Nvidia’s blockbuster earnings; SpaceX files for IPO - what’s moving markets

Nvidia posted record April-quarter sales of $81.6B, up 85% year over year, and net income of $58.3B, more than tripling from last year, though shares were muted because the outlook excluded China and only modestly topped estimates. U.S. futures rose about 0.2% as investors balanced the AI earnings strength against geopolitical risk and falling oil prices, while SpaceX filed for a potential $80B+ IPO and OpenAI was reported to be preparing a filing as early as September. Brent crude last traded at $106.34 a barrel amid hopes for a U.S.-Iran peace deal and reopening of the Strait of Hormuz.

Analysis

NVDA remains the cleanest expression of the AI capex cycle, but the more important signal is that demand is broadening from model training into persistent inference workloads. That matters because inference is less episodic and more margin-durable: if agentic AI actually drives enterprise deployment, spending should become recurring rather than project-based, extending the runway for GPU allocation and software attach. The market’s muted reaction suggests the bar is now set less by beat size and more by whether supply, memory, and networking constraints can keep up without compressing realized margins.

The second-order winner is likely not GS/MS from underwriting fees alone, but the broader private-markets ecosystem that will intermediate these mega-listings. If SpaceX and OpenAI move toward public markets, expect a valuation reset across late-stage AI and frontier-tech private rounds; listed names with credible AI monetization should trade with a richer scarcity premium, while cash-burning peers without infrastructure control may face a tougher funding tape. The flip side is that mega-IPOs can temporarily siphon risk capital away from semis into a few headline trades, creating short-term relative underperformance for the rest of the AI basket.

On geopolitics, the oil setup is less about the absolute Brent print than the probability of a rapid de-risking of energy supply. Even a partial reopening of trade routes would likely trigger a swift unwind in oil volatility and the inflation-risk premium embedded in rates, which is constructive for duration-sensitive growth and semis. The contrarian read is that the market is too focused on a binary peace outcome and not enough on the path dependency: a fragile ceasefire still leaves shipping insurance, tanker routing, and Treasury term premium elevated for weeks, so the disinflation impulse may lag the headline.

The main risk to the bullish AI narrative is not demand failure but supply normalization: if foundry capacity, advanced packaging, and HBM availability catch up faster than expected, NVDA’s pricing power can compress even while unit volumes rise. That is the setup for a “good but not good enough” derating over 1-3 months. In other words, the current move is strongest as a relative trade, not a blind outright long.

Key Decisions for Investors

• Stay long NVDA vs. SOXX on a 1-3 month horizon; prefer a relative-value position because the next leg is likely driven by share gains and supply bottlenecks rather than multiple expansion. Risk/reward: 2:1 if inference demand keeps compounding; stop if semicap equipment starts outperforming semis on improving supply visibility.
• Buy NVDA upside exposure via call spreads into the next 4-8 weeks rather than outright stock. Use defined-risk structures because the market appears to be discounting perfection; the payoff is strongest if guidance implies sustained demand beyond China limitations.
• Long GS/MS only as a basket if private-market issuance accelerates; otherwise treat them as low-beta beneficiaries. Better implementation is a 6-12 month call on capital markets activity, with upside from IPO/underwriting and downside limited by diversified fee streams.
• Pair long QQQ / short XLE for a 1-2 month geopolitical unwind trade if Iran headlines de-escalate. The thesis is that falling oil volatility should compress inflation risk premium faster than growth assumptions change, supporting duration and megacap tech.
• Avoid chasing the broad AI basket until post-IPO valuation read-through is clearer; use any weakness in non-NVDA AI beneficiaries as a better entry point. Consensus is likely overestimating how much second-tier AI names will benefit if capital consolidates into a few public winners.