Storage and semiconductor stocks rallied sharply in early 2026 despite widespread AI and tech sector declines. SanDisk stock jumped 315.3% while Western Digital shares climbed 166.1%, marking a clear separation between infrastructure providers and software companies.
Micron Technology reported record revenues and expanding margins in its fiscal Q1 2026 earnings. The memory chip maker's performance reflects sustained demand for data center infrastructure as AI model training and inference workloads intensify.
Two companies disclosed a combined $400 billion in capital expenditure commitments for AI infrastructure buildout. Nebius separately projected $16-20 billion in capex over the coming period, with 60% funded from operational cash flow rather than external financing.
The stock performance gap signals investor repositioning toward companies selling physical infrastructure—chips, storage, data centers—and away from AI application developers. Hardware providers generate immediate revenue from infrastructure deployment, while software companies face longer monetization timelines and market uncertainty.
Analysts characterize the broader tech selloff as indiscriminate, hitting profitable and speculative companies alike. The infrastructure rally suggests investors distinguish between companies building AI systems and those attempting to profit from AI-powered products.
Storage companies benefit from dual demand drivers: expanded data center capacity and higher storage density requirements for large language models. Training runs for frontier AI models now require petabytes of high-speed storage, creating sustained order backlogs.
The valuation split creates a bifurcated market where price-to-earnings ratios and revenue multiples compress for software firms while expanding for hardware makers. Infrastructure companies trade on tangible deployment metrics—chip shipments, storage capacity sold, data center square footage—rather than speculative user adoption curves.
Capital flows into infrastructure reflect confidence in near-term AI spending regardless of application-layer success. Tech giants and cloud providers committed to multi-year buildouts must acquire chips and storage systems before launching services, guaranteeing hardware revenue even if consumer AI products underperform.
The trend positions semiconductor and storage firms as lower-risk AI plays compared to software developers betting on unproven business models.