There’s a familiar story playing out in AI investing right now. Everyone is chasing GPU stocks. Everyone is watching Nvidia. But here’s the thing — some of the smartest capital is quietly piling into something else. They’re buying the company that Nvidia itself wrote a $2 billion check to lock up. They’re buying the business that is literally turning away orders it cannot fulfill. Not because it doesn’t want the revenue. Because it physically cannot make its products fast enough. That company is Lumentum Holdings (NASDAQ: LITE). And the story is more interesting than most people realize.
The Demand Nobody Can Satisfy
Here’s a number worth sitting with. Lumentum’s management has disclosed that demand for its indium phosphide lasers exceeds supply by more than 30%. The gap is growing — not closing. The company is increasing laser output by 50% year-over-year at its existing fabs. It shipped twice as many laser chips as it did a year ago. And it still cannot keep pace. Customers keep calling. Orders keep stacking up. The waiting list is not getting shorter.
The scale of demand is hard to grasp. One customer recently approached Lumentum asking for 1 billion CW lasers per year. Starting with a B. It used to be millions. Now it’s billions. CEO Michael Hurlston has stated plainly in public forums: “We are sold out through calendar 2027.” That’s before the new Greensboro, North Carolina fab even comes online. Management doesn’t describe this as a temporary blip. The supply-demand imbalance is structural — and expected to persist for years.
This scarcity translates into real pricing power. Lumentum has been actively raising prices on its most constrained product lines. It is also negotiating long-term agreements with take-or-pay provisions and upfront prepayments. Customers are being asked to help fund capacity expansion in exchange for supply guarantees. This is not commodity-supplier behavior. It’s the behavior of a company whose customers cannot afford to be left without supply. That dynamic tends to be very good for margins.
The Infrastructure Bet Nvidia Made Before Everyone Else
In March 2026, Nvidia made a $2 billion equity investment in Lumentum — alongside a multi-billion dollar purchase commitment. This wasn’t passive. Nvidia was locking up a scarce, critical resource: indium phosphide laser capacity. It made a parallel deal with Coherent. Together, the two transactions cornered a significant chunk of the world’s indium phosphide manufacturing. The signal was hard to miss: optical interconnects are the next bottleneck in AI infrastructure.
Why does this matter? Because Nvidia doesn’t write $2 billion checks casually. This is a company that maps its supply chain years in advance. The investment signals that photonics is now viewed as mission-critical. Lumentum’s ultra-high power lasers enable AI chips to communicate at speeds copper cables simply cannot match. At higher data rates and longer distances, copper hits a physical wall. Optics doesn’t. That is the core thesis.
The commercial relationship goes beyond the equity deal. Lumentum is currently sole-sourced for certain co-packaged optics (CPO) laser applications. A multi-hundred million dollar purchase order is slated for fulfillment in H1 2027. The company is also developing a turnkey External Light Source (ELS) module. It’s a ready-to-deploy package for hyperscalers that lack deep in-house optical expertise. For every ELS shipped, Lumentum generates roughly 2x the revenue of a standalone laser sale.
The Hidden Products & Opportunities Most Investors Are Missing
Most investor attention on Lumentum focuses on EML lasers and cloud transceivers. Both matter. But two other product lines deserve more attention — and get far less. The first is the Optical Circuit Switch (OCS). The second is what management calls the “scale-across” portfolio — components that synchronize high-bandwidth traffic across multiple data center buildings. Both are ramping. Both carry margins meaningfully richer than the transceiver business.
The OCS story has moved fast. One year ago, management was discussing $100 million in quarterly OCS revenue as an aspiration. Today, it’s targeting $400 million across the back half of calendar 2026 — and expects OCS to exceed $1 billion in calendar 2027. A new multi-year, multi-billion dollar OCS agreement was recently signed with one of its three hyperscaler customers. The product replaces power-hungry electrical packet switches with low-latency, low-power optical alternatives.
The scale-across portfolio is similarly underappreciated. Pump laser shipments grew 80% year-over-year in Q3 FY2026. Narrow linewidth laser shipments grew over 120% year-over-year — for the ninth consecutive quarter. Management has flagged that the supply-demand imbalance in scale-across may actually be worse than in EMLs — a surprise even to internal teams. These are high-margin components feeding directly into coherent optical networking buildouts at the world’s largest hyperscalers. They are effectively sold out.
Margins, Momentum, & The Path To $2 Billion
Not long ago, Lumentum was generating roughly $300–$350 million per quarter and barely breaking even at the operating level. Q3 FY2026 looked different. Revenue hit a record $808 million — up 90% year-over-year. Non-GAAP operating margin reached 32.2%, a gain of over 2,100 basis points year-over-year. Q4 FY2026 guidance points to a revenue midpoint of $985 million — another record. The pace of financial transformation has been notable.
The gross margin story deserves attention. Non-GAAP gross margins reached 47.9% in Q3 FY2026 — up 1,270 basis points year-over-year. The gains came from better factory utilization, a richer product mix, and selective price increases. The company’s stated long-term target is 40% non-GAAP operating margins at a $2 billion quarterly revenue run rate. Management believes this milestone can be reached within 18 to 24 months of its OFC investor event in early 2026.
The path to $2 billion quarterly revenue runs through four growth drivers. Transceivers, OCS, scale-out CPO, & scale-up CPO. None are fully ramped today. Scale-up CPO is estimated to be 3 to 4x larger than the initial scale-out opportunity. The Greensboro fab — expected online in 2028 — adds a potential $5 billion in annual revenue capacity. That figure is excluded from current financial targets. If execution holds, the stated $2 billion quarterly run rate represents a floor, not a ceiling.
Final Thoughts: A Note On Valuation
So, is this a story worth watching? Almost certainly. Is it a story worth paying any price for? That’s the harder question — and it deserves a direct answer.
As of May 11, 2026, Lumentum trades at 30.4x LTM revenues and a trailing P/E of 200x. The LTM EV/Gross Profit multiple sits at 74.4x. On forward numbers, the picture is more measured. NTM EV/Revenue of 15.1x and NTM EV/EBITDA of 36x. But those multiples assume the $2 billion revenue path materializes on schedule. The NTM levered free cash flow yield stands at just 1.5%. There is limited margin of safety embedded at current prices.
The business fundamentals are real. The supply-demand imbalance is documented. The Nvidia partnership carries weight. The margin expansion is visible in reported numbers. But at these multiples, the market has already priced in a great deal of flawless execution. A supply chain disruption, a CapEx pullback from hyperscalers, or a delay in CPO adoption could weigh heavily at these levels.
This is a high-conviction growth story — trading at a high-conviction growth price. The opportunity appears genuine. So does the bar that needs to be cleared to justify it. Worth watching closely. Worth sizing thoughtfully.
Disclaimer: We do not hold any positions in the above stock(s). Read our full disclaimer here.
