What happened
Micron commits $200 billion to expand memory chip production for AI data centres. This investment shifts Micron from commodity manufacturing to high-margin AI hardware. The capital funds new fabrication plants to meet surging demand for High Bandwidth Memory (HBM). Micron targets the memory bottleneck currently limiting GPU performance. This follows a December 2025 sales surge and aligns with record profit forecasts across the memory sector.
Why it matters
Data centre operators can scale compute clusters without memory-bound constraints because Micron increases HBM capacity. Hardware architects currently face a critical memory bottleneck that limits large language model performance. Therefore, procurement teams gain long-term supply stability, reducing the need for component hoarding seen at Lenovo in late 2025. This $200 billion commitment matches Amazon’s recent AI spend. Result: memory moves from a commodity risk to a predictable architectural component within the $660 billion AI infrastructure cycle.
