Micron Technology Surges on Memory Chip Supply Constraints and AI Demand

What happened

Several Wall Street analysts raised their price targets and reaffirmed bullish ratings on Micron Technology. DA Davidson maintained a $1,000 price target, citing increased conviction in a 'virtuous cycle' for memory suppliers to AI infrastructure. Deutsche Bank issued a matching $1,000 target, while Mizuho forecast DRAM contract pricing could rise approximately 355% year-over-year and NAND pricing up 510%. Market research firm Gartner projected a 125% full-year increase in DRAM prices, with Q1 contract prices expected to jump 58% to 63%. Analysts pointed to supply constraints as a key driver, with production limited to three major manufacturers—Samsung, SK Hynix, and Micron—while demand from AI applications continues to accelerate. The supply shortage is expected to persist at least through the end of 2027, as new facilities like Micron's Singapore plant will not begin volume production until late 2028 and may take an additional eight years to reach full capacity.

Why it matters

Memory chips have emerged as a critical bottleneck in AI infrastructure development. Large language models require substantial memory capacity in intermediate processing stages, and as AI models grow larger and incorporate longer context lengths, memory requirements increase exponentially. High-bandwidth memory (HBM) demand is particularly strong, with each gigabyte consuming three times the wafer capacity of traditional DRAM. Counterpoint Research estimates HBM demand from custom AI processors could increase 35-fold between 2024 and 2028, while SK Hynix projects the HBM market will grow at 30% annually through 2030. For Micron, favourable pricing dynamics translate directly to earnings growth. Analyst consensus estimates show fiscal 2026 earnings per share of $58.11—a significant jump from fiscal 2025's $8.29—with fiscal 2027 projections reaching $101.78 per share, representing 75% year-over-year growth. The company reported second-quarter fiscal 2026 adjusted earnings that increased nearly eightfold year-over-year.

Bigger picture

The AI infrastructure buildout shows no signs of slowing, with data centre capacity expected to double between 2025 and 2030. Barclays projects record low vacancy rates and steady rent growth, with the majority of new capacity pre-leased. The memory market has shifted from cyclical to structural demand patterns, reducing traditional volatility. However, not all analysts agree on the magnitude of the opportunity. Bernstein assigned Micron an outperform rating with a more conservative $510 price target, reflecting historical concerns about memory demand volatility. The competitive landscape remains concentrated, with DRAM production limited to three manufacturers who control their own capacity, contrasting with the increasingly crowded CPU market where Nvidia and Arm are competing for foundry capacity at Taiwan Semiconductor Manufacturing Co. Market observers increasingly reference a potential memory supercycle extending beyond 2027, though views on duration and magnitude vary.

What to watch

Near-term catalysts include quarterly DRAM and NAND contract pricing updates, which directly impact revenue and margin trajectories. Progress on Micron's Singapore facility construction and ramp timeline will signal supply expansion pace. Broader indicators include HBM adoption rates across custom AI processors, enterprise AI deployment velocity, and data centre capacity utilisation metrics. Competitive dynamics among Samsung, SK Hynix, and Micron regarding HBM production capacity and pricing discipline will influence industry profitability. Additionally, any signs of demand normalisation or inventory adjustments among hyperscale customers would affect the supply-demand balance and pricing environment.

This article was generated by Quantli AI using publicly available news sources.