Zacks Investment Ideas feature highlights: Meta Platforms, Amazon, Alphabet, Microsoft and Nebius

AI capital spending is accelerating sharply, with hyperscaler CAPEX projected to rise from under $500B in 2025 to $646B in 2026 and $1T+ in 2027. The article argues that markets are increasingly decoupling from Middle East geopolitical shocks and refocusing on fundamentals, even as crude remains above $100 amid Strait of Hormuz tensions. It also highlights Microsoft’s upfront payment to Nebius for AI infrastructure and cites historical data showing the S&P 500 has rarely peaked in May.

Analysis

The market is starting to price this as an infrastructure supercycle rather than a headline-driven trade. That matters because the incremental buyer set shifts from momentum investors reacting to AI narratives into long-duration capital allocating against visible demand, which supports multiples for the largest cloud platforms and vendors with the cleanest balance sheets. The second-order winner is the pick-and-shovel layer: power, networking, memory, and specialized compute vendors should see better pricing power than software, where monetization lags spend.

MSFT is the cleanest beneficiary because its capex intensity is not just defensive; it is tied to preserving scarce AI capacity and customer lock-in. NBIS is higher beta but structurally important because advance payments from a hyperscaler validate neoclouds as balance-sheet bridges in the supply chain, which could compress financing spreads for the space and widen the investable universe of AI infra lenders and lessors. The flip side is that increasing capex intensity can eventually pressure free cash flow optics, so any multiple expansion is most fragile if revenue conversion or utilization slips over the next 2-3 quarters.

The geopolitical setup is becoming less market-relevant unless it spills into shipping insurance, semiconductor lead times, or a sustained input-cost shock. If equities continue to ignore higher oil, that is a bullish signal for index breadth because it implies the market is treating energy as a tax that does not yet break demand; however, that complacency can unwind quickly if crude stays above $100 for another 4-6 weeks and starts feeding through consumer inflation expectations. The consensus is underestimating how much of the AI spend is pre-committed and how little flexibility hyperscalers have to slow without ceding share.

The contrarian trade is not to fade AI capex outright, but to own the beneficiaries with the most operating leverage while fading names where spending is still ahead of monetization. That favors MSFT over the broader Mag 7 basket, and NBIS as a tactical high-beta expression, while AMZN/GOOGL are more exposed to scrutiny around cash conversion if capex keeps accelerating faster than revenue revisions. If the market starts rewarding capex growth independent of near-term FCF, the strongest tape should persist for several months rather than days.

Key Decisions for Investors

• Long MSFT vs. equal-weight Mag 7 basket over 1-3 months: best risk/reward into a capex-upcycle because it combines scale, customer retention, and the highest probability of monetizing spend without multiple compression.
• Tactical long NBIS for 2-8 weeks with a tight stop: high-beta confirmation trade on AI infrastructure demand; attractive if hyperscaler prepayments keep validating financing capacity, but vulnerable to any contract/timing disappointment.
• Overweight META and GOOGL only on pullbacks, not strength: both can participate in the AI capex theme, but near-term upside is capped if the market keeps demanding evidence of revenue conversion before paying for incremental spend.
• Pair trade: long MSFT / short a low-quality AI-infra proxy or unprofitable software name over 1-2 quarters: express the view that capital will flow to the few names with balance-sheet durability and real demand visibility.
• Avoid outright shorting the Mag 7 here; use downside hedges instead: the risk is a multi-month rerating of AI infrastructure assets if capex remains above the current run-rate and geopolitics stays a background factor.