Prediction: This Is Where Nvidia Stock Will Be When September Ends

Nvidia’s fiscal Q2 results surged, with revenue up 106% YoY to $96.2B and non-GAAP EPS up 120% to $2.22 (vs. Street $2.09 EPS on $92.3B revenue). The company guided for ~70% fiscal 2028 revenue growth (supply-constrained) versus analysts expecting ~44%, prompting nearly a 9% jump in the stock after the Aug. 26 release. The article also highlights bullish AI infrastructure demand signals (cloud capex backlog >$2T) and argues NVDA trades at a discounted valuation (29x earnings vs. Nasdaq-100 at 34x), implying further upside into September.

Analysis

The key market implication is not that AI demand is still healthy; that is already consensus. The more important signal is that the bottleneck is shifting from demand to industrial capacity, which tends to extend the cycle for the true enablers: NVDA gets a longer earnings runway, while TSMC captures the capital-intensity premium via higher utilization, advanced-node mix, and pricing power on constrained wafers. In the next 1-3 months, that should keep momentum factor money in the large-cap AI complex, but the cleaner relative trade is likely upstream semicap infrastructure rather than just the headline beneficiary.

Second-order effects: if TSMC’s capex actually converts into usable capacity, the group that benefits most over 6-18 months may be the equipment and packaging ecosystem, not just GPU vendors. The risk is that investors extrapolate peak growth rates too far ahead of the supply ramp; once capacity arrives, NVDA’s revenue growth can decelerate faster than the multiple can absorb, especially if hyperscaler procurement shifts toward internal accelerators or networking budgets start substituting for GPU spend.

Contrarian view: the market may be underpricing how much of NVDA’s near-term upside is already de-risked by the post-earnings re-rating, while underpricing the likelihood of a broad semicap capex upcycle. The thesis breaks if next-quarter sequential growth slows, if TSMC delays capacity additions, or if hyperscaler capex prints below the implied trajectory. This is a momentum-favorable setup, but not a blind chase: the right question is whether the earnings surprise can migrate from NVDA into the supply chain without margin dilution.

Key Decisions for Investors

• Long NVDA on 3-5% pullbacks over the next 2-4 weeks; use a tight thesis stop if next-quarter guidance stops inflecting higher or if the stock fails to hold post-earnings breakout levels. Expect continued multiple support for 1-3 months, but do not underwrite the full 12-month upside without another upward revision cycle.
• Add TSM as a lower-beta way to own the AI capex cycle for 6-18 months; the cleaner payoff is capacity monetization and foundry pricing, with less valuation sensitivity than NVDA if the market rotates from growth to cash flow durability.
• Pair trade: long NVDA / short SOXX or SMH for the next earnings season if you want to express idiosyncratic share gains rather than the whole sector beta. The trade works best if NVDA keeps beating while the broader semi index stalls after a strong 2026 run.
• Watch-list only: consider long semiconductor equipment on weakness (AMAT, LRCX, KLAC) if TSMC’s capex hike proves durable in the next two quarters. This is the higher-upside second-order beneficiary, but it needs confirmation that capex is being converted into usable capacity rather than just headline spend.