This Could Be the Most Underrated Artificial Intelligence Stock to Buy Right Now
Source: Nasdaq

Qualcomm announced an AI-infrastructure partnership under which Amazon could purchase up to $60 billion of its AI chips, while warrants could allow Amazon to acquire up to $4 billion of Qualcomm stock at $161.26 per share. Qualcomm shares are up roughly 6% year to date and trade at 21x trailing earnings, or about 18x projected earnings, despite revenue declining 4% in its latest quarter. The Amazon relationship could expand Qualcomm's position in custom AI chips and provide a growth catalyst, although the potential purchase volumes remain contingent rather than booked revenue.
Analysis
The market should not capitalize the reported arrangement as data-center revenue until Qualcomm discloses binding purchase commitments, delivery schedules, and chip-level economics. The key issue is whether this represents merchant silicon revenue with gross margins comparable to QCT, or lower-margin custom-design/NRE work whose revenue is recognized over a multi-year qualification cycle. Qualcomm’s near-term valuation re-rating is therefore more likely to come from proof of a credible non-handset AI revenue stream than from material EPS contribution in the next two quarters.
The warrant structure creates asymmetric signaling but also a potential ceiling: Amazon has incentive to concentrate purchases and support ecosystem adoption, while incremental dilution and the effective strike level can limit upside if the stock approaches that level before revenue converts. The more important second-order effect is validation for Qualcomm’s IP, connectivity, and edge-inference stack; that could improve its negotiating position with PC OEMs and automotive customers, where design wins have longer duration and less concentrated customer risk than hyperscale infrastructure.
Consensus may overstate the read-through as direct share loss for NVDA. Hyperscalers can add custom accelerators for tightly defined workloads while retaining Nvidia for training, software portability, and rapidly changing model architectures. A sustained QCOM rerating requires evidence that its AI mix raises consolidated gross margin and reduces dependence on Android handset replacement cycles; absent that, this is a narrative catalyst rather than a durable earnings inflection over the next 1-3 months.
AllMind Terminal
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request TrialMarket Sentiment
Overall Sentiment
moderately positive
Sentiment Score
0.48
Ticker Sentiment
Key Decisions for Investors
- Establish a small long QCOM / short SOXX pair over the next 1-2 weeks only if Qualcomm confirms binding volume commitments or provides a fiscal revenue contribution range. Target 10-15% relative upside over 6-12 months; exit if management cannot identify a production ramp by the next two earnings calls.
- Do not chase QCOM on the initial announcement. Accumulate on post-event weakness below its pre-announcement valuation range only after checking warrant dilution, vesting milestones, and whether Amazon’s purchase obligation is enforceable rather than discretionary.
- Use a QCOM long / NVDA short only as a short-duration event hedge, not a structural substitution trade. Cover the NVDA short if hyperscaler capex commentary or Nvidia guidance indicates continued training-demand acceleration; Nvidia’s software moat makes broad displacement unlikely within 12 months.
- Monitor Qualcomm’s data-center/AI revenue disclosure, segment gross margin, and handset guidance at the next two reports. A lower gross-margin custom-chip mix, no disclosed backlog, or renewed Android inventory weakness would falsify the bullish thesis and argue for closing QCOM exposure.
More News
- Asian shares fall as Alphabet earnings miss drags Wall St futures lower
- As Waymo expands overseas, it remains a niche cash-burning business
- Wall Street looks past Fed pause to tech earnings and AI rebound
- Microsoft’s cloud growth accelerates, and so too does the spending
- SK Hynix says the memory-chip shortage will last until 2030—and more than half its production is booked for the next two years
- U.S. stock futures drop as Alphabet slides on soft earnings, Microsoft firms