CNBC Daily Open: Nvidia may still be tech’s king, but Micron just stole a scene

Micron posted an 84.9% gross margin, up from 74.9% last quarter and 39% a year ago, while Qualcomm raised its fiscal 2029 non-handset revenue outlook to $40 billion from $22 billion, sending both stocks up 15%. SK Hynix also filed for a potential $29.4 billion Nasdaq ADR listing, the second-largest U.S. listing on record after SpaceX. Offsetting the chip rally, U.S. crude fell below $70 as tensions with Iran eased, while the White House requested $87.6 billion in supplemental funding tied to the Iran conflict and farm aid.

Analysis

The key read-through is that AI demand is no longer a single-name story; it is bifurcating into compute, memory, and edge inference, and the market is starting to price the second-order beneficiaries differently. QCOM’s step-up in long-dated non-handset revenue matters because it signals a path where data-center exposure becomes a material earnings bridge before the market fully rerates it as an AI infrastructure supplier; if adoption is real, the multiple expansion can outpace the earnings ramp because the revenue base is still comparatively small. The implication for NVDA is not immediate share loss, but a gradual broadening of AI capex that lowers the “one-vendor” premium and shifts investor attention toward components with lower power draw and system-level integration.

Micron’s profitability inflection is the more important signal for the supply chain: when memory margins get this high, it typically triggers capacity discipline in the near term and capex response in the medium term. That creates a setup where the current earnings upgrade cycle can persist for several quarters, but the back half of the cycle becomes vulnerable if HBM pricing starts to normalize faster than the market expects. For hyperscalers, the hidden risk is that memory inflation can quietly pressure GPU deployment economics even if headline AI budgets remain intact.

SK Hynix’s planned listing is strategically relevant because it creates a liquid U.S. proxy for an enterprise that is more levered to HBM demand than many U.S.-listed peers. If the deal sizes near the upper end, expect a scarcity premium initially, but also eventual de-rating risk once the float expands and investors benchmark it against the same cycle dynamics as Micron rather than treating it as a pure AI scarcity asset. The cleaner expression is not chasing the listing itself, but using it as a signal that Asia-based memory capacity is moving into the center of the U.S. AI allocation trade.

The geopolitical overlay is more tactical than structural. Lower crude is a near-term tax cut for transport, industrial, and consumer discretionary, but the bigger point is that easing supply risk reduces inflation volatility and gives the Fed more room to tolerate capex-heavy AI investment. The contrarian risk is that any renewed friction in the Strait of Hormuz would snap crude higher quickly, so energy hedges should be treated as event protection rather than a high-conviction directional long here.

Key Decisions for Investors

• Overweight QCOM vs NVDA for a 3-6 month horizon: long QCOM / short NVDA as a relative-value pair, betting that Qualcomm's AI data-center optionality rerates from a low base while Nvidia is already priced for near-perfect execution; stop if QCOM execution slips or NVDA accelerates incremental share gains.
• Buy MU on pullbacks for a 6-12 month cycle trade, but hedge with short-dated calls or a partial short in memory-adjacent semis if HBM pricing begins to roll over; the reward is continued margin expansion, the risk is that capacity adds in 2H compress the multiple before earnings catch up.
• Use the SK Hynix listing as a sentiment catalyst rather than a core long: consider a small tactical long on any U.S. IPO/ADR pricing weakness only if valuation comes at a discount to MU on forward gross margin, and take profits into first-week strength as scarcity premium normalizes.
• Add a tactical short-duration downside hedge on crude-sensitive sectors via XLE put spreads or long USO puts for 1-3 months; the thesis is that easing Gulf risk and sub-$70 oil supports multiple expansion in transport and discretionary, but any renewed shipping disruption would invalidate the trade quickly.
• If looking for the cleanest AI second-order beneficiary, favor networking/power-efficient inference beneficiaries over frontier AI leaders on a 12-month view; QCOM is the cleaner expression, with asymmetric upside if data-center design wins translate into recurring revenue rather than one-off wins.