After painful breakup, Qualcomm tries to replace Apple with AI

Qualcomm faces a revenue reset as Apple phases out Qualcomm chips, with Q2 revenue down 3% and next-quarter revenue expected to fall 7%, or about $700 million below analyst expectations. The company is trying to offset the iPhone-related decline by expanding into AI data centers, including new AI200/AI250 accelerator chips and a recently completed $2.4 billion Alphawave Semi acquisition, but execution and competition from Nvidia remain major concerns. Shares jumped 15% on data-center optimism, but the core business outlook is still under pressure.

Analysis

QCOM is facing a classic “duration mismatch” problem: the market is pricing a long-term option on AI/data center exposure, while the near-term cash engine is rolling over faster than the replacement business can scale. That creates a valuation trap—any rally on AI headlines is vulnerable unless management can show booked backlog, not just product roadmaps, because the Apple loss is a step-function revenue headwind that is arriving before inference wins can compound.

The more important second-order effect is competitive positioning in the broader edge-AI stack. If Qualcomm can translate its mobile modem / power-efficiency DNA into data-center inference, it could become a niche beneficiary in workloads where watts-per-query matters more than raw FLOPS; that would pressure AMD and some custom silicon efforts at the margin. But if memory inflation and software friction persist, the real winner is NVDA, because every delayed or failed entrant effectively extends CUDA’s moat and preserves pricing power longer than consensus expects.

AAPL’s transition is not just a supplier change; it is a margin and strategic control upgrade that likely improves hardware gross margin over time and weakens Qualcomm’s negotiating leverage across the handset ecosystem. The more subtle spillover is that other OEMs will be emboldened to internalize more of the stack, which can compress QCOM’s mix and force heavier R&D spend just to defend share. That increases the probability of a multi-quarter earnings reset, not a one-quarter air pocket.

The contrarian view is that the market may be underestimating how big a 1% share gain in AI infrastructure can be if it lands in a constrained supply environment. Still, the burden of proof is high: until Qualcomm proves it can ship at scale and not just announce partnerships, the setup remains a “show me” story with asymmetric downside if June investor day disappoints.

Key Decisions for Investors

• Short QCOM into June investor day; use call spreads or outright equity short if borrow is manageable. Risk/reward favors downside because execution gaps can reprice the stock 10-15% quickly if no major customer win is disclosed.
• Pair trade: long NVDA / short QCOM for a 1-3 month horizon. If Qualcomm’s data-center launch slips, the market will continue to pay for the incumbent software moat and supply certainty at NVDA while discounting QCOM’s roadmap risk.
• If you want to express the Apple breakup without fighting the full QCOM tape, consider long AAPL / short QCOM on a 3-6 month basis. The thesis is that supplier insourcing benefits Apple’s structural margin profile faster than it creates credible upside for Qualcomm’s replacement businesses.
• Wait for pullbacks to add to AMD relative-value longs versus QCOM. AMD has clearer data-center credibility and better investor trust; QCOM needs multiple proof points before the market will award it a scarcity premium.
• For event-driven traders, buy downside protection on QCOM heading into the June catalyst rather than chasing the post-rally momentum. The stock can remain supported in the near term, but the skew is attractive if management disappoints on customer naming, shipment timing, or memory economics.