共有:
GPU(Less scarce)HBM(Bottleneck)Power(Bottleneck)AI InferenceCapital flows: GPU → HBM / Power
AI Infrastructure · Chip Economics

Wall Street Eyes Micron as the Next Nvidia

The "just buy Nvidia" era may be shifting. With HBM and power emerging as the real constraints, capital is moving upstream in the AI supply chain.

2026-06-29 · AI Navigate Editorial · 4 min read

What Shifted

For two years, AI investment meant Nvidia stock. But GPU supply has been less constrained than HBM memory and power, which were quietly becoming the real bottlenecks.

Wall St. is framing Micron as next Nvidia as AI bottlenecks shift from GPUs to HBM/CoWoS/power, redirecting capital to upstream supply-chain winners.

For PMs and business leads, tracking HBM and power constraints — not just GPU availability — sharpens infrastructure decisions. Direct impact for engineers is minimal.

Supply Chain Shift

Old bottleneckGPU supply (H100 etc.)
New bottleneckHBM memory + power infrastructure
Rising playerMicron (HBM production)
Related techCoWoS advanced packaging

How to Factor This In

When planning AI cloud or GPU cluster procurement, factor in HBM supply tightness and power costs as first-class constraints — focusing only on GPU specs can lead to operational bottlenecks.

Business leads evaluating long-term AI infrastructure investment should track HBM memory makers (Micron, SK Hynix) and power infrastructure trends alongside GPU news.

Source: AI Semiconductor / GPU Economics