Memory Chip Shortage Persists as MSFT, META, GOOGL, AMZN, AAPL Cite Constraints

A striking convergence appeared across earnings calls in late April and early May: the CEOs of Microsoft, Meta, Google, Amazon, and Apple all publicly stated that memory, particularly for AI inference and training, remains severely constrained and shows no signs of relief soon. This consistency across five of the world's largest technology companies is rarely seen and underscores the severity of the shortage.

Despite this unanimous chorus, the memory-chip market cap remains depressed. Micron Technology ($MU) trades at roughly 7x trailing earnings, a valuation typically reserved for mature, slow-growth sectors. This gap between the reality described by major cloud and consumer-tech firms and the market's willingness to pay for memory suppliers suggests traders either underestimate the constraint's durationBond price sensitivity to interest rate changes., doubt profitability will follow from higher chip prices, or are waiting for evidence that supply is truly expanding.

The shortage extends beyond DRAM to packaging, substrate, and specialized memory for AI workloads. Samsung and SK Hynix face similar supply pressures but trade at slightly higher multiples in Seoul. The arbitrage opportunity implies either that US investors are more skeptical of the thesis or that they expect competitive normalization faster than the forward guidanceCompany-issued forecasts of future financial performance. suggests. Major chipmakers have announced capacity expansions, but these typically take 18-24 months to come online.

Industry sceptics point out that earnings-call commentary can reflect near-term demand spikes that don't necessarily sustain pricing power long-term. If foundries like TSMC and Samsung manage to free up capacity for memory production, or if AI model training shifts to more memory-efficient architectures, the shortage narrative would evaporate quickly.