positive.

Customer concentration worsens and the largest customer reduces orders. The 10-K already discloses material customer concentration. A loss of a large OEM design win would remove revenue with little short-term replacement.

Competitor filings show share gains at Lumentum’s expense. Coherent’s 10-K is a direct competitor source. If Coherent reports strong datacom/optical growth while Lumentum’s Cloud & Networking growth remains low, that would contradict the investment thesis and show Lumentum is not participating fully in AI-linked demand.

Pricing power erodes further. The 10-K itself warns of average selling price erosion. If this intensifies in high-speed datacom modules—because of merchant silicon photonics or new capacity—Lumentum cannot raise prices and the margin story reverses.

Goodwill/impairment charges recur. The large FY2024 GAAP loss includes impairment charges. Additional impairments would indicate prior acquisitions are not generating expected cash flows.

If any of these appear in subsequent 10-Q/10-K or competitor 10-K disclosures, the Lumentum long thesis would weaken materially.

This section states what would prove the investment thesis wrong. Since no specific buy/sell thesis is asserted, I frame disconfirmation around the qualitative bull case: Lumentum has differentiated photonic products, cloud/datacom exposure, and a recovery in telecom spending.

Telecom spending does not recover. If carrier capital expenditures remain depressed or Lumentum’s telecom customers continue inventory digestion, revenue rebound will be delayed or absent.

Apple concentration risk materializes. If Lumentum loses share or pricing in consumer 3D sensing — for example, if Apple moves to an in-house design, second-sources the VCSEL supply, or changes sensing architecture — revenue and margin could fall sharply.

Datacom growth fails to offset telecom decline. If Cloud Light’s data-center transceivers do not scale, or if AI-related optical demand shifts toward internally developed or silicon-photonic modules, the strategic acquisition thesis fails.

Gross margin stays below the level needed for operating leverage. If non-GAAP gross margin does not recover with revenue, the company may remain close to breakeven or generate losses.

Coherent or other competitors win share in WSS/ROADM or coherent components. The moat from qualification and manufacturing is real but not absolute. Competitor innovations or aggressive pricing could erode market share.

Industrial laser demand deteriorates further. If manufacturing PMI-linked demand remains weak, the Industrial Tech segment will drag consolidated results.

Inventory charges or goodwill impairment recur. If past acquisitions underperform, further impairments would signal that the acquisition strategy destroyed value.

Customer vertical integration accelerates. If hyperscalers or telecom equipment makers bring more optical components in-house, Lumentum’s addressable market would shrink even if end-demand grows.

Pricing erosion exceeds volume growth. In commodity transceivers and standard lasers, annual price declines could overwhelm unit-volume gains.

A verified competitor filing contradicts Lumentum’s positioning. If Coherent’s 10-K shows stronger cloud growth, better industrial demand, or share gains in WSS/coherent components, the relative-positioning argument weakens.

Open questions

[ ] What percentage of LITE's revenue is actually LiDAR/sensing lasers vs. AI datacom vs. consumer 3D sensing?
[ ] Is the LiDAR laser business growing independently or just riding the overall photonics boom?
[ ] When does InP capacity expansion relieve the sold-out supply constraint?
[ ] Q4 FY2026 actuals vs. $987.7M / $2.97 estimates — did they beat and raise?

[ ] What is the exact Cloud & Networking vs Industrial Tech (VCSEL/LiDAR) segment revenue split in the FY26 10-K (filed ~late Aug 2026)?
[ ] Confirm the AXT InP substrate agreement terms (duration, $87M prepayment) from the FY26 10-K.
[ ] How much share dilution did the Q4 FY26 convertible equitization actually produce (share count pre/post)?
[ ] Does Coherent's InP capacity ramp close the supply gap and start compressing the GM premium?

[x] Cloud & Networking vs Industrial Tech split — 10-K retired the two-segment reporting; FY26 is Components 66.5% / Systems 33.5%. VCSEL/3D sensing is not broken out.
[x] AXT InP terms — confirmed in AXT 8-K: through 2031-12-31, $43.5M + $43.5M deposits, annual take-or-pay.
[x] Dilution from Q4 equitization — 74.6M FY26 basic WAS vs 89.7M common on 2026-08-14 (+20.2%); Q4 non-GAAP diluted ~101M.
[ ] Does Coherent's InP ramp close the 200G EML supply gap and compress the 50.4% GM? Watch COHR next print vs LITE GM.
[ ] Named identity of Customer A (26.6%) and whether OCS >$100M is that same buyer.
[ ] Cloud Light module GM vs InnoLight — does calendar 2027 manufacturing fix actually land?

Research verdict

Excellent business, right vehicle, wrong thesis labeling. LITE is an AI data center photonics play that happens to sell lasers to LiDAR makers. The picks-and-shovels logic is correct — you don't need to pick the LiDAR winner — but the shovel factory's main customer is AI datacom, not LiDAR. Buy LITE for the multi-year sold-out AI laser bottleneck. The LiDAR exposure is a volatility kicker.

STRENGTHENED (upgraded from the 08-03 "excellent business" narrative). The Q4 beat-and-raise plus the new CPO/ELS/NPO/OCS catalyst cluster confirms the InP-laser bottleneck thesis AND extends it from pluggable interconnects into in-rack co-packaged optics — a materially larger TAM. Thesis-label correction: LITE is a Connectivity/RF (Layer 6) optical-interconnect bottleneck, not primarily a Perception/Sensing (Layer 7) LiDAR name — the VCSEL/LiDAR business is the sidecar.

Primary and supporting material

Sources

18 sources preserved from the latest qualified research update.

Historical long-form synthesis

Lumentum: The Sold-Out Laser Bottleneck Powering AI Data Centers

1,027 words · Research as of 4 Aug 2026

Preserved research context: this long-form synthesis reflects the evidence and valuation snapshot available on 4 Aug 2026. Use the current snapshot above for the latest signal, conviction, target and market-cap values.

Investment Thesis

Lumentum occupies one of the tightest supply bottlenecks in the AI infrastructure buildout: the indium phosphide (InP) lasers that drive 800G and 1.6T optical interconnects between GPU clusters. These lasers are the physical layer that converts electrical signals to light and back — and Lumentum's lasers are sold out through 2027, potentially 2029. The CEO has stated the company is "significantly under-shipping demand." When a critical component is on allocation and capacity cannot be added quickly (InP wafer fabs take 2-3 years to build and qualify), the supplier owns pricing power and revenue visibility that few semiconductor companies can match.

The business has transformed under CEO Alan Lowe (formerly of MaxLinear). Revenue has roughly tripled in ~5 quarters as the AI datacom cycle ignited. Q2 calendar 2026 delivered approximately $426M (a beat), and full-year guidance was raised to $1.75-1.8B. Non-GAAP operating margins of 32-35% in strong periods reflect the pricing power of a sold-out supplier. The Q4 FY2026 report (Aug 11) is expected to show $960M-$1.01B in revenue — roughly half the annual total in a single quarter, underscoring the acceleration.

The LiDAR laser business (VCSEL and edge-emitting lasers for Ouster, Luminar, Innoviz, Hesai, and others) is a genuine but secondary thesis component. It is the picks-and-shovels logic in its purest form: Lumentum sells lasers to every LiDAR maker, so LiDAR adoption benefits Lumentum regardless of which architecture wins. But the shovel factory's main customer is AI datacom, not LiDAR — the LiDAR exposure is a volatility kicker, not the core thesis.

Physical AI / Value-Chain Relevance

Lumentum spans two layers of the physical-AI stack: Perception & Sensing (VCSEL and edge-emitting lasers for LiDAR and 3D sensing) and AI/ML Infrastructure (InP-based EML, CW, and coherent lasers for 800G/1.6T transceivers). The AI datacom business is the dominant revenue driver, but both layers are authentic physical-AI exposure.

The LiDAR relevance is real but should be sized correctly. Every LiDAR unit — whether from Ouster, Luminar, Innoviz, or Hesai — contains laser diodes that Lumentum or Coherent likely supplied. As LiDAR volume scales from millions to tens of millions of units, laser content per unit grows (higher channel counts, longer range, better resolution). However, LiDAR laser revenue is dwarfed by AI datacom revenue — the ratio is approximately 10:1. This is not a critique of the LiDAR thesis; it is a calibration. Buy Lumentum for the multi-year sold-out AI laser bottleneck. The LiDAR upside is a free call option.

Catalysts

Q4 FY2026 earnings (Aug 11, 2026). The immediate catalyst. Guidance of $960M-$1.01B would represent explosive sequential growth. Any LiDAR design-win commentary, InP capacity expansion updates, or guidance for FY2027 will move the stock.
InP capacity expansion reaching production. Lumentum has secured long-term InP wafer supply through AXT (agreement through 2031, $87M prepayment). As new capacity comes online, the sold-out constraint eases — but demand is growing faster than supply, so sold-out conditions may persist even with capacity additions.
800G → 1.6T transition. As NVIDIA shifts from 800G to 1.6T optical interconnects for next-generation GPU clusters, the laser content per transceiver increases. Higher data rates require more sophisticated (and more expensive) lasers — a content-per-unit growth story.
LiDAR laser design wins at Western automakers. While secondary to the AI thesis, LiDAR design wins at volume OEMs (Toyota, Hyundai, GM) would validate the sensor-buildout option and could re-rate the LiDAR portion of the business.

Positioning / What the Market May Be Missing

At $779.89, Lumentum trades at a forward P/E of ~44x — reflecting the growth premium but not fully pricing the structural supply-demand imbalance in InP lasers. The stock is down -14.4% in the last week and -16.0% in the last month, caught in the AI infrastructure rotation alongside Coherent and other photonics names.

Three factors the market underappreciates:

"Sold out through 2027" is not marketing language. The InP laser supply chain is genuinely capacity-constrained at the substrate level (AXT, Sumitomo) and the device level (Lumentum, Coherent). Adding capacity requires years of fab construction and customer qualification. This is a structural shortage, not a cyclical spike.
The AXT agreement through 2031 is a strategic asset. $87M in prepayments locks in InP wafer supply for the better part of a decade. Competitors without similar agreements will be substrate-constrained.
LiDAR is a free option on a business that already works. The AI datacom thesis does not require LiDAR adoption to succeed. If LiDAR volumes disappoint, Lumentum's revenue and earnings still grow on AI photonics alone. But if LiDAR volumes surprise to the upside, it is incremental margin on existing manufacturing infrastructure.

Risks and What Invalidates the Thesis

AI capex cycle reversal. This is the number-one risk. If hyperscaler spending on GPU clusters slows, the core growth engine stalls. Book-to-bill provides a buffer (backlog converts to revenue even if new orders slow), but the multiple would compress sharply.
CPO (co-packaged optics) disruption. If co-packaged optics are deployed broadly before 2028, it could reduce or eliminate demand for the pluggable transceivers that Lumentum's lasers go into. This is a technology-transition risk, not an imminent threat — CPO is still in early adoption and faces thermal, yield, and serviceability challenges.
Competition from Coherent and Chinese suppliers. Coherent is the primary competitor in InP lasers, and Chinese photonics suppliers are investing aggressively. Lumentum's moat is manufacturing scale and qualification cycles, not unique technology — if Chinese suppliers close the quality gap, pricing power erodes.
Telecom cyclicality. The legacy telecom laser business (~15-20% of revenue) faces demand lumpiness and carrier capex cycles. This is a drag during telecom downturns, even if AI datacom is booming.

What to Watch Next

Q4 FY2026 earnings (Aug 11). Revenue vs. $960M-$1.01B guidance. Datacom vs. telecom vs. 3D sensing segment splits. LiDAR design-win commentary. FY2027 outlook.
InP substrate supply chain. Monitor AXT earnings and capacity announcements — Lumentum's growth is gated by InP wafer availability.
NVIDIA 1.6T optical interconnect roadmap. Any shift in NVIDIA's transceiver architecture (pluggable vs. CPO, silicon photonics vs. InP) directly affects Lumentum's addressable market.
LiDAR volume data. Ouster, Luminar, and Innoviz quarterly unit shipments — a proxy for Lumentum's LiDAR laser revenue trajectory.

Connectivity & Positioning exposure in Physical AI.

Catalysts

Invalidation

Latest change

Earlier changes

Reported holdings as-of

Disclosed investment history

Moat & position

Deep research update

1. Telecom optical components and modules

2. Datacom optical transceivers and laser chips

3. 3D sensing and laser-based sensing

4. Commercial and industrial lasers

Sources

Lumentum: The Sold-Out Laser Bottleneck Powering AI Data Centers

Investment Thesis

Physical AI / Value-Chain Relevance

Catalysts

Positioning / What the Market May Be Missing

Risks and What Invalidates the Thesis

What to Watch Next

Related companies

Applied Optoelectronics

Coherent Corp

Cisco Systems Inc

Arista Networks

Get a weekly digest of what’s changing.