Western Digital (WDC.US), one of the three leading U.S. storage product manufacturers alongside SanDisk and Seagate Technology PLC, reported quarterly results and future outlook after the U.S. market close on Thursday. The data revealed that the storage giant's quarterly performance and guidance exceeded Wall Street analysts' expectations, driven by what appears to be an "insatiable" and explosive global demand for nearline high-capacity HDDs from tech enterprises. Following the exceptionally strong results from memory chip manufacturers Samsung and SK Hynix, and fellow HDD storage leader Seagate Technology PLC (STX.US), Western Digital's across-the-board beat significantly reinforces the "storage super cycle" growth narrative being led powerfully by these storage titans. The global construction of hyperscale AI data centers, including projects like "Stargate," is advancing rapidly. Consequently, propelled by the torrent of AI computing demand fueling the continued explosive expansion for enterprise-grade high-performance HDDs and enterprise data center SSDs (eSSDs), flagship enterprise storage systems from Seagate and Western Digital are experiencing an unprecedented phase of simultaneous volume and price increases. This explains why Western Digital's stock surged 285% throughout 2025 and has already gained 60% in 2026. Following the earnings report from its rival Seagate, which saw its stock jump nearly 20% in a single day by Wednesday's close, Western Digital's results further ignited investor enthusiasm.

For its fiscal second quarter ended January 2, 2026, Western Digital's total revenue surged 25% year-over-year to $3.02 billion, surpassing the average analyst expectation of approximately $2.95 billion. This quarterly revenue figure significantly exceeded the constantly upward-revised consensus estimates that have followed the soaring stock prices of Western Digital, Seagate, and other storage leaders since Q4 2025. The company's non-GAAP adjusted gross margin for the quarter was 46.1%, higher than the average analyst forecast of 44.5%. Free cash flow was approximately $653 million, also beating the average estimate of $637 million.

Regarding other profitability metrics, on a non-GAAP basis, this leader in enterprise nearline high-capacity HDDs reported adjusted earnings per share (EPS) of $2.13 for the quarter, exceeding the average analyst estimate of about $1.93. Non-GAAP adjusted net profit reached approximately $807 million, a substantial 92% increase year-over-year, while adjusted operating profit soared to $1.019 billion, up about 72%. Under GAAP accounting standards, Western Digital's second-quarter operating profit was approximately $908 million, a 62% increase, and net profit skyrocketed to approximately $1.802 billion, a staggering 296% surge compared to the prior-year period.

Focusing on the critical forward guidance, Western Digital's management expects diluted EPS for the current quarter (fiscal Q3 2026) to be in the range of $2.15 to $2.45, far surpassing the Wall Street consensus estimate of approximately $1.99. The company anticipates total revenue for the quarter to be between $3.1 billion and $3.3 billion, again exceeding the recently elevated average analyst expectation of around $3.0 billion. Capital expenditures for the quarter are projected to be approximately $380 million to $390 million, reflecting the HDD duopoly's disciplined approach to not rushing into significant capacity expansion during the storage super cycle, instead opting for a measured, steady increase in HDD production.

"Our business continues to grow significantly. We expect revenue to continue its strong growth trajectory and profitability to improve, driven by sustained AI data center demand and accelerating penetration in the enterprise high-capacity hard drive segment," stated Western Digital's Chief Financial Officer, Chris Senerthal, in an earnings release. The world's largest tech giants, 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 seemingly endless surge in storage demand resulting from this explosive expansion in AI compute power 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 massive AI training projects.

During the nearly year-long bull market for storage stocks, besides the significant rallies seen at memory chip manufacturers SK Hynix, Samsung, and Micron, the three major storage product giants—Seagate, SanDisk, and Western Digital—also posted robust gains in 2025, each rising over 200%. SanDisk, the enterprise SSD storage system leader spun off from Western Digital, saw a staggering increase of nearly 600%. These leaders across storage chips and product lines have significantly outperformed the broader U.S. stock market and global equities. The core logic behind the powerful stock performance of Western Digital, Seagate, and SanDisk lies in the fervent construction of AI data centers, which not only fuels soaring demand for HBM memory but also causes a shift in memory chip capacity away from consumer electronics towards the more complex manufacturing and packaging of HBM. Simultaneously, all three tiers of the AI data center storage stack—the hot tier (NVMe SSD), warm/nearline tier (HDD), and cold tier (object and backup storage)—are experiencing exponential expansion. This, combined with long-term supply discipline from the HDD oligopoly, a recovering NAND cycle, and multi-year volume commitments from cloud providers, has led to a simultaneous leap in volume, pricing, and order visibility for these three companies.

Western Digital – The "Purest Nearline HDD Play" Amid the AI Wave. Since the spin-off of its flash business into SanDisk in February 2025, Western Digital's financial reporting no longer consolidates NAND flash or eSSD operations. The company's historically significant "enterprise data center SSD (eSSD)" super product line has been entirely transferred to the NAND storage giant SanDisk, which began trading publicly in February 2025. As a result, Western Digital as a listed entity has become arguably the "purest nearline HDD stock play" within the unprecedented AI wave. The underlying logic for Western Digital's benefit from the AI super-cycle closely mirrors that of Seagate: when enterprises and cloud providers massively expand or build new data centers to meet the "insatiable" storage demands of AI training and inference, the initial storage theme driven by "volume" is often enterprise DRAM—centered on HBM systems—followed by data沉淀, replay, archiving, and reuse. This latter need is met by nearline HDDs, which offer the "lowest TCO per TB" to handle exponentially growing data. In its latest earnings report, Western Digital explicitly attributed its growth and profit improvement to the scaled delivery of high-capacity HDDs and data center demand under the "AI-driven data economy," providing stronger guidance fueled by data center needs and high-capacity drive penetration.

From a product line perspective, Western Digital's primary beneficiary in this cycle is almost entirely the generational upgrade to high-capacity HDDs and the surge in high-density storage demand within the data center/cloud infrastructure sector. This aligns with the logic of Seagate's benefit from the AI infrastructure cycle via nearline HDDs. Hyperscale/cloud data center nearline HDDs represent Western Digital's core product line benefiting from the current "storage super cycle." Examples include the 32TB Ultrastar DC HC690 (UltraSMR) and 26TB-class CMR/UltraSMR HDDs, used for object storage, warm/cold tiers, deep content storage, and massive data lake expansion. Western Digital's official materials directly position these drives as low-TCO, high-performance storage expansion tools for hyperscalers and cloud service providers (CSPs). In AI training/inference systems experiencing exponential growth, AI training data, retrieval-augmented generation (RAG) corpora, observability data (logs/traces/metrics), compliance retention, and backups ultimately reside in the capacity tier with the lowest TCO. Therefore, when prices for data center-grade SSDs, dominated by SK Hynix and Micron, are pushed significantly higher due to tight supply and demand, large data centers are incentivized to more actively adopt a tiered/hybrid architecture 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 to use relatively fewer SSDs for caching/hot tiers and rely on HDDs to handle capacity/cold-warm tiers, which in turn enhances the resilience of HDD demand. For workloads like AI Agent/RAG/vector search/online inference, which involve "small-block 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, and archiving," object storage/HDDs hold a significant unit cost advantage.

The HDD Duopoly's Disciplined Approach to Capacity. Whether it's Google's massive TPU AI compute clusters or vast clusters of NVIDIA AI GPUs, all require fully integrated HBM memory systems alongside the AI chips. Furthermore, the current accelerated construction and expansion of AI data centers by tech giants necessitates large-scale purchases of server-grade DDR5 memory and enterprise-grade high-performance SSDs/HDDs. Storage leaders like Samsung Electronics, SK Hynix, Micron Technology, plus Western Digital, Seagate, and Kioxia are precisely positioned across key segments of these three core storage areas: HBM, server DRAM (including DDR5/LPDDR5X), and high-end data center SSDs/HDDs. These storage giants are the most direct beneficiaries within the "AI memory + storage stack," effectively feasting on the "super红利" (super dividends) of the AI infrastructure wave.

The primary reason Western Digital and Seagate are benefiting so greatly from the unprecedented global torrent of AI compute power and its seemingly "endless" storage demand lies in the explosive expansion needs of hyperscale AI data centers, like "Stargate," for Seagate's data center-grade nearline HDDs and its high-performance eSSDs. AI training and inference stretch the chain of "data generation -> cleansing -> versioning -> replay -> archiving" to an exponential degree. Data centers require storage solutions that balance cost and scale at the exabyte level. At this scale, nearline HDDs remain one of the optimal solutions for "cost per TB / capacity per watt." Therefore, as cloud providers and enterprises expand their AI infrastructure, they allocate substantial budgets to high-capacity nearline hard drives to meet the demands of data lakes, object storage, cold/warm tiering, and long-term retention.

As for why these two HDD super-oligopolies are not rushing to accelerate HDD capacity expansion, the core logic lies in the combination of "supply discipline," as evidenced in Western Digital's financials, and "industry engineering realities." They are actively seeking to avoid repeating the old cycle of "capacity expansion -> oversupply -> price wars -> profit collapse" that has plagued the HDD industry. Western Digital's management was quite blunt during the earnings call: they are not currently focused on adding unit capacity. Instead, they are using product structure and technology iteration to match demand—essentially delivering more exabytes through density and structural upgrades while maintaining price and profitability through supply discipline, rather than plunging back into a cyclical downturn via crude capacity expansion. Consequently, the two HDD giants prefer to use higher areal density roadmaps (e.g., Seagate's HAMR, Western Digital's ePMR/Ultra SMR/OptiNAND) to convert the "same unit capacity" into more exabytes, while keeping capital expenditures within a more controllable range.

Nomura's latest assessment reaffirms that this "storage chip super cycle" will last at least until 2027, with the brokerage explicitly stating in its research report that "meaningful supply additions are not expected until 2028 at the earliest." The report repeatedly emphasizes that expanding capacity for storage chips and HDD/SSD components is not something that can be done instantly, involving complex timelines for greenfield/brownfield projects and customized semiconductor equipment upgrade plans. Using SK Hynix as an example, Nomura points to specific constraints such as the long lead times for cleanroom construction/wafer capacity, significant delays in yield and cycle times due to advanced process node upgrades, and limitations on building/upgrading overseas chip factories. These factors lead to the conclusion that supply increases will be slower and the supply-demand gap will persist for a longer duration.