700 Billion Reasons to Stay in AI This Summer

Q1 2026 S&P 500 earnings growth is tracking at 27.1%, up from 13.1% estimated on March 31 and potentially the strongest since Q4 2021, but the gains are heavily concentrated in Tech/AI. AMD reported revenue of $7.44B, up 36% year over year, with Data Center revenue surging 57%, while the S&P 500 Equal Weight Index is up 9% versus 16% for the S&P 500 and 32% for XLK since the late-March lows. The article argues that hyperscaler AI capex is approaching $700B in 2026 and could exceed $1T in 2027, supporting continued leadership in AI-related stocks despite broader market caution.

Analysis

The tape is being driven by a narrow but unusually durable capex cycle, not just a valuation re-rating. The key second-order effect is that hyperscaler spending is now functioning like a quasi-public utility order book for the AI supply chain: chip designers, power equipment, grid modernization, cooling, and data-center infrastructure all have earnings visibility that the rest of the market lacks. That matters because it compresses fundamental risk premia for beneficiaries while widening it for cyclicals whose demand depends on consumers or rates.

The market’s internal split suggests this is less a broad bull market than a capital-expenditure supercycle concentrated in a few balance-sheet winners. That should continue to reward names with direct exposure to AI deployment bottlenecks—especially power generation, transmission, and electrical equipment—because those are the areas where spend is hardest to defer once capacity planning is locked. It also implies that downstream software/platform winners can lag hardware and infrastructure if monetization remains a later-phase story.

The main risk is not immediate demand collapse; it is a policy or funding hiccup that slows the pace of capex into 2027. If hyperscalers start defending margins, or if macro volatility forces even a modest pause in incremental buildout, the highest-multiple AI beneficiaries could de-rate quickly because the market is pricing a multi-year earnings runway. Geopolitically, lower oil is a short-term tailwind for beta and margin, but if energy prices stabilize too high, it could eventually pressure the non-AI consumer cohort and keep the earnings dispersion theme intact.

The contrarian takeaway is that the consensus may be underestimating how concentrated the current earnings engine is. That is bullish for leadership persistence, but it also means index-level participation can remain mediocre even as headline earnings look strong. In other words: the market can keep going up, but the average stock may not participate unless it is tied to AI infrastructure, electrification, or balance-sheet-driven secular growth.

Key Decisions for Investors

• Stay long AMD over the next 3-6 months; use pullbacks toward support to add, because it has direct leverage to the capex cycle and still has upside if AI infrastructure budgets re-accelerate in H2.
• Initiate a basket long in GEV/MSFT/AMD versus a short in equal-weight S&P proxies (e.g., long XLK vs short RSP) for a 1-3 month relative-value trade; thesis is earnings dispersion, not broad market beta.
• Buy GEV on any 2-4% weakness with a 6-12 month horizon; the best risk/reward is in the power-and-grid bottleneck names where order visibility is strongest and cancellation risk is lowest.
• Avoid chasing non-AI cyclicals that are trading on multiple expansion rather than earnings revision; if you want market exposure, prefer a pair long AI infrastructure / short consumer-discretionary or rate-sensitive names.
• For tactical upside, use call spreads in AMD or GEV over the next 1-2 quarters rather than outright longs; this captures upside from continued capex surprises while limiting drawdown if the market starts discounting 2027 slowdown risk.