SK Hynix Just Expanded Coherent's AI Upside

The key catalyst is SK hynix’s co-packaged optics roadmap (published in Nature Electronics) that extends optical links beyond switch networking to direct processor-to-pooled-memory connectivity—expanding Coherent’s potential addressable market. Coherent is already seeing strong execution: Data Center & Communications revenue rose 59% YoY and 19% sequentially to $1.62B in the latest quarter, with total quarterly revenue up 34% YoY to $2.05B and non-GAAP gross margin up 215 bps to 40.2%. Management guided fiscal Q1 2027 revenue to $2.2B–$2.4B and non-GAAP EPS to $1.85–$2.05, while Coherent estimates CPO TAM could exceed $15B by 2030. Overall, the roadmap adds long-term optionality while current results support near-term growth, though the processor-to-memory optics plan is still a roadmap with adoption-timing risk.

Analysis

The important shift is not another optical headline; it is that photonics is migrating from the network edge into the compute/memory core. That favors suppliers with breadth and manufacturing depth more than pure-play component exposure, because the economic value shifts toward alignment, thermal management, reliability, and packaging integration rather than simple bit-rate scaling. In that framework, COHR looks better positioned than vendors tied to a single bottleneck, while TSM gets a quieter but real second-order benefit from advanced packaging complexity.

Near term, this is mostly a sentiment and multiple story, not an immediate earnings reset. The market may extrapolate the roadmap faster than revenue can follow, so the first real catalyst is proof that current 800G/1.6T ramps keep translating into sequential margin improvement and utilization gains; absent that, the stock can give back news-driven gains once the attention cycle fades. If CPO remains confined to switch-to-switch links, the TAM expands less than bulls assume and the trade becomes more about product mix than category creation.

Contrarianly, the consensus may be too eager to treat "AI optics" as a single winner-take-most theme. The more likely path is a fragmented stack where laser sources, silicon photonics, packaging, and fiber attach each capture different economics; that makes COHR a better all-weather exposure than a more concentrated peer, but it also means investors should not pay as if the roadmap is already commercialized. The thesis is falsified if hyperscaler commentary stops short of deeper optical integration, or if COHR’s next two quarters show revenue growth without meaningful gross-margin leverage.

Key Decisions for Investors

• Go long COHR on any 5-10% pullback over the next 2-6 weeks; the setup is a 3-6 month re-rating as the market prices deeper optical penetration, but only if management confirms sequential margin expansion and capacity utilization gains.
• Pair trade: long COHR / short LITE for a 3-6 month horizon. COHR has broader exposure across lasers, silicon photonics, and optical subassemblies, while LITE is more concentrated in the high-power laser bottleneck; cover if laser scarcity proves more durable than expected.
• If already long COHR, consider financing upside with covered calls or a call spread into the next earnings print rather than chasing spot after the headline. The near-term catalyst is narrative-driven, so implied volatility may overstate the durable fundamental delta.
• Treat TSM as a secondary beneficiary watch item rather than a primary trade. If the next wave of CPO commentary starts emphasizing advanced packaging and photonic interposers, TSM should capture incremental wallet share; if not, the market impact should stay limited.