Nvidia by the Numbers: 3 Remarkable Facts in Its Q2 Report

Nvidia reported Q2 revenue of $96.2B (+106% YoY), topping the $92.2B consensus, with gross margin rising to 75% (from 72.4%) and adjusted EPS up 120% to $2.22 (vs. $2.09 expected). CEO Jensen Huang guided for at least 70% revenue growth next year versus Wall Street modeling ~45%, driving a reversal to +4.4% in after-hours trading. Data center revenue grew 117% to $89B, reinforcing continued outperformance amid AI chip demand.

Analysis

This is less a one-off beat than a re-anchoring of the AI cycle around infrastructure scarcity. The immediate beneficiary is NVDA, but the bigger implication is that only the largest balance sheets can keep funding the buildout without blowing up free cash flow, which concentrates the spending power in MSFT, GOOGL and AMZN and leaves everyone else mostly as narrative beta. That concentration should keep pricing power high for the chip supply chain and adjacent bottlenecks, while application-layer names may struggle to convert the AI story into near-term earnings torque.

Near term, the market reaction should be driven by short covering and multiple expansion over the next few sessions to weeks; the stock is now trading against a much higher bar for perfection. The 1-3 month test is whether customer capex is still accelerating or simply being pulled forward; if cloud spend growth moderates, the entire AI complex can de-rate quickly because expectations now assume a multi-quarter supercycle. Over 6-18 months, the real risk is digestion: depreciation, power costs, and export restrictions can pressure hyperscaler margins even if revenue stays strong.

Consensus may be missing that scarcity can persist longer than monetization. The better risk/reward is not to chase the basket indiscriminately, but to own the infrastructure monopoly on pullbacks while being selective on the rest of the AI trade. Falsifiers are clear: a guide below the current growth path, a gross-margin plateau, or evidence that order growth is normalizing faster than investors now expect.

Key Decisions for Investors

• Buy NVDA on a 3-5% pullback or first post-earnings consolidation over the next 1-3 weeks; target 10-15% upside over 1-3 months, with the thesis invalidated if forward growth or gross margin stops expanding.
• Pair trade: long NVDA / short TSLA for the next 3-6 months. NVDA has direct earnings torque, while TSLA’s AI exposure is mostly optionality; risk is a broad AI momentum squeeze that lifts all high-beta names together.
• Do not chase AAPL as an AI beneficiary here; if you need a relative-value hedge against AI enthusiasm, AAPL is a cleaner under-exposed proxy than the hyperscalers, but only as a modest short and only if the market starts rotating away from infrastructure.
• Set an alert on the next MSFT/GOOGL/AMZN capex commentary cycle: if customer spending starts to decelerate, reduce NVDA exposure quickly because the entire AI trade is then likely pricing in too much 2025-2026 demand.