Micron, Sandisk, Seagate, and Western Digital Diverge on Technology Even as AI Boom Lifts All Four

Four memory and storage companies — Micron, Sandisk, Seagate, and Western Digital — have become among the most closely watched equities in the S&P 500 this year, each riding the surge in spending on artificial intelligence data centers. All four rank among the index's ten best performers year to date, though their underlying businesses, and their recent trajectories, differ considerably.

Sandisk leads the group with a year-to-date gain of roughly 575%. Micron has risen approximately 240%, Seagate about 210%, and Western Digital more than 170%. Measured from the start of 2025, the returns are even more pronounced: Micron is up roughly 920%, Western Digital up 820%, and Seagate up nearly 740%, against a 25% gain for the S&P 500. Sandisk, which began trading independently in February 2025 following its separation from Western Digital, has surged more than 4,300% since then.

Despite those headline figures, all four stocks have pulled back sharply from their late-June peaks, falling roughly 20% to 40%.

The companies occupy distinct segments of the memory and storage ecosystem, distinguished primarily by the trade-off between speed and cost across three core technologies: DRAM, NAND flash memory, and hard disk drives.

DRAM sits at the fastest and most expensive end of that spectrum and forms the core of Micron's business. NAND — slower but cheaper, and non-volatile, meaning it retains data without power — is the sole focus of Sandisk, which also benefits from a longstanding joint manufacturing partnership with Japan's Kioxia. Seagate and Western Digital concentrate on hard disk drives, the slowest but least expensive option for storing large volumes of data.

D.A. Davidson analyst Gil Luria told CNBC that these technologies are not interchangeable. Modern data centers "need all of them," he said. "They do different things."

Micron CEO Sanjay Mehrotra made a similar observation, saying AI is generating demand across the entire memory hierarchy. "AI is driving a whole hierarchy of memory requirements, from high-bandwidth memory to DRAM to SSDs, and this is the exciting era for memory and critical need that all AI systems have," Mehrotra said on "Squawk on the Street," speaking from the site of the company's planned $10 billion research facility in Boise, Idaho.

HBM — high-bandwidth memory, which is constructed by stacking DRAM chips — has drawn particular attention given its role in feeding data to the GPUs and CPUs that power AI model training and inference. Luria argued that HBM has shifted memory's role in computing. "It's no longer just storage," he said. "It's actually how the model works. It's the guts of the model now. It's not just a periphery. It's not an accessory."

That shift has implications for how memory is sold. Rather than customers soliciting bids from whichever supplier offers the lowest price — the dynamic that historically drove brutal boom-and-bust cycles — HBM increasingly must be co-designed with the processors it supports. That deepens supplier-customer relationships and improves visibility into future demand.

Mehrotra said supply constraints remain acute. "All our customers across our end markets will buy everything that we make," he said, adding that data-center customers currently want roughly 50% more supply than Micron can commit. He appeared on "Mad Money" to add that the value equation around memory has "totally changed": "Today there is no AI without memory. AI systems need more memory. They need higher performance memory. They need lower power memory."

To reduce cyclicality further, Micron has pursued long-term strategic customer agreements, announcing 16 such deals alongside its June earnings report and signing additional ones since, Mehrotra said.

The rally across all four stocks comes against a backdrop of broader market caution. JPMorgan technical strategist Jason Hunter published a note Friday warning that AI stocks are showing similarities to the 1999–2000 communications equipment cycle and flagging potential downside heading into post-Labor Day seasonality — a signal that the sector's next leg may face more friction than the one that preceded it.

Disclaimer