Tech Giants Signal Memory Bottleneck: MSFT, META, GOOGL, AMZN, AAPL CEOs Cite Shortage

Within a 48-hour window last month, the CEOs of Microsoft, Meta, Google, Amazon, and Apple all made nearly identical remarks on earnings calls: memory is constrained and this bottleneck is not ending soon. This chorus of guidanceCompany-issued forecasts of future financial performance. is significant because it validates a structural supply shortage that has been building for quarters and points to sustained, peak-level capex spending ahead.

Micron has been the primary beneficiary of DRAM and NAND supply constraints, yet the stock trades at just 7x forward earnings, a steep discount to historical multiples and a fraction of the valuation commanded by AI-infrastructure pure-plays. This disconnect suggests the market is either pricing in a sharp deceleration in capex or is simply overlooking memory as a bottleneck asset class.

The memory shortage thesis cuts across multiple end-markets: hyperscaler data centers (where NVIDIA and other GPU vendors are inventory-constrained by memory availability), AI training workloads (where bandwidth becomes a bottleneck), and consumer devices (where memory costs inflate margins for phone and PC makers). If memory supply genuinely cannot keep pace with AI buildout, then Micron's earnings power should remain elevated well into 2026 and beyond.

The counter-argument is that spot memory prices are falling, suggesting supply is healing. Additionally, there is risk that the market misinterprets a temporary supply crunch as structural demand and pulls back capex expectations if GPU shortages ease or if AI model training efficiency improves faster than expected.