NVDA Q2 Guide of $91B Strips Out China: What the Beat Is Actually Saying

Nvidia reported earnings that shattered historical benchmarks on absolute metrics, yet the stock slipped in after-hours trading, a signal that execution excellence no longer guarantees valuation lift. Revenue came in at $81.6 billion against $79.2 billion consensus, while earnings per share hit $1.87 versus $1.78 expected, with Data Center segment revenue doubling year-over-year to $75.2 billion. The company also announced an $80 billion share buybackA company repurchasing its own shares from the open market. program, effectively signaling confidence in long-term fundamentals.

But the inflection point was buried in the forward guidanceCompany-issued forecasts of future financial performance.. Nvidia guided Q2 2025 revenue to $91 billion, well above the $87-88 billion Street consensus, yet explicitly stated this number assumes zero contribution from mainland China data center compute. This represents management's clearest acknowledgment to date that Chinese hyperscaler demand, previously a growth backstop, is now a known operational headwind. Prior to this earnings, the market had largely assumed China would remain a secondary lever, not an excluded variable.

The macro cross-asset implication cuts across semiconductors, cloud infrastructure names, and AI-dependent equities. If Chinese capex demand materializes below Nvidia's zero-assumption, upside surprises become possible; if geopolitical risk escalates further and U.S. export controls tighten, the $91B guide itself comes under pressure. This also tests whether the Magnificent 7's 38% weight in the S&P 500 can sustain momentumThe empirical fact that winners keep winning over the medium term. when a pillar stock admits to geographical constraint.

Rival chipmakers including AMD, Broadcom (AVGO), and ARM are now trading on the assumption that Nvidia's China exclusion is permanent or near-term unresolvable. Meanwhile, energy markets have rallied on Iran tensions, raising power costs for data center operators. The debate is no longer whether AI capex is real, but whether it can be absorbed at current cost structures and geopolitical frictions.