Lumentum Just Doubled Its Revenue. It Still Can’t Ship Fast Enough.

There is a category of earnings beat that looks strong on the surface and turns out to be conservative once you read the fine print. Lumentum’s fiscal Q4 report, released after the close on August 11, belongs in that category.

The headline numbers are already extraordinary. Revenue hit $1.01 billion for the quarter ended June 27, a 109% jump year over year. Adjusted EPS of $3.23 came in $0.28 above the consensus estimate of $2.95. Non-GAAP gross margin expanded 1,260 basis points to 50.4%, and operating margin more than doubled to 36.6%. By any conventional measure, Lumentum delivered one of the cleanest beats in the AI infrastructure trade this earnings season.

But the number that actually changes the investment calculus is buried in management’s candid admission about what Lumentum cannot do. CEO Michael Hurlston has said the company continues to lag demand, with the supply-demand imbalance for EML laser chips and pump lasers running greater than 30%. The company is, in effect, revenue-constrained by its own manufacturing capacity, not by customer appetite. Every dollar of Q4 revenue represents a floor, not a ceiling, for what the market is trying to buy.

The Q1 Guide Is More Aggressive Than It Looks

Management guided Q1 fiscal 2027 revenue to a range of $1.225 billion to $1.275 billion, with a midpoint of $1.25 billion. That midpoint sits roughly 8% above the analyst consensus of $1.16 billion heading into the report. Adjusted EPS guidance of $4.05 to $4.35, with a midpoint of $4.20, topped consensus by nearly $0.60.

Hurlston said the Q1 guide brings Lumentum to its target operating model more than a quarter ahead of schedule. The company’s mid-term financial targets call for a $2 billion quarterly revenue run-rate at 40% non-GAAP operating margin, on an approximately 18-to-24-month timeline. At the current sequential growth rate, that timeline looks increasingly conservative.

The revenue trajectory in fiscal 2026 alone illustrates the pace: Q2 came in at $665.5 million, Q3 at $808.4 million, and Q4 at $1.01 billion. Each quarter brought a new record. The Q1 2027 guide extends that streak toward $1.25 billion. Four quarters, four records, and demand still running ahead of supply.

What the 1.6T Transition Actually Means for Margins

The most underappreciated driver in Lumentum’s earnings story is not volume growth. It is average selling price inflation driven by the shift from 800-gigabit to 1.6-terabit networking.

The 1.6T transition requires 200-gigabit-per-lane EML laser chips rather than the 100-gigabit versions that dominated 800G systems. The ASP for a 200G EML runs roughly twice that of a 100G EML. Because Lumentum supplies EML chips to virtually every major transceiver manufacturer globally, that ASP tailwind flows through regardless of which transceiver brand a hyperscaler ultimately buys. The company is the toll road, not the vehicle.

This dynamic is already visible in the margin line. Non-GAAP gross margin went from 47.9% in Q3 to 50.4% in Q4, a 250-basis-point sequential improvement on top of the 540-basis-point gain the quarter before. The revenue surge is not simply a volume story. Pricing is doing real work.

Nvidia Anchored the Supply Chain. Greensboro Is the Wildcard.

In March, Nvidia committed $2 billion to Lumentum as part of a broader $4 billion strategic investment split between Lumentum and Coherent. The deal includes a multibillion-dollar purchase commitment and future capacity access rights for advanced laser components. That commitment gives Lumentum the balance-sheet visibility to expand capacity with confidence, and it anchors the demand side of the equation at the highest level of the AI infrastructure stack.

The capacity response is real but slow. Lumentum is converting a 240,000-square-foot former Qorvo facility in Greensboro, North Carolina into a new U.S. indium phosphide manufacturing site. The facility is expected to ramp production in mid-2028. Hurlston has been direct about the timeline mismatch: even as new capacity comes online, the company expects to remain supply-constrained because demand is accelerating at roughly the same pace as capacity is being brought online.

That is not a bullish qualifier. It is the core of the investment case. A company that is sold out through 2028 and building a new InP site while still under-shipping by more than 30% is not a commodity supplier facing pricing pressure. It is a structural bottleneck in a multi-trillion-dollar buildout.

OCS and CPO: Two Growth Engines Still in Early Innings

The current revenue base is almost entirely driven by EML laser chips, pump lasers, and 1.6T transceivers. Two additional product lines, optical circuit switches and co-packaged optics, have barely started contributing and represent the next phase of the growth curve.

OCS technology reduces data center power consumption by routing light instead of electrical signals, with published estimates suggesting OCS deployments can cut a data center’s power draw by up to 40%. Lumentum’s OCS backlog exceeded $400 million as of earlier this year, with management expecting most of that to ship in the second half of calendar 2026. The OCS total addressable market, per Mordor Intelligence, sits at $8.06 billion in 2026 and is projected to reach $12.71 billion by 2031.

CPO, which integrates optical connections directly onto the switch or server chip rather than keeping them as external pluggable modules, is the longer arc. Lumentum has secured a multi-hundred-million-dollar order for CPO light engines, with shipments expected to scale in 2027. Hurlston has sized the CPO opportunity at over $5 billion in incremental revenue for Lumentum alone. Barclays has argued broad CPO adoption may take until 2029 or 2030 to fully materialize, which is the realistic bear case on timing but not on direction.

The Revenue Ramp the Market Is Underpricing

Consider the sequential math. Q4 at $1.01 billion. Q1 guided to $1.25 billion. That is a $240 million sequential step-up in a single quarter. If OCS shipments accelerate on schedule and the 1.6T transceiver ramp continues, the path to Hurlston’s $2 billion quarterly target could close faster than the 18-to-24-month timeline implies.

Narrow-linewidth laser shipments grew more than 120% year over year in Q3. Pump laser shipments grew 80%. Both products are sold out for the foreseeable future. The pricing dynamic on 200G EMLs is accretive to margins. And Lumentum is negotiating take-or-pay and prepayment agreements with key customers to lock in long-term capacity access, a structural feature more common in capital-intensive commodities than in photonics companies.

The supply picture tells a simpler story than the income statement. When a company reports 109% revenue growth and then explains that demand still exceeds supply by more than 30%, the revenue number is not the measure of demand. It is the measure of what the factory can currently produce.

Risks Worth Taking Seriously

The obvious risks are customer concentration and execution. Two customers account for the majority of expected OCS volume. The 1.6T transceiver ramp depends on yields improving and component shortages across DSPs and TIAs not compressing shipments from the supply side. Barclays and others have flagged the possibility that CPO adoption timelines could slip, deferring a significant chunk of the incremental opportunity to the 2029-2030 window.

There is also a valuation question that no analyst can fully resolve at this growth rate. Shares have rallied dramatically over the past year, reflecting a re-rating from legacy optical company to AI infrastructure essential. That re-rating compresses the margin of safety for investors entering now.

Insider selling has been notable, with 22 recent transactions running net negative. That is worth flagging, though insiders frequently sell for reasons unrelated to fundamental outlook at companies where stock-based compensation is a significant portion of total pay.

Finally, Greensboro is a 2028 event. The two-year gap between now and meaningful new fab output leaves Lumentum’s growth trajectory dependent on squeezing more from existing fabs and on the Rose Orchard facility, which is already described as sold out indefinitely.

The Bigger Picture: Photonics Is the New Bottleneck

The AI infrastructure buildout has moved through a sequence of bottlenecks. First it was advanced packaging. Then HBM memory. Then power. Photonics is increasingly being identified as the next constraint, and Lumentum’s own CEO compared the indium phosphide shortage to the memory-chip crunch at its worst.

Goldman Sachs has projected that AI demand will continue to exceed compute center capacity for years. That demand does not flow through GPUs alone. It flows through the optical interconnects that connect GPUs to each other across scale-up clusters and scale-out networks. Every AI training run, every inference request at hyperscale, travels through components that Lumentum makes. That is not a temporary position. It is a structural one.

Nvidia recognized this when it committed $2 billion to secure Lumentum’s capacity. Hyperscalers are recognizing it through the order backlog and take-or-pay negotiations. The market is recognizing it through LITE’s stock performance. The question worth asking now is not whether the demand is real. The Q4 results answered that. The question is whether the market has fully priced the gap between what Lumentum can currently ship and what its customers are trying to buy.

Final Thought

A 109% revenue beat with Q1 guidance 8% above consensus is the kind of result that generates headlines about AI momentum and optical networking. Those headlines are correct. They are also incomplete.

The more precise framing is this: Lumentum just reported the best quarter in its history, raised guidance above what Wall Street expected, and simultaneously confirmed it cannot manufacture products fast enough to meet actual demand. That combination, record revenue plus a greater-than-30% supply gap plus a $2 billion Nvidia commitment plus OCS and CPO barely contributing yet, describes a company where the reported numbers are not the upper bound of the opportunity. They are the current production limit.

For investors watching the AI infrastructure buildout and trying to identify where physical constraints will shape the next phase of the trade, the photonics bottleneck is worth watching closely. Lumentum appears to be at the center of it.

Disclaimer: This editorial is for informational purposes only and does not constitute investment advice. The information presented may not be complete or accurate. Investing in securities involves risk, including the possible loss of principal. Past performance is not indicative of future results. Always conduct your own due diligence and consult a licensed financial professional before making any investment decisions.