Qualcomm: Licensing floor funds a risky bet on becoming the second source for AI silicon

Stevie AI on QUALCOMM Incorporated (QCOM-USA | qualcomminco)

10/26/2026

Summary

Qualcomm is a semiconductor and IP licensing company transitioning from a smartphone-dependent royalty and chip business into a diversified supplier across automotive, IoT and now data center custom silicon. The structural insight is that QTL licensing — running at roughly 69% EBT margins — provides an earnings floor that is largely insulated from the Apple modem exit and the handset downcycle, giving management the confidence to fund an aggressive, multi-year diversification push without jeopardizing the balance sheet or the buyback program. The market appears to be pricing QCOM as a melting-ice-cube handset supplier losing its largest customer, when the more accurate framing is a licensing-funded option on a $40B non-handset business by FY2029. FY2025 actuals show the transition pain clearly: revenue grew to $44.3B from $39.0B in FY2024, but net income nearly halved to $5.6B from $10.3B and EPS fell to $5.01 from $8.98, reflecting Apple-related headwinds, margin compression, and mix shift costs hitting before the offsetting growth engines fully ramped. FY2026 guidance implies a sharp recovery — EPS of $9.75 — as automotive grows ~60% YoY and cost actions take hold, before FY2027 dips slightly to $9.68 as the Apple exit fully completes and Android handset revenue declines an estimated 20% YoY, carrying an explicit >$1.50 EPS headwind. Our price targets are built off a 14x forward P/E applied to FY2029 EPS of $12.61, reflecting a blended multiple between a legacy handset/semiconductor supplier (10-12x) and a company with a credible, high-margin licensing floor and emerging data-center optionality (18-20x) — discounted back at 10% annually through the trough years. This yields a FY2029 target of $176.65, with FY2028 at $160.59, FY2027 at $146.00, and FY2026 at $132.73 on a pure earnings basis — though we would flag that this methodology undershoots the current $177.23 price in the near term, implying the market has already priced in much of the recovery and is effectively underwriting data center optionality for free. Given the asymmetry — a funded floor business plus a free call option on a $15B data center TAM target — we rate QCOM a BUY, while acknowledging near-term EPS-based targets lag spot price and the thesis requires patience through FY2027's trough.

Thesis

1. **The Licensing Floor Is the Real Asset, Not the Chips** QTL generated extraordinary margins — 69% EBT in Q3 FY2026 — on a revenue base that is largely decoupled from Qualcomm's own chip shipment volumes, since royalties are assessed on device ASP across the broader handset industry regardless of whether Qualcomm supplies the silicon. This means that even as QCT handset revenue erodes from the Apple modem exit, QTL continues generating high-margin, high cash-conversion royalties. This floor is what allows Qualcomm to fund aggressive R&D investment in automotive, IoT, and data center custom silicon, plus a buyback program large enough to hold EPS flat-to-up through FY2026-27 despite choppy net income. Few companies in semiconductors have a royalty stream this durable and this insulated from their own execution risk. 2. **Automotive Is Already a Proven, De-Risked Growth Vector** Unlike the unproven data center bet, Qualcomm's automotive silicon business is already compounding at ~60% YoY for multiple consecutive quarters, backed by multi-generation platform lock-in agreements with BMW and Stellantis. Management's target of automotive plus IoT exceeding $24B by FY2029, within a broader $40B non-handset target, is underwritten by design wins already secured, not aspirational pipeline. Qualcomm is tracking toward becoming the #1 automotive semiconductor supplier by revenue — a durable, multi-year revenue stream with structurally better visibility than consumer handsets ever offered. 3. **Data Center Is the Free Option the Market Is Mispricing** Qualcomm's data center custom silicon ramp, starting December 2026, targets >$15B of revenue by FY2029 — nearly double the prior $22B non-handset guide now raised to $40B. The company is explicit that this will come at structurally lower margins (a 1.5-2pt drag on QCT gross margin), and we treat this as execution risk rather than embedded upside. But because the licensing floor funds this investment without diluting shareholders or straining the balance sheet (net debt falls from $7.0B in FY2026 to a net cash position of $1.7B by FY2029), shareholders are getting this option largely for free. If HBC Gen 1 silicon validates in customer evaluations through early 2027, this becomes a re-rating catalyst; if it fails, the core licensing and automotive business still supports the current valuation. 4. **EPS Resilience Through Buybacks Masks Volatile Net Income, But Isn't Fake** Net income is genuinely choppy — $10.3B in FY2024, collapsing to $5.6B in FY2025, recovering to $10.5B in FY2026, dipping to $10.0B in FY2027, before accelerating to $11.6B by FY2029. Yet EPS tells a smoother, upward story ($8.98 → $5.01 → $9.75 → $9.68 → $10.99 → $12.61) because of continuous, large-scale buybacks funded by FCF that grows from $12.9B to $15.3B over the forecast period. This is not share-count financial engineering layered on a shrinking business — FCF generation is real and growing, net debt is falling to a net cash position, and the buyback is simply redistributing genuine owner earnings to fewer shares. 5. **The Market Is Anchored to the Apple Headline, Not the Replacement Math** The Apple modem exit is the single most discussed overhang on the stock, yet management's own guidance shows non-handset QCT revenue growing >60% YoY in FY2027 specifically to fully replace Apple's departed revenue within that single year. The market's bearish framing treats the Apple exit as a structural hole; the data suggests it is a fully-funded bridge to a more diversified, albeit lower-margin, revenue base. The stock's current $177.23 price, against an EPS trough of $9.68 in FY2027, implies the market has not fully priced in the magnitude of this replacement, creating the entry point for patient capital. 6. **Catalysts Are Concentrated and Near-Term** The November 4, 2026 Q4 FY2026 earnings release will be the first quarter with quantified data center custom silicon revenue contribution and will confirm whether Apple revenue decline and Android growth are tracking to plan. HBC Gen 1 performance demonstrations in late 2026/early 2027 are the binary near-term catalyst for the data center optionality. Both catalysts arrive within a 6-month window, compressing the timeline for thesis validation or invalidation.

Risks

1. **Data Center Execution Risk Is Real and Unproven** Qualcomm is a new entrant in a capital-intensive, deeply entrenched market dominated by Nvidia, with hyperscaler custom silicon teams, Intel, and AMD all competing for the same design wins. HBC and CPU products remain pre-revenue and fully dependent on silicon performance validation that has not yet occurred. If the first two custom ASIC wins fail to scale or additional customers do not adopt the platform, the entire $15B FY2029 data center target — and the associated multiple expansion embedded in our valuation — evaporates, leaving Qualcomm as a slower-growth licensing and automotive story trading at a lower multiple. 2. **Apple Revenue Decline Could Be Steeper or Longer Than Modeled** The assumption that non-handset growth fully replaces Apple revenue within FY2027 is management's target, not a certainty. Any slippage in automotive or IoT ramp timing, or an accelerated Apple internal modem rollout beyond current assumptions, could push the margin and EPS trough further out or deepen it below the modeled $9.68 FY2027 EPS. 3. **Gross Margin Recovery Is Guided, Not Guaranteed** Management expects QCT gross margins to return to the historical 48-50% band as price increases take effect over multiple quarters, but this assumes successful pricing power in a competitive handset and IoT market. Input cost inflation (particularly memory) and the structural 1.5-2pt data center mix drag could keep margins below the historical band for longer than guided, compressing net income even if revenue targets are met. 4. **Buyback-Dependent EPS Growth Is Sensitive to Capital Allocation Discipline** The EPS growth path assumes continued aggressive buybacks at a fixed ~$176.50 price assumption. Any disruption to FCF generation — from data center capex overruns, M&A (Modular and Alphawave integration costs), or a slower-than-expected licensing renewal cycle — could force a pullback in buyback pace, directly slowing the EPS trajectory that underpins our price targets. 5. **Android Handset Decline Carries Explicit, Quantified EPS Risk** Management has already flagged QCT Android revenue declining ~20% YoY in FY2027 with a >$1.50 EPS impact — a headwind large enough to erase nearly all of the FY2026-to-FY2027 EPS growth otherwise expected from automotive and cost actions. Any further deterioration in Android flagship share or pricing could extend this drag beyond FY2027. 6. **Macro and Licensing Renewal Risk** Weak consumer spending, inflation, and soft capex cycles could further elongate handset replacement cycles beyond current assumptions, pressuring both QCT volumes and the ASP base against which QTL royalties are calculated. Any erosion in licensing terms upon renewal with major handset OEMs would directly impair the high-margin floor that underwrites the entire diversification strategy.

📈 Price Targets