Nvidia May Have Just Eliminated Its Biggest Risk

Nvidia delivered a blowout quarter with revenue up 106% to $96B and profit up 126% to $59B. The key competitive overhang from customer in-house chips eased as AWS agreed to buy an additional 2 million Nvidia GPUs over the next two years (with potential purchases of millions of CPUs). Overall, the earnings beat plus deeper AWS commitment reinforces Nvidia’s AI platform dominance and supports a bullish outlook for NVDA.

Analysis

The market should read this less as a demand surprise and more as evidence that custom silicon is additive, not substitutive, at the hyperscaler level. That matters because it weakens the bearish narrative that every internal chip project is a direct unit-share loss for NVDA; in practice, customers are segmenting workloads, keeping Nvidia for frontier training and burst capacity while using house chips to lower steady-state inference cost.

Near term, the biggest beneficiary is NVDA’s multiple, not just its revenue line: it reduces perceived customer concentration risk and supports a premium as investors extend the duration of growth. The second-order loser is the "Nvidia is being disintermediated" trade in AMD and INTC, whose share-gain cases now look more dependent on price/perf gains than on a broad hyperscaler migration. The broader supply chain still wins because total AI capex remains additive; advanced packaging and high-bandwidth memory stay constrained, which keeps pricing power with the ecosystem.

Contrarian view: consensus is probably overreading this as a clean victory for NVDA. The more important implication is bargaining power—large buyers will keep dual-sourcing to preserve leverage, which can cap gross margin expansion over 6-18 months even if demand stays strong. The falsifier is not a single order, but next-quarter data-center growth and gross-margin guide: if Nvidia’s mix or margin steps down while AWS capex flattens, the bullish read-through fades quickly.

Key Decisions for Investors

• Long NVDA vs. SMH on any 1-2 day post-news fade; hold 1-3 months. Target is relative outperformance as this reduces the bear-case discount, but cut if next earnings guide implies data-center growth deceleration or margin compression.
• Pair trade: long NVDA / short AMD for 1-3 months. Thesis: the market is underpricing Nvidia’s centrality and overpricing AMD’s ability to displace it in hyperscaler budgets. Invalidated by a major AMD design win at a top cloud provider or a sharp AMD gross-margin inflection.
• Avoid chasing AMZN outright on this headline. Use it as a watch item for the next AWS capex and free-cash-flow print; if AI spend lifts revenue without margin erosion, AMZN can work, but if capex intensity rises faster than monetization, the stock should lag.
• Alert on NVDA gross margin and cloud-customer concentration commentary next earnings. If management confirms pricing discipline despite customer self-supply, add to longs; if discounts increase, reduce exposure.