Lumentum (LITE.US) earnings are about to be released under the spotlight! Will the "short storage, long optics" trade get another boost?

According to Zhishang Finance APP, Lumentum (LITE.US), the core participant in the “Google TPU AI Computing Power Chain” and a key supplier of optical/optical module components for both the “Nvidia and AMD AI GPU AI Computing Power Chains,” will announce its FY2026 Q4 results after the US market closes on August 11th, Eastern Time. If Lumentum, which has received strategic investment from Nvidia and is bound by a long-term optical component procurement agreement, releases results and guidance that significantly exceed the market consensus, it will be a major positive catalyst for Lumentum, the “AI Super Bull Stock” which has surged 140% since 2026 and rocketed 345% throughout 2025. This will also benefit the recent global surge in the investment theme of “Standing in the Light”—the hot “AI Data Center Optical Interconnect/Optical Communication Industry Chain” sector.

For this upcoming quarter, according to institutional compilations of Wall Street analyst expectations, Lumentum’s Q4 revenue (calendar Q2) is expected to range from $960 million to $1.01 billion, with NON-GAAP EPS forecasted between $2.85 and $3.05.

Consensus compiled by Zacks Investment Research shows analysts expect revenue of about $989 million, implying a substantial YoY increase of approximately 105.65%. Consensus EPS is about $2.99, revised up by 3 cents over the past 30 days; in contrast, Lumentum’s EPS in the same period last year was just $0.88.

As shown above—the combined chart of Lumentum’s stock price and its EPS surprise performance—Lumentum has beaten Zacks-consensus Wall Street analyst EPS estimates for four consecutive quarters, with an average beat of 9.59%. For a leading AI computing power chain company with ever-rising analyst growth forecasts, this is already a strong result.

The main reason Lumentum has been an “ultimate bull” in 2025 so far is its ability to benefit from both the Google TPU AI computing power chain and the Nvidia-dominated AI GPU power chain. Whether it’s the Nvidia AI GPU infrastructure cluster or Google’s TPU cluster (TPU adopts the AI ASIC technical route), “the optical interconnect and CPO silicon photonic high-speed interconnect capability represented by Lumentum” is indispensable.

Lumentum has been included in the S&P 500 index as well as the Nasdaq 100 index—known as the “global tech stock bellwether”—meaning Lumentum now qualifies for passive allocation from two core US benchmark indices. This further strengthens its institutional investor base as a core growth stock in the AI optical interconnect space.

Inclusion in the benchmark indices elevates Lumentum from a “niche supplier in the AI data center optical interconnect chain” to a global core technology asset pool, indicating that the AI computing power investment narrative is rapidly moving from “GPU/ASIC single-point computing race” to an “AI agent-driven full-stack inference computing system.” In this narrative shift, data center CPUs, storage chips, and the optical interconnect/optical communication industry chains could be the big winners.

The strong recent stock performance of Lumentum and other data center optical product suppliers also stems from market news indicating that “AI chip champion” Nvidia (NVDA.US) is adjusting its next-generation AI rack computing system architecture. The next-gen AI cluster, Rubin Ultra, might reduce per-rack HBM configuration, switching instead to an optical interconnection architecture for linking AI racks. With the continued surge in HBM/DRAM and NAND prices, the focus of AI infrastructure investment is shifting from GPU/TPU single-point performance and HBM capacity to overall data center architectural efficiency, where high-speed optical interconnect is likely becoming the critical direction.

From 800G to 1.6T and 3.2T: Lumentum Bets on “Optics in, Copper out” in AI Data Centers, High-End Lasers Enter Supply Shortage Cycle

Lumentum’s FY2026 Q4 results are expected to benefit from persistently high demand for high-speed optical devices driven by its cloud and AI product portfolio, with both component and systems segments performing strongly. Of the 21.3% sequential revenue increase expected for the reported quarter, over half is projected to come from data center optical components, with the remainder mainly driven by new contributions from high-speed optical transceivers (modules) and OCS (Optical Circuit Switches).

Demand for EML laser chips and scale-across products—including pump lasers and narrow linewidth laser modules—may remain especially robust. In FY2026 Q3, narrow linewidth laser shipments grew 120% YoY and pump laser shipments increased 80%; for the foreseeable future (2–3 years), these products are effectively sold out.

Another major catalyst for FY2026 Q4 is likely the ramp-up of 1.6T optical transceiver volume. Management expected higher 1.6T shipments entering the quarter, including the integration of internally sourced continuous-wave (CW) lasers. At the same time, consistently improving yields and declining waste rates are expected to boost transceiver profitability. The migration from 800G to 1.6T provides high per-unit value and favorable pricing for the company: Lumentum notes that the 200G EML for 1.6T averages double the sales price (ASP) of the 100G EML currently used for 800G modules.

Concurrently, a favorable product mix, disciplined pricing, high capacity utilization, and operating leverage are expected to further drive profit margin expansion. These factors lifted Q3 non-GAAP gross margin by 540 bps QoQ to 47.9%, and operating margin by 700 bps to 32.2%.

Despite favorable demand, supply constraints may limit Lumentum’s ability to fully meet customer needs in FY2026 Q4. Management notes the EML supply-demand gap exceeds 30%, and pump laser shortages are even more severe, forcing the company to allocate capacity between customers. Meanwhile, booming demand from hyperscalers like Google for Lumentum’s OCS product lines is also held back by supply chain tightness.

As shown above, Lumentum (LITE.US) stock has dramatically outperformed its sector. Year-to-date (YTD), Lumentum has surged 127.4%, far ahead of the S&P IT Computers & Tech sector’s 16.9%. It has outperformed optical interconnect/module rivals Coherent (COHR.US) and Ciena (CIEN.US), but lagged Marvell Technology (MRVL.US), which posted YTD stock returns of 81.1%, 72.7%, and 147.6% respectively.

Lumentum and Coherent are competitors in the high-speed optical module/transceiver field for data center and AI infrastructure optical interconnects. Ciena is a leading supplier of optical networking hardware, software, and services for inter-data center connections. Marvell Technology is also a strong competitor in the AI/data center optical networking sector, squaring off with Lumentum, Coherent, and Ciena in AI optical products.

Lumentum is not cheap by valuation metrics. On a next-twelve-month (NTM) price/sales (P/S) basis, Lumentum currently trades at 10.93x, above the sector and most peers. The sector’s average is 6.54x, with Ciena at 7.63x and Coherent at 6.42x. Marvell Technology trades at about 13x P/S, so there’s a significant Lumentum discount to Marvell.

As shown above, Lumentum (LITE.US) stock is trading at a premium. LITE’s outlook depends on whether AI computing infrastructure demand can remain robust for the long term.

As AI infrastructure increasingly shifts from copper to optics, the per-unit value of optical components is poised for a multi-year expansion, with Lumentum in a favorable position to benefit. Beyond its EML and optical transceiver leadership, the company sees four key growth drivers—AI data center OCS, scale-out, scale-up, and transceivers—several of which currently have only a limited revenue contribution. Management expects OCS and scale-out will become more significant, with scale-up possibly seeing a major inflection starting in 2027.

The company has highlighted the market migration towards 1.6T, and ultimately 3.2T, connection technology, which will drive up the value of high-speed optical devices in AI training/inference systems. Lumentum has a strong franchise in differentiated high-performance lasers, with barriers to entry such as reliability, performance, and client yields likely allowing substantial market share capture.

Scale-up interconnections and CPO/NPO present extremely attractive long-term opportunities. Lumentum anticipates that as optical interconnects move deeper into and between AI racks, scale-up will greatly increase component intensity—potentially tripling it—and create a market far larger than scale-out. According to management, once scale-up is deployed at scale, optical usage could rise at least threefold; in parallel, data center CPO demand remains robust, as does demand for CW lasers. With continuously expanding customer collaborations and new application scenarios, OCS-related product lines have immense growth potential, including more port configurations and potential in-rack applications.

In the longer term, Lumentum plans to extend its optical component portfolio from lasers further into photonic integrated circuits (PICs), photodiodes, laser drivers, and possibly transimpedance amplifiers (TIAs), expanding its addressable content in future optical systems.

Applied Optoelectronics Earnings Ignite “Long AI Optical Interconnect” Trade: Will Lumentum, With All-Stack EML to CPO Coverage, Spark a Further Rally?

Lumentum enters FY2026 Q4 with strong business momentum, underpinned by surging AI optical demand, accelerating 1.6T transceiver volume, tight EML and scale-across supply, a favorable product mix, and pronounced operating leverage. Crucially, many of its largest opportunities—OCS, scale-out CPO, especially scale-up optics—are still in early-stage adoption. Lumentum’s differentiated laser portfolio, expanding data center high-speed optics footprint, and deep exposure to growing AI power infrastructure connectivity needs give it a solid base for long-term revenue and profit growth.

Recently released results from Applied Optoelectronics (AAOI.US) provide a powerful “demand cross-validation” for the optical interconnect sector. Q2 revenue reached $191.9 million (up ~86% YoY), with data center revenue rising from $44.8 million to $107.7 million; 800G product shipments more than doubled sequentially, and the company expects 800G and 1.6T demand to outstrip capacity through at least mid-2027. Notably, current monthly capacity is only ~200,000 units but is expected to reach 650,000 by year-end 2026—highlighting that the top industry issue is not orders, but securing enough lasers, DSPs, TIAs, and packaging lines to build modules in time.

Post-earnings (August 7), AAOI jumped about 9%, Coherent 13.4%, and Lumentum 6.5%, indicating that capital markets view AAOI’s supply shortfall as a heavy demand signal for LITE, COHR, and other upstream optical platform companies, whose high-speed component capacity is lagging fast-growing demand for AI training/inference cluster connectivity.

With coverage in EML/CW laser/VCSEL/CPO/OCS, Lumentum is emerging as a “pure optical tool vendor” during the AI infrastructure boom. That’s because AI cluster expansion has moved from worrying about “is there enough compute chips” to “can tens or hundreds of thousands of GPUs move data fast and efficiently enough, at low power.”

Lumentum is not just selling standard optical modules; it's positioned at three levels simultaneously: on the scale-out side, offering high-speed EML chips and 800G/1.6T optical transceivers; on the scale-up side, advancing 1060nm VCSEL arrays to push optics further inside and between racks; and on the CPO/silicon photonics front, providing high-power CW lasers, DWDM external sources, and more. For even longer reach (scale-across/DCI), it covers pump lasers, narrow linewidth lasers, tunable lasers, and WSS devices. The 1.6T DR4 module Lumentum showcased at OFC this year already uses four 400G differential EMLs and sets the stage for 3.2T; ultra-high-power lasers are directly targeted at CPO and silicon photonics architectures. In short, **whether data centers opt for pluggable modules, CPO, or ultimately drive optics into the scale-up fabric, Lumentum stands to benefit at the core.

According to recent market news, Nvidia is revising its next-generation AI rack computing system, potentially reducing per-rack HBM and linking rack clusters optically in the upcoming Rubin Ultra, mainly due to surging memory chip prices. At the same time, leading data center optical component companies such as Lumentum—focusing on AI data center optical interconnect systems—stand to benefit from a “triple multiplier” of port count + per-port speed + higher-speed optical penetration. This, fundamentally, is why the optical interconnect theme has recently outperformed memory chips.

As AI models shift from training to massive inference, to MoE and Agentic AI, collective communication, KV/cache migration, and model-parallel communication between computing nodes all multiply east-west traffic. That's as 800G upgrades to 1.6T and towards 3.2T, pushing high-speed SerDes to the limits (insertion loss, power, signal integrity) and making optics more economical for scale-up connections once dominated by copper. Nvidia has already integrated silicon photonics directly into Spectrum-X/Quantum-X switches, touting 1.6Tb/s class optical switching for higher bandwidth density, lower power, and greater reliability—just as Lumentum accelerates VCSEL scale-up solutions co-packaged with host ASICs.

The data center storage chain has already seen extreme price hikes and profit booms, and the market is now starting to price in a future slowdown in ASPs. In contrast, AI optical interconnect is graduating from 800G to 1.6T and moving from scale-out to scale-up/CPO, so upward revisions to earnings forecasts are accelerating. Put simply: as GPU counts, model parallelism, and data center distances expand, bandwidth, lasers, optical modules, and CPOs are becoming the new marginal bottleneck—not that storage chips are unimportant, but the market is trading on “which supply bottleneck is tightening faster.” This explains the recent “short memory, long optics” trading trend.