Bernstein’s latest 2026 U.S. IT hardware outlook highlights how ai data storage will capture the largest share of value from the ongoing artificial intelligence build-out.
Bernstein forecasts an AI-driven data surge through 2026
In its new 2026 report, research firm Bernstein argues that the AI-driven data explosion will continue through 2026 with no signs of slowing. Analyst Mark Newman writes that model developers have little incentive to curb spending as AI capabilities keep improving and data requirements expand.
According to Newman, developers must feed ever-larger datasets into training runs to stay competitive. Moreover, the firm notes that AI training workloads, content creation and long-term enterprise retention are all combining to drive unprecedented demand for capacity.
“Memory and hard-disk drive suppliers stand to benefit most from these trends,” Newman said, calling the current phase an “unprecedented memory and storage super cycle.” That said, Bernstein stresses that the structural demand backdrop, rather than short-term cycles, is the core driver.
Memory and storage suppliers emerge as top beneficiaries
The report concludes that memory and storage suppliers are the clear winners as AI workloads expand. Bernstein says buyers are effectively price insensitive because they must secure storage capacity regardless of cost, pushing prices higher across the sector.
SanDisk receives Bernstein’s highest praise for 2026. The firm raised its SanDisk target price to $580, citing severe NAND shortages and price increases it describes as unprecedented in the recent history of storage markets.
Seagate also earns a top pick designation. Moreover, Bernstein characterizes Seagate as “a less volatile beneficiary of the data explosion” relative to other storage suppliers, positioning it as a steadier way to gain exposure to rising capacity needs.
The firm expects the memory storage super cycle to run through at least 2026. It argues that AI training, inference, richer media formats and lengthening corporate retention policies will keep utilization and pricing elevated.
Broadcom secures leadership in AI ASIC market
Bernstein also highlights Broadcom as a structural winner in the AI accelerator stack. Analyst Stacy Rasgon reaffirmed an Outperform rating and set a $475 price target on January 9, underscoring confidence in Broadcom’s AI roadmap.
The firm met with Broadcom leadership to address questions about rising AI competition and customer-owned tooling. However, Bernstein came away convinced these threats are exaggerated, arguing that rivals cannot quickly replicate Broadcom’s accumulated advantages.
Those advantages include 3D chip stacking, 400G serdes technology and in-house substrate manufacturing. In addition, Broadcom is developing techniques that could reduce or eliminate the need for interposers, which would further differentiate its AI ASIC platform.
Bernstein notes that Broadcom benefits from Nvidia‘s rapid innovation cycle. The company enables XPU customers to track Nvidia’s roadmap, making Broadcom the only supplier that can provide this level of compatibility in the broadcom ai asics segment.
IBM positioned for quantum computing upside
Beyond classical AI infrastructure, IBM stands out in Bernstein’s outlook for early quantum computing adoption. The firm estimates a potential 23% upside for IBM tied to initial commercial use cases expected to emerge around 2026.
Bernstein links that opportunity to enterprises exploring specialized workloads where quantum systems may deliver performance or optimization advantages. However, it still frames quantum as an early-stage, targeted driver rather than a broad replacement for traditional compute.
OEM hardware makers face margin and earnings pressure
While component and infrastructure suppliers benefit from rising demand, original equipment manufacturers are under strain. Bernstein warns that surging memory and storage costs are compressing margins for major PC and server vendors.
HP Inc. and SMCI could see earnings fall by as much as 19% as component inflation outpaces their ability to raise system prices. Moreover, Dell and HPE also face downward pressure on profitability as they absorb higher bill-of-materials costs.
Apple has partial protection in the near term thanks to long-term supplier agreements that lock in pricing. However, Bernstein expects Apple may eventually need to increase device prices later in 2026, a move that could slow unit growth even if revenue holds up.
These dynamics illustrate how the same data boom that benefits storage and component makers can create hp dell earnings pressure when passed through to hardware manufacturers with less pricing power.
AI workloads reshape enterprise data storage strategies
Bernstein emphasizes that the surge in AI training and inference is fundamentally reshaping enterprise data strategies. Companies are retaining more information for longer periods to feed future models and to support richer content workloads.
Consequently, the data storage requirements for ai extend beyond hyperscaler training clusters into broader corporate IT estates. That said, the firm argues that only a subset of vendors are positioned to monetize this shift consistently.
In Bernstein’s view, the primary_keyword ai data storage companies include leading memory producers, hard-disk drive manufacturers and advanced networking chip providers that directly enable massive-scale data pipelines.
Outlook: storage-led cycle to define AI infrastructure economics
Bernstein concludes that elevated memory and storage pricing will likely persist through 2026 as AI development accelerates. Moreover, it believes buyers will continue to prioritize capacity and performance over near-term cost optimization.
As a result, component suppliers such as SanDisk, Seagate, Broadcom and IBM appear best positioned to benefit from the AI data surge. Meanwhile, OEMs including HP Inc., SMCI, Dell and HPE must navigate rising input costs and potential unit demand risks.
Overall, Bernstein’s 2026 outlook paints a hardware landscape defined by an enduring data boom, where those controlling critical storage and compute technologies capture the lion’s share of AI-related value.
Source: https://en.cryptonomist.ch/2026/01/16/ai-data-storage-2026-hardware/
