Record chip rally adds $2 trillion in combined value to Micron, Intel and AMD in second quarter

AI-chip “enablers” surged in Q2 as investors rotated out of Nvidia’s hyperscalers into semis: Micron more than tripled (+240%, +$920B market cap), Intel jumped 216% (+$480B), and AMD nearly tripled (+$615B). Combined gains across Micron, Intel, and AMD totaled about $2T, while Nvidia rose only 15% as a proxy for slower share performance. The VanEck Semiconductor ETF (SMH) climbed 71%, and Barclays attributed the move to semis-led euphoria as AI capex expands across data centers, lifting memory, CPUs, networking (Marvell +200%), and Arm (+134%).

Analysis

This is less a “new AI winners” story than a capital-allocation rotation: investors are moving up the stack from platform owners to the picks-and-shovels names that monetize each incremental dollar of AI capex. That is bullish for semis in the near term, but it also means the market is paying cyclical peak multiples for businesses whose earnings are still highly sensitive to ordering cadence and supply additions. In other words, the first-order trade is momentum; the second-order risk is that the industry starts discounting a demand plateau before fundamentals actually roll over.

The cleanest beneficiary is MU because memory pricing has the steepest operating leverage and the fastest path from supply tightness to margin expansion. But that same leverage cuts both ways: memory is the easiest part of the chain to overbuild, so this is the segment most exposed to a 6-12 month air pocket once customers normalize inventories. AMD and ARM look more durable than INTC because they have cleaner secular share gain vectors, while INTC’s rerating is more dependent on execution and policy optionality than on demonstrated earnings power.

Contrarian view: the consensus is underestimating how quickly this trade can rotate back into NVDA if hyperscaler capex stays elevated or if packaging/power constraints keep training hardware scarce. The recent move may be too much of a short-term relative-value adjustment and not enough of a durable earnings revision story. The main falsifier for the bullish enabler thesis is any evidence that memory pricing, server demand, or AI capex growth inflects down over the next two earnings seasons.

Key Decisions for Investors

• Long AMD / short INTC for 3-6 months: express secular share-gain versus a rerating that is still partly turnaround/policy optionality; risk/reward favors AMD unless Intel shows real margin inflection or major foundry wins.
• Long MU on pullbacks, not strength: best upside leverage to AI memory tightness over the next 1-2 quarters, but size modestly because this is the most cyclical part of the basket and can de-rate fast if supply expands.
• Long ARM / short META for 1-3 months: ARM benefits from broadening inference and device-side AI while META remains a capex-heavy platform with weaker near-term operating leverage; stop the pair if META re-accelerates ad growth or capex guidance falls.
• Avoid chasing SMH after the quarter’s run; if entering, use a staged buy on a 5-8% drawdown and pair with a hedge via QQQ or XLK to isolate semiconductor beta from broader tech multiple risk.
• Set an alert on next hyperscaler capex commentary: if AMZN/GOOGL/MSFT/MSFT-like spend growth slows, fade the enabler trade quickly; if it accelerates, revisit NVDA relative strength as the likely catch-up winner.