KeyBanc downgrades Cirrus Logic after Asia trip, sees memory and analog strength

KeyBanc downgraded Cirrus Logic to Sector Weight from Overweight, citing limited growth prospects over the next 6–12 months as roughly 80% of revenue is tied to Apple, which cut TSMC wafer starts equivalent to 15–20 million iPhones. Analyst John Vinh expects Q4 DRAM prices to rise 15%–20% quarter over quarter and NAND prices to increase 20%; he also forecasts high-bandwidth memory prices up 80%–100% in Q1 2027. Persistent AI and data-center demand supports memory and analog suppliers, though memory makers’ capacity shifts could reduce PC and smartphone bit supply by 20%–30% next year.

Analysis

The clearest asymmetry is memory pricing versus downstream unit economics: constrained bit supply can lift supplier realizations, but higher component costs may force PC and handset makers to raise prices, absorb margin pressure, or ship fewer units. That makes the pricing outlook for MU more investable than the reported long-dated customer discussions, which do not by themselves establish committed volumes or take-or-pay protection. Watch OEM guidance and inventory/contract disclosures for evidence that pricing converts into earnings rather than demand destruction.

Over 1–3 months, the analog read-through is positive but less clean. Longer lead times can support pricing and factory loading at ADI, MCHP, TXN and NXPI; they can also reflect customer precautionary ordering, leaving these names vulnerable if orders normalize. CRUS has a more specific near-term risk: concentration in AAPL plus weaker handset production makes estimate downside more plausible, while memory inflation could further defer PC diversification. Do not infer that TSMC wafer-start changes map one-for-one to CRUS shipments or TSMC revenue.

Over 6–18 months, memory capacity allocation toward data-center products could structurally constrain client-device supply, but substitution, customer qualification and new capacity can unwind scarcity. Contrarian risk: the most bullish memory price forecasts, especially far-dated HBM, may be difficult to monetize if AI investment or customer budgets cool. The article’s headline references Fed policy, but its body supplies no basis for a rates trade.

Key Decisions for Investors

• Relative-value watch: consider long MU versus short CRUS, preferably on a pullback or after checking current valuation and positioning. The thesis is memory pricing leverage against CRUS’s concentrated handset exposure, not a claim that the businesses are directly hedged. Reassess if MU’s contract pricing or gross-profit outlook fails to improve, or CRUS guides to resilient growth despite softer AAPL demand.
• Avoid chasing analogs solely on lead-time commentary. Track order growth, backlog conversion, inventory and guidance over the next quarter for ADI, MCHP, TXN and NXPI; favor names whose reported orders and margins confirm demand rather than extended lead times alone.
• Monitor AAPL and DELL/HPQ device guidance for pass-through versus demand destruction. If component inflation prompts material shipment or margin cuts, reduce exposure to device OEMs and reassess the memory long; if pricing is passed through without unit deterioration, the supplier thesis strengthens.
• Treat announced or discussed long-duration supply relationships as a catalyst to verify, not booked revenue. Look for contract duration, volume commitments, pricing terms and customer concentration in disclosures; broadening supply or easing spot/contract prices would falsify the scarcity thesis.