Jensen Huang Just Told Investors to Buy This Under-the-Radar Artificial Intelligence (AI) Chip Stock (Hint: It's Not Marvell)

Nvidia CEO Jensen Huang publicly endorsed Qualcomm, saying Nvidia is not well positioned for mobile devices and recommending investors buy QCOM. The article argues Qualcomm has strength in on-device AI, with automotive revenue up 38% year over year and IoT growth in the high single digits, while also trading at a modest forward P/E. The piece is broadly constructive on Qualcomm’s AI-inference and edge-computing outlook, though the market impact is likely limited to stock-specific sentiment.

Analysis

This is less a headline about one CEO endorsing another stock and more a signal that the AI value chain is bifurcating into two durable pools of profit: compute-heavy training and low-power inference. The second-order implication is that edge AI monetization may accrue to vendors with system-level integration and power efficiency, not just raw model performance. That favors QCOM as a “good enough AI” beneficiary in handsets, autos, and embedded devices while leaving NVDA dominant where latency is less important than scale.

The market risk is that investors over-rotate on the endorsement and pay up for a narrative that is still in early monetization. QCOM’s upside from edge AI likely unfolds over 12-24 months, while the nearer-term catalysts are mix shift in automotive/IoT and any evidence that AI PCs or on-device assistants lift ASPs without cannibalizing margins. The main reversal risk is competitive compression: if handset OEMs standardize AI features and silicon differentiation fades, edge AI becomes a feature, not a moat.

The contrarian read is that NVDA’s comment is not bearish on NVDA; it is a tacit admission that the addressable market is expanding faster than any single chip architecture can monopolize it. That means the more interesting trade may be not long-only QCOM, but expressing the spread between edge inference beneficiaries and incumbents exposed to cyclical handset demand. The broader setup also supports suppliers of memory and connectivity that sit adjacent to inference-heavy devices, especially if AI workloads become persistent rather than episodic.

Key Decisions for Investors

• Initiate a tactical long QCOM vs short NVDA pair for the next 3-6 months; thesis is valuation re-rating in edge AI versus already-discounted perfection in data-center AI. Target 8-12% relative outperformance for QCOM, but trim if NVDA multiple re-expands on a new GPU cycle.
• Buy QCOM on pullbacks with a 6-12 month horizon; use a 5-7% downside stop because the stock’s edge-AI optionality is real but the near-term handset recovery is still the dominant earnings driver.
• Add a small long basket in MRVL/MU as second-order beneficiaries of inference-heavy devices and memory-rich architectures; these names can outperform if on-device AI increases memory bandwidth and connectivity spend. Hold 6-9 months.
• Avoid chasing NVDA on this headline; use strength to sell upside via covered calls or call spreads over the next 30-60 days if you already own the name, since this news mostly confirms a broadening opportunity rather than incremental share gain.
• For higher-conviction expression, consider QCOM Jan-2027 call spreads to capture a 12-24 month edge-AI adoption cycle with limited capital at risk; this is best if you expect AI PC and automotive design wins to compound faster than consensus.