Lumentum vs Coherent: One AI-Optics Thesis, Two Multiples — 28x Sales Against 12x
Lumentum (NASDAQ: LITE) and Coherent (NYSE: COHR) are the two Western names every AI-optics conversation eventually circles back to. Both have been pulled out of telecom-cyclical obscurity and re-rated into large-caps by the same force: NVIDIA’s data-center buildout and its need to move colossal amounts of data between accelerators with optics instead of copper. Both took a $2 billion equity investment from NVIDIA in March 2026, the identical capital-plus-purchase-commitment template, on the same day the company committed to locking in its photonic supply chain.
And yet the market prices them very differently. That gap — same thesis, two multiples — is the entire story.
Two businesses, one cycle
Lumentum splits into Components (laser chips, pump lasers, the VCSELs that go into iPhones) and Systems (optical modules, optical circuit switches, industrial lasers). Its edge is depth in the laser chip itself, plus a vertical move into finished transceiver modules via its Cloud Light acquisition. The growth legs management points to are optical circuit switches (OCS), where backlog already runs well past $400 million, and co-packaged optics (CPO), where an incremental multi-hundred-million-dollar order is deliverable in the first half of calendar 2027.
Coherent is the larger, more diversified, vertically integrated player. It designs and manufactures its own InP EMLs, CW lasers, VCSELs, silicon photonics, detectors, and finished modules in-house, outsourcing only the DSP. Its datacom transceiver franchise spans both 800-gigabit and 1.6-terabit products, and it carries legacy industrial-laser and materials revenue that the optics growth has to drag along.
The valuation gap
The two have converged on similar market caps despite very different revenue bases. At recent prices — LITE around $891, COHR around $393 — the picture looks like this:
| Metric | Lumentum (LITE) | Coherent (COHR) |
|---|---|---|
| Price | ~$891 | ~$393 |
| Market cap | ~$69B | ~$77B |
| Revenue (TTM) | ~$2.5B | ~$6.6B |
| Price / sales | ~28x | ~12x |
| Avg analyst target | ~$1,113 | ~$381 |
| Implied vs target | ~+25% | roughly flat / slightly underwater |
| 52-week high | $1,085.68 | $440.00 |
The core divergence: nearly the same market cap, but Coherent does roughly 2.6x Lumentum’s revenue. So on a sales basis the market pays a much steeper multiple for Lumentum — around 28x sales versus 12x — pricing the slope of its growth rather than its current scale. Coherent is the bigger, lower-multiple name whose legacy revenue optically dilutes the optics growth.
The re-rating
Both have moved violently. Lumentum’s 52-week range runs from the low $80s to over $1,085; a year ago the street’s average price target sat around $135. Coherent printed an all-time high of $440 in early June 2026 and is up several hundred percent over twelve months. Both have since pulled back into the zone JPMorgan flagged as a buying opportunity — partly real volatility, partly a Coherent shortage warning that put execution risk back in the frame.
A note on the targets: they lag. Analysts have spent this entire cycle revising upward after the fact, wrong-footed by NVIDIA-capex step-changes that linear models did not anticipate. Lumentum’s average target still shows headroom to spot; Coherent’s price has run up to, and at times past, its last revision. That cuts both ways — targets lag on the way up, and they would lag on any down-cycle too.
The moat: it comes down to indium phosphide
The structural moat for both is not the finished transceiver. Anyone can assemble a module. The moat is the InP EML laser chip underneath it. Chinese assemblers remain dependent on third-party EML chips drawn from the same constrained InP fabrication base, and the highest-performance applications require hyperscaler qualification cycles of 12 to 18 months that those assemblers have generally not completed for the most advanced products. That qualification lag, plus the supply constraint, is the real barrier.
Within that shared moat, the two differ:
- Lumentum holds an estimated 50 to 60 percent share of high-end laser chips. Its vertical integration into modules gives it better cost positioning when supply is tight, because it can prioritize its own production.
- Coherent is betting on a manufacturing leapfrog. In August 2025 it began production on what it calls the world’s first 6-inch InP platform, claiming roughly 4x the devices per wafer at about half the cost of legacy 3-inch lines, ramping across multiple sites. Lumentum is migrating 3-inch to 4-inch; Broadcom is believed to be on a similar path. If Coherent’s yields hold at scale, that is a genuine cost-structure advantage — and it is the single most important variable in the LITE-versus-COHR debate.
There is also a geopolitical layer: Western hyperscalers face growing pressure to source critical AI components from trusted, non-China supply chains, and the InP chokepoint is a Western-controlled node in a way module assembly is not.
Growth projections
Splitting company fundamentals from stock targets:
Fundamentals. Lumentum is roughly doubling revenue off a smaller base — Q3 FY26 revenue grew about 90% year over year, with a Q4 guide implying an annual run-rate near $4 billion. Coherent grows off a 3x-larger base: consensus has revenue moving from roughly $7.05 billion in FY26 toward $9.45 billion in FY27, around 34% growth, with management signaling FY27 growth will exceed FY26 and the 4x InP capacity expansion improving both supply and margins. EPS estimates scale accordingly — Coherent toward the high single digits in FY27 and into double digits by FY28.
Stock targets. Lumentum’s average target still sits above spot, implying meaningful upside. Coherent’s average target has been overtaken by the price — not because analysts turned bearish (ratings remain overwhelmingly positive) but because the stock outran the last revision cycle. Expect those targets to be marked up at the next print, the same way Lumentum’s were ratcheted from $135 to over $1,100 across a year.
The silicon photonics question
The most cited long-tail risk is silicon photonics displacing InP. The framing deserves a correction: silicon cannot efficiently emit light, so a silicon-photonics or CPO system still needs an external laser source — and that source is almost always InP. The disruption mostly hits the pluggable transceiver module, not the laser chip that is Lumentum’s and Coherent’s deepest moat. Even the silicon-photonics winners have to buy lasers.
The names worth tracking in that ecosystem:
- Platform giants: NVIDIA and Broadcom, the two poles — one the holistic architect embedding photonics into the compute fabric, the other the modular scaler propagating CPO through established switch supply chains. Both already in initial CPO production.
- Acquirers: Marvell, which agreed to buy Celestial AI for up to $5.5 billion; Cisco, which holds Acacia; Intel, the original silicon-photonics incumbent.
- Disruptor startups: Ayar Labs (optical I/O chiplets,
$3.75B valuation, AMD and NVIDIA backing), Lightmatter ($4.4B, shipping its Passage photonic interposer), Celestial AI, and nEye Systems (wafer-scale OCS). - Enablers: TSMC, whose COUPE packaging platform is becoming the default; plus GlobalFoundries, Corning, Fabrinet, and advanced-packaging houses.
The real risk to Lumentum and Coherent is therefore not that InP disappears. It is margin compression as they shift from selling a high-value finished module to selling a laser engine into someone else’s package.
What it comes down to
Both are clear winners of the supercycle. The open question that defines the trade is whether Lumentum’s laser-chip leadership survives the 1.6T transition and the CPO era, and whether Coherent’s 6-inch InP gamble proves out on yield.
Lumentum is the purer, higher-multiple, higher-beta bet on chip leadership. Coherent is the broader, lower-multiple, vertically integrated bet whose InP cost gamble could re-rate it up toward Lumentum’s multiple — or disappoint on execution. The two estimates are not independent: they are two expressions of one thesis — sustained AI-factory buildout — priced at roughly 2.4x the sales multiple on Lumentum for the steeper growth slope. If you believe the capex cycle holds for years, LITE expresses it with more torque; if you want the same tailwind with revenue cushion and a cheaper entry, COHR is the lower-multiple version.
This is analysis, not investment advice. Figures move with every revision and were current as of mid-June 2026; valuation multiples and price targets in particular should be re-checked around the next earnings prints.