Amazon Cites Higher Memory Costs for Raising Capital Expenditures to $220 Billion. Here's Why the Micron Stock Price Still Dropped.

Amazon raised full-year 2026 capex guidance to $220B from $200B, citing higher memory costs, but CEO Andy Jassy warned AI infrastructure capacity will still fall short of demand through 2026-27. Micron (MU), a key AI-memory supplier, still saw the stock drop 5.9% the day after Amazon’s earnings (from $874.66 on July 30 to $823.03 on July 31), likely offset by earlier gains and profit-taking after an 18.3% rally. The article also flags potential rate-hike pressure after a July 29 Fed decision with dissent toward a 25bp hike, adding caution despite bullish memory-supply narratives extending into 2028.

Analysis

The cleanest read is not "AI demand is strong" but "cost inflation is moving upstream." When hyperscalers are forced to raise capex because memory is scarce and expensive, the first-order beneficiary is the memory vendor, while the second-order loser is the buyer's free-cash-flow conversion and return-on-capital narrative. That matters more for AMZN than for the broad semiconductor group, because the market can tolerate capex when it is growth-funded, but it tends to punish it when margin assumptions have to be revised down.

MU and SSNLF still look like the relative winners, but the trade is increasingly about duration of pricing power rather than unit growth. The risk is that investors extrapolate a straight-line AI supercycle and miss the transition from spot scarcity to contract discipline: that is bullish for revenue visibility, but it usually caps upside to gross margin at some point well before demand rolls over. If memory tightness extends into 2027-2028, the structural reset is a higher-quality earnings stream for suppliers, not a permanently higher peak multiple.

The contrarian point is that the consensus may be underestimating rates sensitivity. A higher-capex, lower-FCF cloud cohort is exactly where multiple compression shows up first if yields back up again, even if the AI thesis remains intact. Near term, the key falsifiers are a rollback in AMZN capex, any sign that AWS demand is not matching spend, or evidence that DRAM/HBM pricing and lead times are flattening; over 1-3 months, those are more important than the headline capex number itself.

Key Decisions for Investors

• Prefer MU and SSNLF on pullbacks over chasing AMZN strength; use a 1-3 month horizon and require follow-through from memory pricing or customer-contract commentary before adding size.
• Initiate a relative-value pair: long MU/SSNLF basket vs short AMZN for 3-6 months, targeting margin-transfer dynamics rather than absolute AI demand; cut if AMZN guides capex lower or AWS margin trends improve without a corresponding drop in memory pricing.
• Do not chase MU after the recent vertical move; wait for consolidation and only add if the next quarter confirms backlog conversion and no sign of ASP erosion.
• If rates continue higher, consider a small tactical short in AMZN into strength as a valuation/FCF hedge against the broader AI-expenditure trade; stop out if the company shows evidence of capex normalizing or better-than-expected operating leverage.