FN's AI Optical Growth Accelerates: Can It Challenge LITE & AAOI?

Fabrinet's fiscal Q4 2026 data-center revenue rose 68% year over year and 13% sequentially to $669 million, representing 51% of total revenue, as AI-driven DCI and HPC demand accelerated. DCI's annualized revenue run rate exceeded $1 billion, and management sees potential for another year of accelerating growth in fiscal 2027, supported by hyperscaler transceiver program ramps. New manufacturing capacity could raise total revenue capacity to $8.5-$9.3 billion initially and $12.5-$14 billion longer term, although Lumentum and Applied Optoelectronics are scaling competing high-speed optical offerings. Consensus EPS is $4.19, up $0.17 in 30 days and implying 43.49% growth, while FN trades at 21.14x forward earnings versus 20.74x for the sector.

Analysis

FN is the cleaner AI-optics volume exposure, but its manufacturing-services model means revenue growth will not translate one-for-one into operating leverage. The key earnings variable is mix: hyperscaler-direct transceiver programs can lift utilization but may carry lower margins and greater pricing pressure than specialized telecom or laser assemblies. The near-term setup is therefore favorable for revenue and estimate revisions, while the 6-18 month debate shifts to whether gross margin holds as new capacity comes onstream and customers gain multi-source options.

LITE is better positioned for scarcity rents if laser and optical-switching components remain the bottleneck, whereas FN is exposed to the risk that component vendors or hyperscalers internalize more manufacturing. AAOI is the highest-beta beneficiary but also the most vulnerable to qualification delays, yield issues, customer concentration, and an abrupt supply normalization once its capacity additions reach production. A broad 1.6T ramp is not automatically additive for all three: it can expand the addressable market while compressing transceiver assembly economics.

Consensus appears to treat customer visibility as demand certainty; it is more accurately a planning signal in a capex cycle subject to rapid architecture changes. Optical circuit switching and co-packaged/near-packaged optics are potentially disruptive to conventional pluggable-transceiver unit growth, though they create a longer-duration opportunity for suppliers with specialized photonics capabilities. The thesis is falsified if FN's next two reports show utilization-led revenue growth without gross-margin expansion, or if hyperscaler capex guidance weakens and lead times normalize.

The stock's relative underperformance creates an entry window only if estimate momentum remains intact; absent a margin or backlog-quality catalyst, modest valuation dispersion versus the technology complex may not independently drive a rerating. Monitor quarterly inventory, customer prepayment/deposit trends, revenue concentration, and management commentary on firm purchase orders versus forecast coverage.

Key Decisions for Investors

• Initiate a 3-6 month long FN position on post-earnings weakness rather than chasing pre-results momentum; target a rerating driven by upward FY2027 revenue/EBIT guidance, with a stop if gross margin declines by more than 100 bps year-over-year or management signals material forecast cancellations.
• Use a long FN / short AAOI pair over 6-12 months for lower-volatility AI-optics exposure: FN offers diversified manufacturing scale and customer breadth, while AAOI embeds greater execution risk from aggressive capacity and product ramps. Reassess if AAOI demonstrates two consecutive quarters of stable gross-margin expansion at high-volume 1.6T production.
• Maintain a tactical long LITE versus FN only if evidence emerges that laser lead times or OCS adoption are constraining system deployment; LITE has greater upside to component scarcity, but the trade should be reduced if component supply improves or cloud systems growth decelerates.
• Set an earnings watch item for FN: upgrade conviction only if management converts forecast visibility into binding orders, backlog, or customer-funded capacity commitments. Without that evidence, treat FY2027 demand assumptions as cyclical capex exposure rather than contracted growth.