Lumentum earnings up next as AI laser demand surges

Lumentum is expected to report fiscal Q3 EPS of $2.26 on revenue of $802.9 million, implying 35% sequential profit growth and 21% revenue growth as AI data center demand drives the business. Investors will focus on NVIDIA-related demand, CPO ramp timing, OCS backlog above $400 million, and whether gross margin can expand from 37% as 200Gbps laser chips gain mix. Shares have already surged more than 1,500% from the 52-week low, so the print could move the stock meaningfully given the 99x forward P/E and lofty expectations.

Analysis

The market is still underestimating how concentrated the upside is in a narrow set of AI plumbing suppliers. LITE is not just a beneficiary of capex; it has become a toll collector on the bottleneck where bandwidth density, power efficiency, and latency all meet, which means every incremental AI rack build can translate into disproportionate pricing power for the right component mix. That also creates a subtle winner/loser dynamic: hyperscalers may keep spending, but a larger share of wallet is likely to migrate away from general networking vendors and toward specialized optical and laser content, compressing margins for laggards that lack the 200G/next-gen roadmap.

The key near-term risk is not demand but expectation compression. At this valuation, even a clean beat can underwhelm if management frames the next 2-3 quarters as supply-constrained rather than demand-constrained, because the stock is already discounting near-flawless execution through fiscal 2026. The most important second-order tell will be whether backlog converts into shipment timing or stays stuck in “strategic commitment” language; if the latter, the multiple can de-rate quickly despite strong headline growth.

On the beneficiary side, NVDA’s investment is less about direct economics and more about securing a supply chain moat around optical interconnects, which could extend NVDA’s control over system architecture and make switching costs higher for cloud buyers. For GOOGL, OCS adoption is a longer-duration lever: if the backlog is real and deployment broadens, it lowers network reconfiguration costs and could support more efficient cluster utilization, which is a subtle but meaningful capex productivity win. The contrarian view is that consensus may be extrapolating the current ramp linearly; if AI networking spend pauses after the initial buildout, the market could be left with a premium multiple on a product cycle that normalizes faster than expected.

Key Decisions for Investors

• Stay long LITE only through the print with a tight post-earnings review window; if guide/timing commentary is merely in-line, trim 25-50% into strength because the stock is pricing near-perfect execution at ~99x forward earnings.
• If you want exposure to the theme, prefer a relative-value long LITE / short a lower-quality optical or networking peer over an outright long; the spread should work if CPO and 200G adoption continue, but reduces beta to a sentiment air pocket.
• Buy NVDA calls or maintain longs into the earnings cycle only as a supply-chain hedge, not as a direct valuation thesis; the embedded option is on control of AI architecture, with a 3-6 month payoff if optical content becomes a larger part of the rack bill of materials.
• For GOOGL, look for a medium-dated bullish structure (3-6 months) only on evidence that OCS backlog is converting to deployments; otherwise avoid chasing, since the catalyst is operational efficiency, not a near-term revenue re-rate.
• Use any post-earnings gap-up in LITE to initiate a partial hedge via short-dated puts or a collar; upside remains if margins expand, but the asymmetric risk is a multiple reset if management signals a slower ramp or customer digestion.