NEW YORK, Feb 21, 2026, 11:02 EST — Market closed.
- Micron finished the day up, as traders zeroed in on AI-fueled appetite for memory chips.
- CEO Sanjay Mehrotra signaled a stronger push into India, with the company relying more on advanced packaging and design efforts tied to DRAM and NAND.
- Nvidia’s earnings, set for midweek, might offer the next cue for memory stocks.
Micron Technology, Inc. ended Friday up 2.6% at $428.17, wrapping up the week with a lift as chip stocks found their footing. The iShares Semiconductor ETF also notched a roughly 1% gain for the day.
This shift is significant: memory stands out as one of the most constrained elements in AI hardware right now. Running AI servers takes DRAM—the fast, volatile “working memory” inside computers—as well as NAND, which acts as flash storage for huge datasets, plus there’s high-bandwidth memory (HBM), specially designed for cutting-edge AI chips.
Micron’s CFO Mark Murphy is pointing to a surge in data center spending—he sees it approaching $800 billion in 2026, versus about $200 billion forecast for 2024. That jump, he says, thrusts supply and pricing discipline right into the spotlight again for memory chip producers.
Speaking at the India AI Summit on Friday, CEO Sanjay Mehrotra praised Micron’s “phenomenal” local team and said the company is ramping up its manufacturing presence in India. Mehrotra confirmed Micron is putting more than $2.7 billion into a new advanced packaging, assembly and test plant in Sanand, Gujarat, with additional research centers in Bengaluru and Hyderabad being expanded. The Times of India
Cash is pouring in on the public side as well. “When Meta says that it’s going to spend $100 billion, it’s going into these companies,” said Adam Patti, chief executive of VistaShares. He’s talking about the draw of AI “infrastructure” names—outside the headline tech giants. Reuters
Macro factors played a role heading into the weekend. On Friday, U.S. data put the personal consumption expenditures price index—favored by the Federal Reserve for tracking inflation—at 2.9% year-on-year in December. Core PCE, excluding food and energy, landed at 3.0%.
Supply’s still the sticking point. This week, The Verge pointed to Samsung Electronics, SK Hynix, and Micron as the heavyweights behind most of the world’s RAM output. With AI demand pushing up against manufacturing limits, prices for certain memory components have jumped.
Even so, memory’s a cycle—you know how it goes. Demand can flip. Should capacity get ahead of schedule, or those cloud clients ease up after their initial spree, pricing could slip, squeezing margins before you know it.
On Monday, attention shifts to MU and whether it can hang onto its recent advance. The rest of the semiconductor group’s moves will likely track with any talk about DRAM, NAND, and HBM contract prices.
Looking ahead to Wednesday, Nvidia’s fourth-quarter report lands on Feb. 25 at 2:00 p.m. PT. Investors will be watching closely, as they typically look to this update to get a read on AI server appetite — and with it, the amount of memory bundled with each new GPU rollout.
