Micron Is the Third-Best Performing S&P 500 Stock in 2026. Here Is Why High-Bandwidth Memory Changed Everything.
Micron's stock has risen 325% year-to-date, trailing only SanDisk and Western Digital among S&P 500 names. The reason is three letters: HBM. High-bandwidth memory is the product that every AI chip requires, and Micron finally has it at scale.
The Number That Defines Micron in 2026
Micron Technology's stock closed June 22 at $1,211.38, up 324.61% year-to-date, making it the third-best-performing stock in the entire S&P 500 in 2026, trailing only SanDisk and Western Digital. To understand how a company that most people associate with commodity DRAM memory chips became one of the best-performing stocks in the world, you have to understand one product: high-bandwidth memory, known as HBM. HBM revenue crossed $1 billion in a single quarter for the first time in Micron's history earlier in 2026. That milestone, and the trajectory beyond it, has fundamentally repriced what kind of company Micron is.
Micron reports fiscal Q3 2026 earnings on June 24 after market close. Wall Street consensus expects revenue of approximately $35 billion and adjusted EPS of $20.57. Those numbers would have seemed disconnected from reality for a memory company as recently as 2023, when Micron was reporting losses during the memory cycle downturn. The transformation reflects both a cyclical recovery in memory pricing and a structural shift: AI compute architecture now requires HBM, and Micron is one of only three companies in the world capable of manufacturing it at meaningful scale.
On June 22, just days before the earnings report, Anthropic announced a multi-year strategic agreement with Micron spanning memory and storage AI architecture design, a supply agreement covering Micron's data center portfolio, and a strategic investment in Anthropic's Series H funding round. "Our compute strategy depends on getting every layer of the stack right, and memory and storage are central to how efficiently we can train and serve Claude," said Anthropic co-founder Tom Brown. That statement, from a frontier AI lab publicly linking its infrastructure strategy to Micron's product roadmap, is not a marketing press release. It is a supply chain signal about how central memory architecture has become to AI system design.
What High-Bandwidth Memory Is and Why It Matters
Standard DRAM memory, the type that has been in computers for decades, is fast enough for most computing tasks. AI workloads are not most computing tasks. Training a large language model or running inference at scale requires moving enormous amounts of data between processors and memory at speeds that standard DRAM cannot provide. High-bandwidth memory solves this by stacking multiple DRAM chips vertically using a technology called through-silicon vias (TSVs), and then connecting that stack directly to the processor die using thousands of microbumps. The resulting memory package sits physically on the same silicon package as the GPU or AI accelerator, providing bandwidth that is an order of magnitude higher than standard off-chip DRAM.
Every NVIDIA H100, H200, and GB200 AI GPU ships with HBM stacked on the package. NVIDIA's forthcoming Blackwell Ultra architecture requires even more HBM per chip. AMD's MI300X AI accelerator uses HBM. Google's TPU v5 uses HBM. Amazon's Trainium2 uses HBM. There is no commercially viable path to training frontier AI models or running high-performance inference at scale without HBM, and there are only three companies in the world that manufacture it: Samsung, SK Hynix, and Micron. That three-company oligopoly in a structurally demanded product is the foundation of the HBM investment thesis.
Micron's Position in the HBM Market
For most of 2024, Micron was the smallest and least advanced of the three HBM suppliers. Samsung and SK Hynix had longer histories in the product and more established relationships with NVIDIA and other AI chipmakers. Micron's HBM3E, its current-generation product, received NVIDIA qualification during 2024, and since then the company has been ramping production aggressively. The key metric to watch in the June 24 earnings report is HBM revenue growth and any guidance on 2026 and 2027 supply commitments. If Micron can demonstrate that its HBM share is expanding and that pricing is holding despite increasing supply, the structural bull thesis remains intact. If HBM revenue growth disappoints or management signals pricing pressure from Samsung's aggressive capacity expansion, the stock will move sharply lower on the news.
The broader memory market context is also relevant. Standard DRAM pricing has recovered from the 2022-2023 cycle trough, and NAND flash pricing has improved. Micron is benefiting from both a secular shift (AI infrastructure driving HBM demand) and a cyclical recovery (general memory pricing normalizing). Separating those two drivers is important for understanding the durability of the earnings improvement. The HBM growth is structural. The DRAM and NAND recovery is cyclical and will eventually revert. Micron's net income projection ranking second only to NVIDIA across the entire PHLX Semiconductor Index for 2026 and 2027, reflects the market's view that the HBM structural story is real enough to sustain above-average profitability even after the cyclical tailwind fades.