Big Tech CEOs Say Memory Chips Constrained; $MU Trading at 7x Earnings

The convergence of messaging from the five largest technology companies represents a rare moment of alignment on a structural market constraint. On consecutive earnings calls, leadership teams at Microsoft, Meta, Google, Amazon and Apple independently highlighted that memory availability remains the binding constraint on their artificial intelligence infrastructure buildouts. None suggested the problem was abating. This confirmation from so many decision-makers at once underscores that the memory shortage is not a cyclical blip but a multi-year structural issue embedded in the AI capex story.

Micron Technologies carries the valuation arbitrage. At 7x forward earnings, memory chip manufacturers appear to be pricing in either cyclical normalisation or maturity, even as the addressable market for AI-trained memory expands. The disconnect between narrative (confirmed memory scarcity across five mega-cap tech buyers) and valuation (cheap relative to earnings and growth) creates a natural flashpoint for money to rotate into semiconductor exposure. Broadcom, AMD, and other AI-adjacent semiconductor plays would benefit from recognition of this demand tailwind.

The implication extends across the AI infrastructure stack. If memory remains the constraint, then companies selling memory or derivatives of memory, from chipmakers to cloud providers to networking firms, face a multi-year revenue ramp that current multiples may not fully reflect. Conversely, AI model-training companies themselves may face persistent margin pressure if they cannot secure adequate memory at acceptable prices, which could dampen near-term profitability even as top-line AI revenue surges.

Skeptics argue that increased production capacity (TSMC, Samsung, Intel foundry expansions) will eventually normalise supply within 18 to 24 months, flattening the arbitrage. But the scale of capex required to satisfy five mega-caps simultaneously suggests the memory shortage persists longer than consensus pricing assumes. This is the thesis driving bullish positioning in $MU and AVGO in particular.