Mag 7 Just Committed $710 Billion To AI Capex

Apple delivered a strong Q2 beat with $111.18B in revenue vs. $109.7B expected and EPS of $2.01 vs. $1.95, then added a $100B buyback and raised next-quarter revenue guidance to 14%-17% YoY; shares rose 3% Friday. Broader markets were mixed but constructive, with the S&P 500 at a record 7,230 and the Nasdaq at 25,114, while the Dow fell 153 points. The article also highlights a highly divided Fed, WTI at $102.28, gas at $4.39, and tariff risks on EU autos that could pressure select sectors.

Analysis

The market is not rewarding “AI spend” uniformly; it is rewarding monetization certainty. That creates a cleaner relative-value setup than a simple index long: names with visible revenue capture from compute scarcity should keep outperforming names where capital intensity is still a promise of future operating leverage. The first-order winner is any platform translating infrastructure spend into near-term revenue or engagement, while the losers are the balance-sheet absorbers whose EPS bridge depends on 2027+ payoff assumptions.

This matters more now because macro is reinforcing the split. A more fractured Fed with energy inflation back in the tape raises the discount rate on duration-heavy stories, especially those with large capex commitments and weaker near-term free cash flow. If oil and gas stay sticky while growth data softens, the market will pay up for cash generative ecosystems and punish “build now, monetize later” narratives more aggressively than it already has.

The next catalyst sequence is unusually tight: AMD and JOLTS can either validate or extend the current semis-vs-software divergence. A weak AI accelerator readthrough would likely hit the entire supply chain, but the second-order damage should be concentrated in software names whose valuations depend on AI attach rates without corresponding hardware proof. Conversely, a strong AMD print could stabilize the build-out trade, but likely only for the semi complex, not the capex-heavy platforms.

The contrarian angle is that the recent selloff in software may already be pricing in an overly linear 2026-2028 skepticism. If management teams slow spend growth even modestly, or if AI revenue contribution inflects faster than expected, the short leg becomes crowded fast. Meanwhile, Apple’s buyback-plus-guidance combination suggests the market still heavily values capital return and execution over narrative spend, which is a warning that the current rotation can persist longer than consensus expects.

Key Decisions for Investors

• Long GOOGL / short META into AMD and JOLTS this week: favor the name with clearer compute monetization and less immediate capex overhang; target 5-8% relative outperformance if AI hardware readthrough remains constructive.
• Short NOW or use put spreads on NOW into the next 2-4 weeks: software multiples remain vulnerable if the market keeps demanding near-term AI ROI; risk/reward improves on any bounce toward prior support after the recent de-rating.
• Maintain long AAPL vs short MSFT pair for 1-2 months: Apple’s buyback and cash return create downside support, while Microsoft remains more exposed to capex scrutiny and margin compression from high-memory costs.
• Trade SMH over IGV tactically for 1-3 weeks: if AMD confirms accelerator demand, semis should keep leadership while software continues to lag; tighten stops if SMH/IGV fails to make a new high by midweek.
• Avoid adding to META until post-AMD confirmation: the name is still the cleanest expression of capex skepticism; only reconsider if management guidance shows spending discipline or ad monetization offsets the investment drag.