Seagate Technology PLC (STX.US), one of the three leading US storage product manufacturers alongside SanDisk and Western Digital (WDC), reported quarterly results and future outlook after the US market close on Tuesday. The data revealed that driven by what seems like an "insatiable" explosive demand from global tech enterprises for high-performance nearline HDDs and enterprise data center SSDs, the storage giant's quarterly performance and guidance comprehensively exceeded Wall Street analysts' expectations. This robust, across-the-board beat from Seagate significantly reinforces the so-called "storage super cycle" growth narrative, championed by major players including the three primary storage chip manufacturers—Samsung, SK Hynix, and Micron Technology (MU)—as well as storage product giants like Western Digital and SanDisk.
Regarding the highly anticipated performance outlook, the guidance provided by Seagate's management for the third quarter of fiscal year 2026, for both revenue and profit, was substantially higher than Wall Street consensus estimates. This fully highlights the persistently strong demand for its enterprise-grade HDD and SSD high-performance data storage equipment, fueled by companies rapidly scaling up AI applications and the resulting exponential expansion in AI computing power requirements. Global tech behemoths, including Meta, Microsoft, and Google, are increasingly aggressively investing trillions of dollars into AI data centers and closely related AI infrastructure, aimed at training and efficiently operating large language models (LLMs) with massive parameters.
The unprecedented surge in storage demand, driven by this continuous explosion in AI computing needs, is benefiting traditional storage hardware suppliers like Western Digital, Seagate, and SanDisk on an unprecedented scale. Their enterprise nearline HDDs and enterprise NVMe SSDs are essential for storing the vast amounts of data required for the efficient operation of these large models and extensive AI training projects. Seagate's management expects third-quarter revenue to be around $2.9 billion, plus or minus $100 million; according to data compiled by LSEG, Wall Street analysts had on average anticipated approximately $2.77 billion, indicating that Seagate's revenue outlook is significantly above the constantly upward-revised consensus expectations following Seagate's stock surge since Q4 2025.
The company anticipates adjusted earnings per share (EPS) for the third quarter to be approximately $3.40, plus or minus 20 cents; this also substantially exceeds analysts' recently revised EPS expectations of adjusted EPS $2.96. Seagate's stock surged 225% throughout 2025 and has already gained 35% year-to-date, ranking among the strongest performers in the S&P 500 index. Following the earnings release, Seagate's stock rose over 10% in after-hours trading. For the second quarter ended early January, Seagate's total revenue was approximately $2.83 billion, a significant 22% increase year-over-year, surpassing analysts' average expectation of $2.73 billion.
On an adjusted basis, the company reported adjusted earnings per share of $3.11, comfortably beating the average analyst estimate of $2.81 per share. More significantly, Seagate's second-quarter gross margin and operating margin both hit record highs. The non-GAAP gross margin surged to a record 42.2%, a marked improvement from 35.5% a year earlier; the non-GAAP operating margin reached a record high of 31.9%, also a substantial increase from 23.1% in the prior-year period. During the nearly year-long bull run in storage stocks, besides the significant share price increases of the three primary storage chip makers—SK Hynix, Samsung, and Micron—the share prices of the three major storage product giants, Seagate, SanDisk, and Western Digital, all surged over 200% in 2025.
Among them, SanDisk, a leader in enterprise SSD storage systems, saw a staggering increase of nearly 600%. These giants in storage chips and product lines have significantly outperformed the broader US stock market and even global equity markets. The core logic behind the strong stock price performance of Seagate and the other two storage product giants lies in the fact that the booming construction of AI data centers not only drives a surge in HBM demand, causing a shift in storage chip capacity from consumer electronics to the more complex manufacturing and packaging of HBM, but also leads to the simultaneous exponential expansion of the three-layer storage stack in AI data centers.
The AI data center storage stack—comprising the hot tier (NVMe SSD), warm tier/nearline HDD, and cold tier (object & backup storage)—is all expanding concurrently. Furthermore, long-standing supply discipline among HDD oligopolists, a recovering NAND cycle, and multi-year volume commitments from cloud providers have collectively caused a simultaneous jump in volume, pricing, and order visibility for these three companies. The explosive expansion in HDD and SSD demand is driving Seagate's performance and outlook beyond expectations.
During the earnings call, Seagate's CEO, Dave Mosley, stated that the quarter saw significant sequential revenue growth in "almost every end market" and set new company records for exabyte shipments, gross margin, operating margin, and non-GAAP EPS. He described storage demand as "exceptionally strong," particularly in the data center end market, and noted sustained substantial growth in demand for high-capacity nearline hard drives from global large-scale AI data centers, alongside improving demand from enterprise edge computing.
In a news release, Seagate CEO Dave Mosley stated, "As AI applications further elevate the creation value and economic value of data, modern data centers increasingly require exabyte-scale storage solutions that combine high performance with high cost-effectiveness." "Our areal density-focused product roadmap positions us to meet evolving storage demands and exabyte-scale data growth while creating significant value for our customers and shareholders for years to come."
Regarding Seagate's 2026 capacity planning, Mosley indicated that Seagate's nearline HDD capacity is "sold out for the entirety of 2026," and the company expects to begin accepting orders for the first half of 2027 in the coming months. The primary reason Seagate is significantly benefiting from the nearly "insatiable" storage demand driven by the global, unprecedented wave of AI computing power is the explosive expansion demand from hyperscale AI data centers for Seagate's data center-grade nearline HDDs and high-performance SSDs.
Seagate's core product for AI data centers, nearline HDD—represented by the Mozaic 3+ / HAMR platform, emphasizing high capacity of up to 30TB per drive—directly addresses the capacity pain points of "data lakes/object storage/archiving" in AI data centers. Seagate's Nytro 5060 (PCIe Gen5 enterprise NVMe)—positioned as a high-performance data center SSD—boasts specifications like up to ~14.9GB/s bandwidth, up to ~3.3M IOPS, a 2.5-inch form factor capacity of up to 30.72TB, and dual-port configuration for high availability.
Labeled as 3D eTLC NAND, it well suits the endurance requirements for "hot data/high-speed writing." In AI training/inference systems experiencing exponential growth, AI corpora, retrieval-augmented generation (RAG) corpus, observational data (logs/traces/telemetry), compliance retention, and backups ultimately reside in the capacity tier with the lowest total cost of ownership (TCO). Therefore, when prices for data center-grade SSDs, dominated by SK Hynix and Micron, are significantly elevated due to tight supply and demand, large data centers tend to more actively adopt tiered/hybrid architectures combining SSDs and HDDs to control costs.
In other words, as prices for high-performance data center SSDs rise sharply due to NAND supply constraints, data centers prefer using relatively fewer SSDs for caching/the hot tier and relying on HDDs to handle capacity/cold & warm tiers, thereby enhancing the resilience of HDD demand. For workloads like AI Agent/RAG/vector search/online inference, characterized by "small random reads, high concurrency, and sensitivity to tail latency," the value of NVMe far exceeds that of HDD/object storage.
Conversely, for "massive raw data, log retention, archiving," object/HDD offers greater cost advantages per unit. Whether it's Google's massive TPU AI computing clusters or vast clusters of NVIDIA AI GPUs, they all rely on HBM storage systems fully integrated with AI chips. Furthermore, the current acceleration by tech giants in building or expanding AI data centers necessitates large-scale purchases of server-grade DDR5 memory and enterprise high-performance SSDs/HDDs. Samsung Electronics, SK Hynix, Micron Technology, plus the three storage product leaders including Seagate, are precisely positioned within or concentrated in certain sub-categories of these three core storage areas.
They are the most direct beneficiaries in the "AI memory + storage stack," essentially capturing the "super红利" from the AI infrastructure wave. On Wall Street, institutions like Morgan Stanley, Nomura, and Bank of America are proclaiming the full arrival of an AI-driven "storage chip super cycle," suggesting that the intensity and duration of this cycle could far surpass the "storage super bull market" driven by the cloud computing era in 2018. Notably, alongside substantial revenue and profit growth, Seagate also demonstrated strong cash generation capabilities.
The storage giant generated $723 million in operating cash flow and $607 million in free cash flow in the second quarter. The company's latest capital allocation strategy has also drawn market attention: on one hand, it repaid $500 million in convertible senior notes, optimizing its capital structure; on the other hand, it paid out $154 million in cash dividends and announced a quarterly dividend of $0.74 per share payable in April 2026. Some Wall Street analysts point out that this dual approach of "deleveraging + shareholder returns" underscores management's confidence in future cash flow growth underpinned by strong HDD and SSD storage demand.
However, it might also暗示 that the company perceives a lack of large-scale, expansionary capital expenditure opportunities with higher returns at the current stage. The "storage super cycle" is expected to extend until 2028! A golden age for HDDs and SSDs has arrived. Analysts from Japan's Nomura also judge that this "storage chip super cycle," beginning in the second half of 2025, will last at least until 2027, with meaningfully new supply additions unlikely to appear before early 2028. Nomura suggests that investors should continue to overweight storage leaders in 2026, focusing on the "price-profit-valuation" triple play as the main investment theme for storage, rather than viewing storage solely as an HBM story.
The institution expects the three major storage chip companies to achieve record-high profits. A forecast from US storage chip giant Micron Technology indicates that the company expects the total addressable market for HBM to grow at a compound annual growth rate of approximately 40%, reaching about $100 billion by 2028, up from around $35 billion in 2025. Overall NAND demand often follows HBM's continued expansion, suggesting that the current storage super cycle could last from 2024 through 2028. A recent report from J.P. Morgan suggests that the AI inference wave is transforming NAND flash from a "strong cyclical commodity" into a high-growth AI infrastructure asset.
This also implies that the three storage giants—SanDisk, Seagate, and Western Digital—are poised for stronger profit growth trajectories than currently anticipated. J.P. Morgan believes that as AI workloads shift from training to inference, and given supply bottlenecks for HDDs in nearline storage, the NAND market surrounding enterprise SSDs is experiencing unprecedented structural growth. Investors are severely underestimating the strategic importance of NAND in the era of AI inference. J.P. Morgan predicts that blended average selling prices in the NAND sector will rise significantly by 40% year-over-year in 2026.
The latest "Context Inference/Long-Context Inference" BlueField platform introduced by AI chip superpower NVIDIA can be seen as a major long-term growth catalyst for "hot/warm tier enterprise SSDs/NVMe" and the cold/warm data ocean. In a technical blog, NVIDIA explicitly positions ICMS as a new G3.5 tier: a pod-level "Ethernet-attached flash tier" specifically designed to store latency-sensitive, reusable inference context, and emphasizes that it provides "petabytes of shared capacity" per GPU pod, using higher bandwidth/higher energy efficiency to support KV cache reuse and prefetching, thereby reducing GPU stall and improving tokens/s.
The figure/architecture description also states plainly: ICMS nodes are "built on SSDs," with BlueField-4 handling the KV I/O plane and efficiently terminating protocols like NVMe-oF and object/RDMA. Translating the aforementioned "Context Inference/Long-Context Inference" BlueField platform into a storage "benefit logic" means: This platform introduces a new high-bandwidth flash tier between local SSDs and shared storage specifically for "context," consequently requiring more enterprise SSDs/NVMe. Another significant benefit is that "as Agentic/Long-Context Inference scales up, the volume of persistent data will overall become larger," and object storage with its underlying capacity media remains the lowest-cost "data ocean."