SanDisk stock has risen 464.5% in 2026 alone. Dell Technologies popped 39% in a single session after its most recent earnings. Western Digital has surged more than 200% year-to-date. The S&P 500 is up 11%.
These are not the names anyone expected to lead the AI supercycle. No foundation models, no chatbots, no billion-parameter training runs. Just drives, racks, and the hardware that keeps data in one place long enough for AI to do something with it. The market, sometimes slow to update, has finally noticed.
The Infrastructure Buildout Creates Demand Nobody Modeled
The AI arms race is, at its core, a physical problem. Large language models do not run on conviction. They run on servers filled with memory chips, connected by copper and fiber to arrays of storage that hold weights, training data, and inference outputs at scale. Every dollar that Alphabet, Microsoft, or CoreWeave commits to AI infrastructure ultimately lands somewhere in the hardware supply chain. In Q1 2026, Alphabet announced an intent to raise $80 billion in fresh equity specifically to fund infrastructure buildout. That money does not stay on a balance sheet. It flows to procurement.
Dell booked $24 billion in new AI server orders last quarter. The company raised its full-year FY2027 revenue guidance to a range of $165 billion to $169 billion, up from a prior guide of $138 billion to $142 billion. Its AI-optimized server revenue expectation for fiscal 2027 was revised to roughly $60 billion, an increase of 144% over prior projections. The stock, which was trading around $85 eighteen months ago, hit $317 on its earnings gap. Its forward price-to-earnings multiple now sits at 32, with trailing P/E around 50.
Seagate said at its last earnings call that its capacity was fully allocated and sold out through calendar 2026. Western Digital's management said the same thing. When the two dominant providers of hard disk storage tell investors there is nothing left to sell, the implication is not subtle. AI data centers are absorbing every platter they can produce.
!Legacy Tech AI Surge, 5-Year Performance (DELL / WDC / SNDK / HPQ vs S&P 500)
SanDisk, spun off from Western Digital in early 2025, has emerged as the most dramatic story in the group. Currently trading around $1,619, the flash memory company now carries a market capitalization near $240 billion. Twenty-two analysts have issued a Buy rating; the average 12-month price target among them is $1,751. The analyst consensus at MarketBeat sits at $1,398, reflecting a divergence in views about how long the AI storage surge can sustain current valuations. The stock's P/E multiple of 54 reflects expectations of earnings growth well beyond the current cycle.
HP Inc. Is the Outlier That Explains the Story
HP Inc. closed at $25.38 on June 8. Its 52-week high is $29.65. The stock is down roughly 16% over the last five years. For an investor looking at the same AI headlines as everyone else, HP's flat trajectory illustrates what separates the hardware beneficiaries from the ones left behind.
The distinction is proximity to inference infrastructure. Dell, Western Digital, and SanDisk are directly in the path of AI capex. Hyperscalers are not building data centers stocked with consumer-grade printers and PC peripherals. HP's business, anchored in personal systems and printing, has not yet found a meaningful connection to AI workload demand. That may change. Major computer vendors including HP are projecting price increases of up to 30% in the second half of 2026 as AI-driven component shortages filter through the supply chain. Whether that translates to margin expansion or unit volume recovery remains to be seen. For now, the stock is a useful control case: exposure to AI buildout is not uniformly distributed across legacy hardware.
What Analysts Think, and Where the Risks Live
Wall Street, broadly, is bullish on the infrastructure cohort. The bull case on Dell rests on the AI server demand curve sustaining the $60 billion trajectory the company has itself guided. Bear cases, which range from $220 to $281 in analyst models, center on a specific and credible risk: margin compression. Gross margin at Dell fell from 21% to 18% as high-volume AI server configurations, dominated by Nvidia GPU racks, mix into the revenue base at lower per-unit profitability. Revenue is growing fast. The question is whether earnings follow at the same pace.
For SanDisk and Western Digital, the risk is cyclicality. Memory storage has historically been one of the most boom-and-bust segments in technology hardware. The current demand signal is real: hyperscale AI infrastructure buildout is consuming NAND flash and spinning disk at a rate that has cleared inventories and sold out production. But supply chains respond. If capital investment in new fab capacity outpaces demand growth in 2027, the same companies that led the 2026 rally could give back gains quickly. The cycle is not new. It is just more intense this time because the underlying demand driver, AI compute, is compounding faster than any previous technology wave.
The more durable argument is structural: every AI model that gets deployed, every inference query that gets answered, every training run that completes leaves behind data that has to live somewhere. Storage is not a discretionary input. It is an operational requirement. The companies that make it may be old. The need for what they make is not.
Follow @UTXOMacro for daily Bitcoin, macro, and AI breakdowns.