3 Undervalued AI Stocks Investors Can Buy Now in September

Motley Fool contributor Parkev Tatevosian highlighted three relatively inexpensive AI stocks as potential investments, although the supplied article text does not identify them. The accompanying Stock Advisor promotion stated that Nvidia was not among its current 10 best-stock selections, while citing a hypothetical $1,000 Nvidia investment from April 2005 growing to $662,392. The item is opinion-driven promotional content rather than a new fundamental, earnings, or market-moving development.

Analysis

This is low-information promotional flow rather than a fundamental catalyst, and neither NVDA nor QCOM should re-rate on it. The more relevant positioning implication is that repeated “cheap AI” framing can pull marginal retail demand toward lower-multiple semiconductor names, but sustained institutional rotation requires evidence that AI inference demand broadens beyond hyperscaler accelerator purchases. Watch relative performance of QCOM versus SOX and NVDA: a durable breakout would need handset AI units, edge inference revenue, or licensing upside to translate into estimate revisions.

For NVDA, the near-term risk/reward remains governed by supply conversion, customer capex budgets, and gross-margin progression as new product ramps—not retail recommendation activity. In a 1-3 month window, any incremental enthusiasm toward AI laggards may be a modest headwind to NVDA’s relative momentum if investors seek valuation catch-up; over 6-18 months, however, the key competitive question is whether inference shifts to ASICs and lower-power edge devices fast enough to reduce accelerator intensity. That thesis is falsified if cloud capex remains concentrated in GPU clusters and NVDA’s next two earnings reports maintain data-center growth and margin guidance.

The contrarian point is that “cheap AI” is often a valuation-screen narrative masking different earnings durability. QCOM has meaningful handset-cycle, China, and OEM inventory sensitivity, while NVDA’s premium reflects materially higher visibility and ecosystem lock-in. Treat any QCOM rally without upward revisions to handset units, AI-enabled device mix, or automotive/IoT guidance as multiple expansion vulnerable to reversal.

Key Decisions for Investors

• No event-driven position based on this article; classify as non-catalytic retail commentary and avoid chasing either NVDA or QCOM at the open.
• Maintain NVDA as the higher-quality AI exposure only on pullbacks tied to broad semiconductor risk-off, not company-specific demand deterioration. Reassess if next-quarter data-center guidance decelerates materially or gross-margin guidance falls below consensus; these are more relevant thesis breakers than relative valuation.
• Set a watch alert for QCOM relative strength: consider a 1-3 month long QCOM / short SOXX or SMH hedge only after consensus handset-unit or EPS estimates rise, supported by verifiable AI-phone sell-through. Without estimate revisions, do not underwrite the catch-up trade.
• For a valuation-rotation expression, prefer a small, defined-risk QCOM call spread rather than outright stock only if QCOM outperforms SOX for 10-15 trading days and breaks above its prior earnings-date high; risk is a renewed China handset or inventory reset, which would compress both earnings and multiple.