Nvidia Stock Is Down 17% From Its High. Is the Artificial Intelligence (AI) Leader Finally Cheap?

The article argues Nvidia could rebound sharply as key July earnings from AI hyperscalers may confirm ongoing AI capex strength. It cites $650B in data center capex this year versus more than $1T next year, and highlights that Microsoft’s FY2027 capex guidance could provide confirmation for Nvidia’s bull case. It also notes Nvidia trades at 21.7x forward earnings (roughly in line with the S&P 500), implying limited valuation headwind ahead of catalysts from MSFT and TSMC.

Analysis

The setup is less about near-term unit demand and more about whether the market believes the AI capex cycle has a second leg. If MSFT and TSM both confirm that spending is still rising and still GPU-heavy, NVDA can re-rate quickly because the stock is trading on revision momentum, not just trailing narrative. The key second-order effect is that strong hyperscaler capex can lift the entire AI supply chain, but the biggest winner is the company with the highest share of incremental wallet spend; that argues for NVDA over lower-quality AI beneficiaries and for skepticism toward names that need multiple expansion without earnings acceleration.

The main risk is that headline capex stays high while economics shift away from merchant GPUs toward custom silicon, networking, and internal optimization. That would keep the AI buildout story intact but leak economics away from NVDA over a 6-18 month horizon, causing the stock to lag even if the sector remains strong. In the next 1-3 months, the trade lives or dies on forward guidance from MSFT and the AI commentary from TSM; the late-August NVDA print is the final confirmation point.

Consensus is likely underweighting how much of the move depends on mix, not just spend. If hyperscalers frame AI returns as improving, the market will forgive a rich absolute multiple; if they emphasize monetization lag, the market will de-rate the whole complex despite continued capex. Falsifiers are simple: a soft FY27 capex guide from MSFT, an AI revenue miss at TSM, or NVDA guidance that fails to show upside versus already-elevated expectations.

Key Decisions for Investors

• Buy NVDA Sep/Oct call spreads into the MSFT/TSM earnings window; best asymmetry if guidance confirms another capex leg, with limited premium at risk if the setup disappoints.
• Pair trade: long NVDA / short AMD for 1-3 months. If AI spend remains GPU-centric, the market should reward the dominant platform provider more than the nearest share-taker.
• Add MSFT only on a post-earnings pullback if FY27 capex guide steps up meaningfully; if the guide is merely inline, treat that as a warning that the AI spend narrative is already fully priced.
• Use TSM as the verification trade, not the alpha trade: a strong AI semi print supports the group, but a miss would be an early signal that spend is rotating away from merchant GPUs toward custom silicon.
• If NVDA gaps higher into earnings without follow-through in MSFT/TSM guidance, take partial profits; the thesis is confirmation-driven, not valuation-driven.