Bet on These Photonics ETFs to Exploit the Next Wave of AI Revolution

Nvidia committed at least $6.5B to photonics companies over the prior three months, including $2B each to Lumentum, Coherent and Marvell, reinforcing optical connectivity as a key AI-infrastructure investment theme. Goldman Sachs forecasts optical-networking TAM will expand more than ninefold, from roughly $15B in 2026 to $154B by 2028, while co-packaged optics penetration could reach 20%-30% by 2028. The article highlights diversified ETF exposure through LYTE, LUMA and LAZR following sharp YTD gains in Marvell (+183.4%), Lumentum (+142.5%), Corning (+68.8%) and Coherent (+60.3%).

Analysis

The investable bottleneck is shifting from general AI capex to the qualified optical components that determine cluster throughput and power efficiency. LITE and COHR have the most direct earnings torque to transceiver, laser and component content, while GLW is a lower-beta beneficiary through fiber and connectivity build-outs; MRVL has a more mixed profile because optical DSP upside must offset its already elevated AI-networking expectations. Standardization lowers customer lock-in at the physical layer, which is strategically positive for component volume but can ultimately cap pricing power for system vendors.

The near-term risk/reward is unfavorable for a broad photonics beta chase after sharp relative performance: 2027-28 adoption is likely to be reflected in valuation well before material revenue, leaving the group exposed to any pause in hyperscaler capex, qualification delay, or evidence that improved copper architectures extend the transition window. The key 1-3 month catalyst is not TAM commentary but order-book conversion, inventory normalization and gross-margin guidance from LITE, COHR and MRVL. A meaningful cut to cloud capex plans from MSFT, AMZN or META would likely compress optical multiples disproportionately because consensus embeds a multi-year ramp.

The contrarian opportunity is GLW: its exposure is less dependent on the technically demanding co-packaged-optics ramp and more tied to physical data-center build-out, where long lead times and customer supply agreements can support visibility. Conversely, the smallest thematic ETFs are poor institutional vehicles: thin liquidity, wide spreads and concentrated overlap can turn a thematic reversal into a liquidity event rather than diversified exposure. Prefer liquid single names and a paired structure over LUMA or LAZR until verified assets, holdings and creation/redemption liquidity improve.

Key Decisions for Investors

• Accumulate GLW on 5-8% pullbacks over the next 1-3 months; target a 15-20% upside over 12 months from AI/data-center connectivity demand with a stop/review if management cuts optical-connectivity growth or AMZN deployment commentary weakens.
• Run a 6-12 month pair: long GLW / short MRVL in roughly beta-neutral dollars. This expresses physical-connectivity volume growth while hedging the more fully valued AI-networking narrative; close if MRVL demonstrates sustained optical revenue acceleration and margin expansion rather than design-win announcements.
• Do not add outright LITE or COHR after momentum-driven moves without confirmation in next earnings of backlog conversion, book-to-bill above 1x and stable gross-margin guidance. If those metrics validate, use 9-12 month call spreads rather than stock to limit downside from a capex-reset de-rating.
• Avoid LUMA and LAZR for scalable execution; their small asset bases and limited trading liquidity create material tracking and exit risk. LYTE is more usable but remains a concentrated proxy for LITE/COHR rather than diversified infrastructure exposure.
• Set a downside trigger across the optical basket if MSFT, AMZN or META guide data-center capex below consensus, or if NVDA platform commentary delays optical interconnect adoption beyond the next architecture cycle; reduce exposure immediately because the valuation thesis is duration-sensitive.