Invest in stocks powered by AI spending surge as S&P 500 earnings beat expectations

S&P 500 earnings are beating expectations broadly, with 84% of companies topping EPS estimates and 81% beating revenue estimates in Q1 2026. AI-driven capex from hyperscalers is surging to $725 billion in 2026, supporting record 13.4% net margins and 28% year-over-year earnings growth for the index. Goldman Sachs still sees 6% S&P 500 upside in 2026, but the article flags concentration risk as 40-45% of market cap is now tied to AI-exposed mega caps.

Analysis

The market is no longer pricing an AI theme; it is pricing a capex flywheel. The subtle but important change is that hyperscaler spending is increasingly self-funding through enterprise demand, which means supplier earnings can stay elevated even if end-market software monetization looks noisy quarter to quarter. That argues for continued relative strength in the picks-and-shovels layer, but not across the whole AI complex equally: infrastructure names should keep outperforming application-layer beneficiaries that still need to prove conversion of usage into durable revenue.

The second-order effect is margin bifurcation inside the index. Leaders with scale, data access, and distribution are converting AI into operating leverage, while laggards are likely to experience a slower, more expensive adoption cycle that pressures G&A and product spend before revenue shows up. That widens dispersion and makes index-level longs less attractive than barbelled exposure: long the monetizers and enablers, short the capital-intensive “me too” adopters that may be forced to spend just to keep parity.

The main risk is not a near-term earnings miss; it is a capex air pocket 2-3 quarters out if management teams start demanding proof of ROI. Because the market cap is so concentrated, even a modest reset in AI spending assumptions could trigger multiple compression faster than EPS estimates fall. The catalyst to watch is guidance at the next two hyperscaler prints: if capex growth slows from expansion to maintenance, the trade shifts from beta expansion to balance-sheet discipline and the market will punish the most levered suppliers first.

Consensus appears too comfortable with the idea that AI spending can remain linear. In practice, these buildouts tend to overshoot before normalizing, and the eventual winners are usually the firms with the best pricing power, not the largest budgets. That makes this more attractive as a relative-value rotation than an outright index chase: the upside is still in the supply chain, but the risk/reward is deteriorating in the most crowded mega-cap AI names.

Key Decisions for Investors

• Maintain/scale long NVDA on any 3-5% pullback over the next 2-4 weeks; risk/reward remains favorable as the clearest operating leverage beneficiary, but trim if hyperscaler guidance implies capex deceleration.
• Go long MSFT vs short AAPL for a 1-2 quarter horizon; MSFT has direct monetization and capex-linked upside, while AAPL is exposed to the weakest AI conversion narrative among mega-caps.
• Pair long AMZN vs short TSLA into earnings season; AMZN captures both cloud demand and AI infrastructure spend, while TSLA lacks a clean near-term link to the capex cycle and trades more on multiple sentiment than fundamentals.
• For a hedged trade, buy MSFT call spreads 6-9 months out financed by short-dated out-of-the-money puts on a basket of AI-capex suppliers if guidance risk rises; this keeps upside to the trend while protecting against a capex pause.
• Avoid initiating new broad index longs at current levels; prefer a basket of MSFT/NVDA/AMZN/GOOGL over SPY because dispersion is likely to increase and the index is too concentrated for clean upside capture.