Skip to content
primarysourced Photonics sector Coherent
COHR
~7 min read · 1,507 words ·updated 2026-04-29 · confidence 70%

Bull case

The Coherent Corp bull case rests on four interlocking pillars that distinguish it from the cleaner two-pillar Lumentum bull case. Coherent’s vertically-integrated breadth — InP devices, silicon photonics, transceiver assembly, ROADM/wave-shaper, SiC materials, industrial fiber lasers — provides a richer optionality space but also a more complex narrative. Each pillar below is independently load-bearing; the combination is what makes COHR a high-conviction AI-photonics name with diversified capital-structure protection. Each pillar has confirmation paths (visible signal milestones that should advance the bull case) and counter-signals (telltales that would weaken it).

Pillar 1 — NVDA-direct alignment via the $2B March 2026 common-stock placement

The single most concrete catalyst that re-rated COHR in CY2026 was the March 2, 2026 NVIDIA common-stock placement: 7,788,161 shares at $256.80/share = $2.0B cash, plus a multi-billion-dollar purchase commitment and future access-and-capacity rights for advanced laser and optical-networking products (Coherent 8-K; NVIDIA press release).

The bull-case interpretation:

  • NVDA committed balance-sheet capital, not just a long-term purchase order — a stronger signal of supply-chain commitment than a take-or-pay contract alone.
  • The capacity dedication aligns Coherent’s incremental investment to NVDA’s roadmap.
  • Multi-year strategic R&D collaboration on advanced optics technology implies CPO co-development at minimum.
  • The bilateral structure (parallel Lumentum investment) preserves duopoly economics rather than tilting toward single-supplier vertical integration.
  • The instrument structure (common stock vs Lumentum’s preferred) reflects relative bargaining position — Coherent’s broader scale and balance-sheet flexibility allowed cleaner terms.

Confirmation paths:

  • NVDA-specific revenue contribution growing as a share of D&C revenue (visible in management commentary or segment disclosure)
  • New / expanded NVDA volume commitments
  • Specific CPO scope identification in follow-on agreements
  • Sherman TX 6-inch InP fab capacity-doubling milestone hit on schedule (Q4 CY2026)

Counter-signals to monitor:

  • NVDA accelerating internal optics development (CPO in-house assembly, in-house EML capacity) beyond the existing partnership
  • NVDA reducing or repricing its purchase commitment to Coherent
  • A third major customer demanding similar capacity-allocation guarantees, signaling supply tightening at Coherent’s expense

Pillar 2 — Vertically-integrated breadth as a structural moat

Coherent’s vertical-integration depth — InP device fab + silicon-photonics process + transceiver assembly + ROADM/wave-shaper + SiC substrates + industrial fiber lasers — is unique in the photonics industry. The bull-case interpretation:

  1. More of the AI-photonics value-stack captured — Coherent ships finished pluggable transceivers (with InP source lasers fabricated in-house), capturing roughly 35–45% finished-module GM rather than just 50–70% chip GM with limited revenue per unit of customer demand.
  2. Industrial-segment ballast smooths cyclical revenue — when AI capex digests, Industrial provides absolute-revenue stability (even if modest growth); when industrial weakens, AI-cycle continues.
  3. CPO optionality at multiple value-stack layers — Coherent has the option to supply CPO at chip layer, SiPh-engine layer, or finished CPO module (vs Lumentum’s chip-focused position); whether the option is realized depends on NVDA architecture choices but it’s a richer optionality space.
  4. Deleveraging completed — net debt approaching zero post-NVDA, post the $437M FY2025 + $400M Q1 FY2026 paydowns + Bain Capital dividend waiver. The capital structure is no longer a thesis-constraining variable.

Confirmation paths:

  • Sustained 38–40%+ non-GAAP gross margin through FY2027
  • D&C share of revenue rising past 75% with margin holding steady (mix-shift compounds)
  • Sherman TX 6-inch InP fab cost-advantage realized at scale (60% die-cost reduction visible in margin expansion)
  • Industrial segment recovering from CY2024–2025 weakness

Counter-signals to monitor:

  • Industrial segment continuing to weaken (would compress operating leverage)
  • SiC pricing pressure intensifying beyond DENSO/Mitsubishi long-term agreement floor
  • Vertical-integration scope generating execution drag (visible in working-capital growth or capex creep)

Pillar 3 — InP EML duopoly economics + 6-inch InP cost advantage

Coherent and Lumentum together control ≥80% of merchant InP EML supply. The structural-margin protection mechanism is supply-constrained pricing power on the leading-edge 200G/lane node, with the duopoly’s combined capacity falling short of demand growth at least through CY2027.

Coherent’s specific differentiator: the world’s first 6-inch InP scalable wafer fab, located in Sherman TX (announced March 2024; ramping CY2025–2026; doubling internal capacity by Q4 CY2026 ✓). The 6-inch wafer-size advantage delivers:

  • 4× more chips per wafer vs 3-inch
  • ~60% reduction in die cost vs 3-inch baseline
  • Generation-ahead vs Lumentum’s 4-inch San Jose / Towcester legacy + planned 6-inch new fabs

Coherent management commentary on Q2 FY2026 framed 1.6T module gross margins exceeding 800G levels “particularly early in the lifecycle,” aided by 6-inch InP cost advantages (Futurum Group) ◐.

Confirmation paths:

  • Continuing supply-constrained framing on quarterly earnings calls
  • 200G EML ASP increases visible in management commentary
  • Sustained 30%+ non-GAAP operating margin through FY2027 (currently ~20%)
  • COHR-LITE pricing parity preserved (no race-to-the-bottom on contract pricing)
  • No publicly-announced third-source qualification at hyperscaler scale

Counter-signals to monitor:

  • Margin expansion stalls or reverses
  • Sumitomo, HG Genuine, or another challenger announces hyperscaler qualification at volume
  • COHR or LITE announces material price concessions to defend share
  • Hyperscaler in-sourcing of EML chip production (e.g., Google, AWS, Microsoft develop captive InP)

Pillar 4 — 2028+ CPO optionality with NVIDIA-named partner status

Co-packaged optics is publicly framed by NVDA and the broader industry as a 2027–2028+ commercial-volume opportunity. NVIDIA’s roadmap (Spectrum-X / Quantum-X Photonics) puts scale-out CPO at 2026–2027 and scale-up NVLink CPO at 2028.

Coherent is a named CPO source-laser partner at NVDA ✓. The bull-case interpretation:

  1. CPO modules consume more InP EML chips per unit of bandwidth than pluggables — net InP-chip-volume positive even as module-form-factor changes.
  2. Coherent’s 200G/lane EML capacity feeds CPO transceivers as much as pluggable transceivers — the source-laser franchise is technology-agnostic to module form factor.
  3. Vertical-integration into CPO module assembly is a credible second-order opportunity — Coherent has the in-house silicon-photonics + assembly capability that Lumentum lacks.
  4. The R&D collaboration with NVDA on advanced optics technology likely includes CPO co-development, locking COHR into the NVDA-stack CPO ecosystem.

Confirmation paths:

  • Spectrum-X / Quantum-X Photonics shipment volumes hit announced milestones
  • Coherent CPO-product announcements (module lines, qual programs, customer wins)
  • NVDA-COHR follow-on agreements specifically identifying CPO scope
  • CPO module assembly capacity additions at Coherent (Sherman TX or other site)
  • Hyperscaler qualifications of Coherent CPO products beyond NVDA

Counter-signals to monitor:

  • CPO timeline slip (NVDA pushes Spectrum-X / NVLink CPO out by 12+ months)
  • Linear-Pluggable-Optics (LPO) or other non-CPO architecture wins more 1.6T/3.2T share than expected, displacing CPO transition timing
  • CPO-engine vendor consolidation outside Coherent (e.g., Broadcom or TSMC SiPh consortium dominate at NVDA’s switch level, excluding Coherent SiPh)

Composite bull-case framing

The composite bull case argues:

  1. AI capex grows ≥30% annually 2026 → 2028, driving optical-transceiver TAM from $26B in 2026 to $40–60B in 2028 (LightCounting baseline ⚠).
  2. Coherent’s D&C revenue grows 25–30% CAGR FY2026 → FY2028, reaching $7–8B by FY2028 (capturing the duopoly’s share of the AI-optics layer + vertical-integration advantage).
  3. Non-GAAP operating margin expands to 24–26% as operating leverage compounds and ASP discipline holds.
  4. NVDA-allocated capacity adds $1–2B+ of incremental revenue from FY2027 onward, with multi-year visibility.
  5. CPO transition is additive (not substitutive) to the EML chip layer, with Coherent’s vertical-integration franchise pivoting successfully to CPO source-laser + potential CPO module assembly.
  6. Free cash flow inflects in FY2028 as capex tail-end coincides with revenue scaling, supporting share-count optimization (Series B convert clearance, potential repurchase post-FCF inflection).
  7. Industrial segment stabilizes — China-capex weakness moderates; Industrial returns to flat-to-modest growth, providing absolute-revenue ballast without compressing margins.
  8. Multiple-compression is delayed by sustained execution; the current AI-photonics-supercycle multiple regime persists into 2027–2028.

This stack drops to FY2028 non-GAAP EPS of approximately $9–11 with a 35–45× P/E multiple — implying $315–500 per share by mid-2028.

What the bull case does NOT require

Worth flagging explicitly:

  • The bull case does NOT require AI capex acceleration beyond the current trajectory — moderation is acceptable
  • The bull case does NOT require NVDA concentration to shrink — current concentration with multibillion purchase commitment is a feature
  • The bull case does NOT require Industrial recovery beyond stabilization
  • The bull case does NOT require Coherent to win every CPO architecture — partial CPO participation is sufficient
  • The bull case does NOT require the SiC franchise to inflect — flat SiC is acceptable; DENSO/Mitsubishi JV insulates the parent capex burden
  • The bull case does NOT require flawless capex execution — modest slip on the Sherman TX fab is absorbed in the timeline

Sources