Skip to content
primarysourced Photonics sector Lumentum
LITE
~7 min read · 1,512 words ·updated 2026-04-29 · confidence 25%

Risk register

A formalized version of the bear-case categories, organized as a risk register with probability and impact bands. Probability is qualitative (low / medium / high). Impact is framed as the percentage hit to FY2028 revenue or EPS in the case where the risk materializes. Mitigation describes the primary structural offset.

Risk categories — summary table

IDRiskCategoryProbability (next 18 mo)Impact (FY2028 base case if realized)Inflection-monitor signal
R1NVDA reduces purchase commitmentCustomer concentrationLowHigh (-25% revenue)NVDA earnings or product roadmap signals reduced optics requirement
R2NVDA in-sources opticsCustomer concentrationLow (near-term) / Medium (3–5 yr)Very High (-40%+ revenue)NVDA acquires InP supplier; internal CPO engine vendor partnerships
R3NVDA’s own datacenter-GPU growth deceleratesCustomer concentrationLow–MediumHigh (-30% revenue)NVDA quarterly revenue decel; AMD/custom-silicon wins
R4Hyperscaler AI capex pauseDemand cycleLow–MediumHigh (-30% revenue, multiple compression)Hyperscaler Q3/Q4 CY2026 capex guides cut
R5CPO timeline slip 12+ monthsArchitectural transitionMediumModerate (-10% revenue, deferred)NVDA Spectrum-X / NVLink CPO product delays
R6LPO captures more pluggable share than expectedArchitectural transitionMediumLow–Moderate (margin headwind on Cloud Light)LITE module-segment ASP commentary deteriorates
R7Alt source-laser tech displaces InPArchitectural transitionLow (near-term)Very High (multi-year decline)Quantum-dot / monolithic-on-Si lasers reach hyperscaler qual
R8Sumitomo / HG Genuine third source emergesDuopoly erosionLow–MediumModerate (margin -300–500 bps)Industry trade-press hyperscaler qual announcements
R9Hyperscaler in-sources EML chipsDuopoly erosionVery LowVery High (decade-scale headwind)Hyperscaler M&A activity in InP fab space
R10New San Jose fab construction slipExecutionMediumModerate (-10% FY2028 revenue, deferred)Lumentum capex commentary; first-tape-out delays
R11Cloud Light pluggable margin compressionExecutionMediumModerate (margin -200–400 bps)LITE earnings module-segment commentary
R12China export-control escalationRegulatoryMediumLow–Moderate (-5–10% revenue)BIS rule changes; III-V additions to ECCN list
R13HSR / CFIUS clearance delayRegulatoryVery LowLow (process risk; doesn’t affect operating)Lumentum 8-K disclosures; FTC second request
R14Convertible-note dilution larger than market expectsCapital structureLowLow–Moderate (EPS -5–10%)Lumentum 10-Q disclosures on diluted share-count; capped-call settlement
R15Multiple compression in macro risk-offValuationMedium–HighVery High (price -30–50% even if operating intact)Real-rate increases; AI-narrative-momentum reversal
R16Apple VCSEL franchise decline acceleratesSegment-specificMediumVery Low (Industrial Tech is 14% of revenue and falling)Apple/Coherent multi-source disclosures
R17Telecom carrier capex slips againSegment-specificLow–MediumLow–Moderate (-3–7% revenue)Tier-1 carrier capex guides cut

Risk descriptions and detail

R1 — NVDA reduces purchase commitment

Probability rationale: Low because the NVDA $2B equity investment is balance-sheet capital, not a take-or-pay revenue contract — meaning NVDA has skin in the game on the success of the LITE relationship. NVDA wouldn’t dilute its own equity returns by de-prioritizing the partnership it just invested in. Renegotiation could happen on price/terms but a wholesale withdrawal is contrary to the structural setup.

Impact rationale: NVDA-allocated capacity revenue is meaningful share of the FY2027–FY2028 incremental revenue ramp; reduction reverses a meaningful share of the bull case. Downside ~$1–2B revenue.

Mitigation: Bilateral structure (parallel COHR investment) means NVDA is committed to the merchant duopoly model. Multibillion purchase commitment is multi-year with implied take-or-pay protections (terms not publicly disclosed; ⚠).

R2 — NVDA in-sources optics

Probability rationale: Near-term low (NVDA’s parallel investment in both LITE and COHR is anti-in-sourcing signal). 3–5 year medium (NVDA’s silicon-photonics capability development is substantial).

Impact rationale: If NVDA captive-optics replaces merchant-EML purchase, LITE’s NVDA revenue goes to zero over 3–5 years and the duopoly’s pricing power weakens.

Mitigation: NVDA’s parallel COHR investment + multiyear strategic R&D collaboration are the structural guards. NVDA’s capital allocation suggests preference for partnership over vertical integration through CPO transition. ⚠ subject to revision if NVDA acquires an InP supplier.

R3 — NVDA’s own datacenter-GPU decelerates

Probability rationale: NVDA datacenter-GPU growth decelerates eventually; the question is timing. Low–Medium probability of meaningful deceleration in next 18 months.

Impact rationale: NVDA’s volume slowdown reduces optical-component pull-through; broader hyperscaler AI capex is correlated.

Mitigation: LITE has hyperscaler-direct customers beyond NVDA-stack; merchant-EML demand beyond NVDA; ROADM/wave-shaper telecom revenue provides counter-cyclical ballast.

R4 — Hyperscaler AI capex pause

Probability rationale: Capex moderation is more likely than capex pause; the AI infrastructure narrative remains robust through CY2026.

Impact rationale: Direct revenue impact on Cloud & Networking segment; multiple compression effect on share price arguably larger.

Mitigation: Order book filled through 2028 per management; AI-cluster-build is multi-year construction (capex doesn’t pause cleanly because builds-in-progress continue).

R5 — CPO timeline slip

Probability rationale: Medium given the engineering complexity of NVLink CPO scale-up. Even Spectrum-X scale-out CPO has execution risk.

Impact rationale: CPO slip is mostly revenue deferral, not destruction. Pluggable transceivers continue selling into the slip window.

Mitigation: Lumentum’s chip-layer franchise is technology-agnostic between pluggable and CPO. CPO slip extends pluggable-module revenue runway, partially offsetting.

R6, R7, R8, R9 — Architectural and duopoly erosion risks

These are slower-moving risks (5+ year horizon) with uncertain probability and high tail-impact. The base-case assumes none of them materialize through FY2028. The bull case assumes the same. The bear case stacks low-probability versions of each.

R10 — New San Jose fab construction slip

Probability rationale: Fab builds typically slip by 6–18 months relative to initial schedule. NVDA-funded urgency may compress but not eliminate slip risk.

Impact rationale: Deferred FY2028 revenue impact; some capex stranded if demand-mix shifts during slip window.

Mitigation: Towcester UK fab provides ramp capacity in interim; existing San Jose fab is being expanded as well.

R11 — Cloud Light pluggable margin compression

Probability rationale: Pluggable-transceiver commoditization is a structural trend; the question is the rate.

Impact rationale: Cloud Light is roughly 25–30% of Cloud & Networking revenue; margin compression there has limited consolidated-margin impact.

Mitigation: Vertical integration with LITE-internal EML chips supplies cost-advantage. Hyperscaler-direct relationships create some pricing protection.

R12 — China export-control escalation

Probability rationale: Escalation is a multi-year pattern. Medium probability of further restrictions that affect Lumentum’s residual China revenue.

Impact rationale: Lumentum’s China revenue exposure is already significantly reduced from pre-2018 levels. Incremental hit is bounded.

Mitigation: Western hyperscaler AI demand is so much larger than residual China demand that further compression doesn’t break the thesis.

R13 — HSR / CFIUS clearance delay

Probability rationale: Very Low — these reviews historically clear in similar transactions.

Impact rationale: Process risk only; doesn’t affect operating business. Could delay NVDA voting/conversion but not the equity-investment proceeds (which were paid at issuance).

R14 — Convertible-note dilution

Probability rationale: Some level of dilution is certain (the convertibles are deep in the money). The magnitude is the variable.

Impact rationale: Already partially absorbed in current diluted-share count of 87.8M. Capped-call hedges limit incremental dilution.

Mitigation: Capped-call hedges; exchange transactions completed in late 2025 reduced uncertainty on settlement mechanics for 2026/2029 notes.

R15 — Multiple compression

Probability rationale: Medium–High over any meaningful holding period; high-multiple AI-photonics names compress in any risk-off period.

Impact rationale: Even with operating execution intact, share price can move 30–50% in a multi-month risk-off window.

Mitigation: Operating cash flow and FCF inflection in FY2028+ provide intrinsic-value support; balance sheet is robust enough to absorb multiple compression without capital-structure stress.

R16, R17 — Segment-specific risks

VCSEL/Industrial Tech and telecom-transport are smaller and don’t drive the thesis. Risk impact is bounded.

Composite risk view

The dominant risk vectors for the LITE thesis are:

  1. R15 (multiple compression) — high probability, moderate-to-high impact, hardest to mitigate
  2. R4 (hyperscaler capex pause) — moderate probability, high impact, partially mitigated by order book
  3. R5 (CPO timeline slip) — moderate probability, moderate impact, well-mitigated
  4. R10 (new fab execution slip) — moderate probability, moderate impact, well-mitigated
  5. R2 (NVDA in-sourcing) — low near-term probability, very high tail-impact, structurally mitigated by bilateral COHR investment

The customer-concentration risks (R1, R2, R3) have low individual probabilities but high tail-impact. The Lumentum bull case effectively pays a premium for the NVDA-direct alignment; the corollary is that the bear case pays a discount when that alignment evolves.

Sources