Qualcomm: AI Boom Ignored

Qualcomm is pivoting beyond handsets into AI data centers, targeting $15B of data-center revenue by FY2029 and $40B in total non-handset sales. Its AWS agreement, involving up to $60B in payments tied to 25M warrants, is expected to accelerate the data-center revenue ramp while limiting upfront cost. The deal materially strengthens Qualcomm's diversification and AI infrastructure growth outlook.

Analysis

The market should value this less as a near-term revenue event than as a potential multiple re-rating trigger: QCOM’s handset exposure has historically imposed a cyclical discount, while credible hyperscaler-qualified accelerator revenue could move the company toward a higher-quality infrastructure earnings mix. The key economic question is whether AWS commitments are incremental, profitable silicon volume or effectively subsidized through warrant dilution and pricing concessions; absent disclosed vesting thresholds, strike price and gross-margin terms, the headline contract value is not a reliable proxy for QCOM shareholder value. Near term, the announcement is more likely to support sentiment and estimate optionality than materially alter FY26 EPS.

Competitive read-through is mixed for NVDA and AMD. QCOM’s best initial opportunity is likely inference and power-constrained workloads rather than displacement of Nvidia’s entrenched training stack, meaning the more immediate pressure falls on AMD’s still-developing accelerator share narrative and ARM-based custom-silicon alternatives. Over 6-18 months, successful AWS deployment would validate a fourth merchant AI-silicon supplier and increase buyer leverage against NVDA, potentially constraining industry gross-margin assumptions even if QCOM’s absolute data-center revenue grows.

Contrarian risk is that hyperscaler warrants often align incentives but defer economic cost: if QCOM shares appreciate materially, dilution can absorb a meaningful portion of the contract’s value, while AWS retains flexibility to prioritize internally designed chips. The thesis is falsified if QCOM does not disclose production deployment, revenue-recognition timing, and data-center gross-margin progression by the next two earnings cycles; a large target without quarterly bookings or backlog disclosure should be treated as strategic aspiration, not a forecastable earnings stream.

Key Decisions for Investors

• Maintain or initiate a modest long QCOM position only on weakness rather than chase the announcement; use a 6-12 month horizon and size for execution uncertainty. Upside requires disclosed AWS deployment and upward revisions to FY27 non-handset revenue, while downside is a return to handset-driven valuation if data-center contribution remains immaterial.
• Express the competitive view as long QCOM / short AMD in equal dollar amounts over the next 3-6 months, contingent on QCOM providing concrete shipment or backlog metrics. The pair benefits if merchant inference demand broadens beyond AMD’s current opportunity set; exit if AMD reports materially stronger accelerator revenue traction or QCOM’s AWS economics indicate low-margin custom work.
• Do not underwrite the reported contract value into EPS estimates until QCOM files warrant terms, including vesting milestones, exercise price and maximum dilution. Set an event alert for the next earnings release: a quantified FY26 data-center revenue contribution and gross-margin commentary would justify increasing exposure; omission is a reason to fade post-announcement strength.
• For existing NVDA longs, retain core exposure but hedge incremental AI-semiconductor multiple risk through a small SOXX put spread or QCOM/AMD relative-value basket over 6-12 months. The risk is not an immediate Nvidia revenue loss, but evidence that hyperscalers can qualify credible alternatives and use them to negotiate lower accelerator economics.