LELOUCH · COHR · LITE

Nvidia Bought Both. Which Laser Do You Own?

In March 2026, Nvidia did something telling: it put roughly $2 billion into each of Coherent (COHR) and Lumentum (LITE), with multi-year purchase commitments attached. These two make the lasers that light up co-packaged optics — the technology that moves data as light right next to the switch and GPU silicon inside an AI data center. You don't invest $2B twice unless the laser is a genuine bottleneck. Both just posted blowout June quarters. But the stocks have split hard: LITE has run roughly 9x in a year and sits near all-time highs, while COHR corrected about 35% off its peak. Same Nvidia anchor, same end market, very different setups. This is the head-to-head — the better business versus the better entry — and the one factor that may decide it: who owns the indium phosphide.

Same Job, Two Different Companies

Both supply the light source for optical interconnects, but they're built differently:

  • COHR — Coherent. The broad, vertically-integrated one. It owns its own indium-phosphide (InP) laser fabs, and supplies the whole stack — CW lasers, external laser sources (ELS), InP modulators, VCSELs, silicon photonics, and the fiber array units (FAUs) that couple fiber to chip. Bigger, more diversified (it still has a legacy industrial-laser business), scale revenue ~$7B.
  • LITE — Lumentum. The focused laser/transceiver pure-play. It makes Nvidia's "prime laser source" for CPO — the ultra-high-power (UHP) laser — plus transceivers and a second GaAs/VCSEL light-source path. Smaller (~$3B revenue) but growing explosively and running at much higher margins.

In one line: COHR is the vertically-integrated arms-maker; LITE is the specialist that happens to hold Nvidia's single most important CPO laser socket. Both matter. They're not quite the same bet.

Where the Prices Are

As of the August 25 close — note how differently the two have traded:

Ticker Price Off 52wk high 1yr range RSI(14)
COHR$288.14−35%$86 – $44045
LITE$885.57−18%$124 – $1,08654

LITE went up roughly 9x off its low and is only 18% below its all-time high — it's the momentum leader, still in its uptrend. COHR made a smaller (~5x) run and has since given back a third of it, sitting on its 200-day. That gap is the whole tension: you're choosing between the name that's working and priced for it, and the name that's cheaper because it stumbled.

What the Quarter Actually Said

Both June quarters were spectacular. Here's the honest part — on the raw numbers, LITE's was the better quarter.

COHR (FQ4'26) LITE (FQ4'26)
Revenue (YoY)$2.05B (+34%)$1.01B (+109%)
Non-GAAP EPS (YoY)$1.74 (+74%)$3.23 (+267%)
Gross margin40.2%50.4%
FY26 revenue$7.12B (+23%)$3.01B (+83%)
Free cash flow (FY26)≈ −$1.0B≈ +$225M
Market cap~$56B~$80B
EV / sales (TTM)~8x~26x
Forward P/E~30x~41x

Figures from company FQ4 FY2026 releases (quarters ended late June 2026). Valuation metrics approximate. COHR's negative FCF reflects roughly tripled capex funding InP/CPO capacity; LITE's balance sheet strengthened this quarter as it equitized ~$1.1B of convertible debt.

Read that table and the paradox is obvious. LITE grew twice as fast, earns ten points more gross margin, and generates cash — yet trades at three times COHR's EV/sales. The market isn't confused; it's paying up for the purer, faster, higher-margin operator that holds Nvidia's prime CPO laser socket. COHR is bigger and cheaper, but it's diluting the growth story with a declining industrial segment and burning $1B a year on capex to build capacity. Two very different risk profiles wearing the same "Nvidia laser supplier" label.

The One Factor That May Decide It: Indium Phosphide

Here's the piece that doesn't show up cleanly in the financials. The lasers at the heart of CPO are built on indium phosphide — a III-V semiconductor that's far harder to manufacture at scale than silicon, and the genuine supply bottleneck for the whole optical-interconnect ramp. Whoever controls InP capacity controls the choke point.

COHR owns its InP fabs — in Sherman, Texas; Sweden; and Switzerland — and is doubling output. That vertical integration is the strongest argument in its favor: in a world where InP is scarce, owning the substrate is a moat, and it's why Nvidia is funding COHR's capacity build directly. LITE, by contrast, is more exposed to the InP bottleneck it openly acknowledges — and has hedged smartly by developing a second, GaAs-based VCSEL light-source path to route around it. So the strategic question underneath the stock question is: does owning InP capacity (COHR) beat having a flexible multi-path laser platform and better current economics (LITE)? Reasonable people land on both sides — and that's exactly the conversation worth having next.

The Verdict

This one is closer than the construction comparisons, so I'll split it by what you're optimizing for.

Better business right now: Lumentum (LITE). It's firing on every cylinder — +109% revenue, +267% EPS, 50% gross margins, positive free cash flow, a de-risked balance sheet, and Nvidia's prime CPO laser socket. If you buy the best operator and are willing to pay up near all-time highs for continued momentum, LITE is defensible. The catch is entirely price: at ~26x sales and ~41x forward earnings, it's priced for years of flawless execution, and it's the name with the most to lose if the optics trade rolls over.

Better risk-adjusted entry: Coherent (COHR). This is the one I lean toward at current levels. You get comparable end-market exposure and the deepest vertical integration in the technology that actually gates CPO scaling — owned InP capacity — at roughly a third of LITE's sales multiple, after a 35% correction that reset the valuation. You're paying for the industrial drag and the capex burn, but you're buying the bottleneck itself, with Nvidia funding the build. For someone who missed the 9x run and wants theme exposure without chasing a stock at its highs, COHR offers more margin of safety per dollar.

How I'd frame it: LITE is the higher-quality, higher-priced momentum name; COHR is the broader, cheaper, more strategically-moated laggard. If forced to own one here, I take COHR for the InP moat and the entry — but I'd understand anyone who wants LITE's cleaner financials and is willing to pay for them. This is a case where both can win if CPO ramps; the difference is how much you're paying for certainty.

The Headwinds That Could Break This

  • AI-capex digestion. Optics is the highest-beta expression of the AI-buildout trade. Both names have already been volatile; a hyperscaler capex pause would hit these harder than almost anything else in the complex.
  • CPO timing. Production co-packaged optics is a 2026–27+ story, and some switch vendors think broad adoption slips to 2028–29, favoring cheaper linear pluggable optics in the meantime. If CPO ramps slower than priced, both de-rate — LITE more, given its multiple.
  • The indium-phosphide bottleneck cuts both ways. It's COHR's moat, but scaling InP output is genuinely hard; a yield stumble hurts COHR directly and constrains LITE's supply.
  • Valuation. LITE at ~26x sales near record highs has priced in near-perfection. COHR's is more reasonable but still rich, and its −$1B FCF needs AI demand to stay strong to justify the capex.
  • Nvidia concentration. Nvidia is the whale for both. The $2B investments are a huge vote of confidence — and also a dependency. A roadmap change or a move toward insourcing would land hard.

Bottom Line

Nvidia buying into both tells you the laser is a real chokepoint and both companies are on the right side of it. Lumentum (LITE) is the better operator today — faster, higher-margin, cash-generative — but you pay all-time-high prices for it. Coherent (COHR) is the better risk-adjusted entry: cheaper after a 35% pullback, and it owns the indium-phosphide capacity that is the genuine bottleneck for the entire CPO ramp. I'd take COHR here for the moat and the margin of safety, with full respect for LITE's execution. Either way, the thing to understand before you buy is the material underneath the laser — which is exactly where this conversation goes next.

Prices and technicals pulled live from Yahoo Finance as of the August 25, 2026 close. Fiscal Q4 2026 fundamentals and the March 2026 Nvidia investments from company releases and StockTitan/24-7 Wall St./Zacks coverage. Valuation metrics (market cap, EV/sales, forward P/E) are approximate and depend on share-count and consensus assumptions. Levels and multiples are point-in-time and will move.