Research
行业4月30日 · Morgan Stanley

AI Phones, Upgrade Cycles & the New Hardware Math: A Selective, Premium-Led Cycle in 2026

AAPL005930.KSQCOM
中文EN

AI Phones, Upgrade Cycles & the New Hardware Math: 2026 as a Selective, Premium-Led Cycle

Core Conclusion

Smartphone upgrade intent hit an all-time high in the most recent AlphaWise survey, yet industry shipment volumes are falling. This divergence is driven by rising memory costs compressing margins and limiting volume at the low end, plus uneven AI monetization that concentrates incremental demand in premium price bands. The market overestimates a broad recovery in 2026. The cycle will be selective—only premium-tier OEMs and chipset suppliers capture upside.

Evidence Chain

Claim 1: Record upgrade intent coexists with declining total shipments. The AlphaWise survey (March 22, 2026) shows consumer upgrade intention at historic peaks. However, the same analysis projects falling industry volumes in 2026. This gap implies a structural shift: higher willingness to upgrade but lower ability to do so at the low-to-mid price points. Investment implication: aggregate demand signals are misleading; the volume-weighted outlook is negative for broad-based handset plays.

Claim 2: Rising memory costs are the dominant margin and volume constraint. The report names “rising memory costs” as the critical headwind for 2026. Samsung Electronics (005930.KS), the memory leader, is valued at ~2x forward P/B, which is the stock's commodity cycle peak multiple. That valuation level implies the market prices in a sustained memory up-cycle, yet the rising cost is a direct squeeze on smartphone OEM margins, especially for non-premium models. Investment implication: memory cost pass-through has an upper limit; mid/low-end units will either lose margin or be deferred, shrinking the addressable TAM for broad market suppliers.

Claim 3: AI monetization is uneven, benefiting only the premium segment. The report underscores “uneven AI monetization” as a defining characteristic of the 2026 cycle. Apple’s $315 target (8.1x FY27 EV/Sales) and Qualcomm’s 14x forward Non-GAAP P/E are both premised on premium-only AI feature uptake. Apple’s multiple relies on iPhone 17 outperform and Apple Intelligence adoption; Qualcomm’s below-historical-average P/E already prices a muted cycle. Investment implication: any sustained uplift from AI is confined to the high end. Broad AI-driven replacement cycles are not materialising.

Key Risks

  • Memory costs rise faster than modeled, compressing margins for all but the top-priced models and forcing further volume cuts at the low end.
  • Consumer willingness to pay for AI features proves insufficient, stalling the premium upgrade cycle and reversing the selective recovery thesis.
  • Geopolitical tensions disrupt supply chains (Samsung, Apple) or accelerate decoupling, impairing both component availability and end-market demand.
  • Apple accelerates internal silicon timelines, reducing Qualcomm’s modem content and undermining its current valuation floor.

Valuation & Trading Implications

2026 demands a selective approach. Long Apple and Qualcomm: both stocks embed premium-cycle expectations and limited volume risk. Apple’s 8.1x EV/Sales and Qualcomm’s 14x Non-GAAP P/E leave room for upside if AI monetisation proves sticky. Avoid pure memory-cycle exposure: Samsung’s 2x P/B offers no buffer if memory costs peak and roll over. The trade is concentrated on the high-end supply chain, not the broad market.

(Character count: 4,987)

Related (同 ticker)