Why is Micron so cheap? AI rally leader has the third-lowest valuation in the S&P 500

The market is still pricing MU as a spot-memory proxy, but the contract mix changes the equity from a pure beta play into a hybrid of cyclical upside and quasi-annuity cash flows. That should compress earnings volatility and justify a higher multiple even if peak-cycle EPS is capped, because the downside case historically embedded in the stock is being structurally reduced. In other words, the market may be underestimating how much of the old bear case was really about earnings instability, not low absolute margins.

Second-order, the main losers from sustained HBM tightness are not just AI chip designers but also the most levered buyers of compute capacity: cloud and model builders with weaker pricing power. NVDA likely remains the best pass-through vehicle, but memory inflation can still pressure gross margin optics and make every incremental AI server node more capital-intensive, which slows adoption at the margin if hyperscalers start rationing deployment. On the supply side, any evidence that Korean rivals or China add meaningful capacity faster than expected is the cleanest way to break the rerating story.

Near term, the stock can stay choppy because flows matter more than fundamentals: semis are still being used as funding legs in factor trades, so MU can sell off even on positive industry prints. Over 1-3 months, the key catalyst is whether management can keep extending contracted coverage and preserve margin floors; over 6-18 months, the debate is whether MU earns a software-like multiple on partially contracted revenue. The thesis fails if HBM/DRAM pricing rolls over before contract coverage reaches scale or if gross margin guidance stops stepping up despite tight industry supply.