Meet the AI Infrastructure Stock That Has Tripled in 2026 (Hint: It Can Still Double)

Seagate reported fiscal 2026 revenue of $12.2 billion, up 34% year over year, and EPS of $15.58, up 92%, as AI data-center storage demand tightened HDD supply and lifted pricing. For fiscal Q1 2027, it guided to $4.1 billion in revenue, implying 56% growth, and adjusted EPS of $7.30, implying 180% growth. With HDD capacity reportedly sold out through at least 2027 and most nearline exabytes allocated into 2028, analysts expect fiscal 2027 revenue to rise 54% to $18.78 billion and EPS to increase 130% to $35.78.

Analysis

STX has shifted from a cyclical HDD recovery into a capacity-allocation story, which can sustain gross-margin expansion longer than the market typically assigns to disk-drive upcycles. The key variable is not aggregate AI capex but nearline exabyte mix, contract duration, and price realization versus unit-cost inflation. With two meaningful global suppliers, incremental hyperscaler demand has unusually high operating leverage; WDC is the cleaner read-through beneficiary, while SSD vendors such as MU and Samsung face a relative disadvantage in bulk cold-storage workloads despite benefiting from broader AI spend.

The immediate risk/reward is less favorable after a 208% YTD move: the stock now requires repeated estimate upgrades rather than merely delivery against elevated expectations. Over the next 1-3 months, the critical catalyst is evidence that FY27 guidance embeds conservative pricing and that backlog converts without customer concentration or qualification delays. A miss in nearline exabyte shipments, gross margin, or commentary that 2028 allocations are non-binding would likely compress the scarcity premium quickly; watch WDC results and hyperscaler capex commentary as leading confirmations.

Consensus may be underestimating the duration of supply discipline, but it is also extrapolating peak-cycle margins too mechanically. HDD supply can respond through yield improvement, manufacturing reactivation, and technology transitions before greenfield capacity is needed, while a hyperscaler shift toward higher-QLC NAND penetration could cap the long-duration upside. The more durable thesis is storage intensity per AI workload, not a blanket AI-capex beta; NVDA remains a demand signal, not a direct economic beneficiary of HDD pricing.

Key Decisions for Investors

• Maintain a tactical STX long only on post-earnings confirmation of nearline pricing, FY27 gross-margin expansion, and unchanged 2028 allocation language; use a 3-6 month horizon and size modestly after the parabolic move. Exit/reduce on a material sequential decline in nearline ASPs or a guide implying margin below prior-quarter levels.
• Prefer a 6-12 month long STX / short WDC pair only if STX demonstrates superior contract pricing and margin conversion; otherwise WDC offers higher beta to the same HDD-tightness thesis and may be the better catch-up vehicle. Reassess if the relative spread widens without corresponding estimate revisions.
• Do not underwrite the published multi-year price target: it assumes both aggressive EPS growth and sustained premium valuation. Treat consensus FY27-FY28 EPS revisions as the tradeable catalyst; if upward revisions stall for two consecutive estimate cycles, take profits rather than waiting for a demand reversal.
• Monitor MU and NAND pricing as a substitution alert rather than short them outright. Sustained QLC cost-per-TB improvement combined with stable enterprise SSD availability would weaken HDD pricing power over a 6-18 month horizon and cap STX/WDC multiple expansion.