An AI slowdown that takes some air out of the exuberance could actually extend the boom, Wall Street veteran says

Alphabet fell about 12% on Thursday, its worst day on record, as investors questioned whether AI returns will justify massive spending; earnings signals were mixed, with Microsoft reporting strong cloud growth and SK Hynix beating forecasts. Yardeni estimates Amazon, Alphabet, Meta, Microsoft and Oracle will invest nearly $800 billion this year and $1 trillion annually from 2027 through 2030, while capital access, electricity and grid capacity, and political opposition could constrain the buildout. Analysts cited both risks to stocks and potential benefits from a slower buildout, including less inflation pressure and lower upward pressure on long-term rates.

Analysis

The market is shifting from “AI exposure” to “who can finance and monetize the buildout.” That favors platforms with demonstrated cloud demand and internal funding capacity over infrastructure borrowers whose returns depend on continued easy capital. The important second-order risk is that capex constraints can hit both sides of the trade: fewer projects ease power and equipment bottlenecks, but also weaken incremental demand for memory and other AI hardware, including SK hynix. Conversely, grid scarcity and local permitting can redirect spend toward power generation, transmission, and electrical equipment rather than eliminate it.

Near term, Alphabet’s sharp repricing raises the chance of forced de-risking across AI-linked names; it does not by itself establish that the investment cycle is ending. Over 1–3 months, watch whether cloud growth and AI-related monetization keep pace with capex, alongside free-cash-flow trajectories and borrowing spreads for less-capitalized infrastructure providers. Over 6–18 months, local permitting and power availability may matter more than federal legislation, while a slower buildout could ease tech-goods inflation and long-rate pressure. That rate benefit is conditional, not yet a clean macro trade.

Contrarian view: treating capex as uniformly wasteful misses the possibility that financing discipline improves returns for the strongest platforms. But the opposite consensus error is assuming hyperscaler balance sheets make every project economic. No evidence here settles return on invested capital or project-level utilization; those are the key verification points.

Key Decisions for Investors

• Consider a measured MSFT-long / GOOG-short relative-value position only after the post-earnings volatility settles. Thesis: favor stronger reported cloud growth and positive free cash flow against Alphabet’s weaker cash-flow signal. Reassess at the next results: reduce or exit if Microsoft’s capex acceleration outpaces cloud growth, or Alphabet stabilizes cash flow without sacrificing growth. This is a relative thesis, not a claim that either name is cheap.
• Do not chase SK hynix’s earnings surprise as a standalone AI demand signal. Track memory pricing, customer order visibility, and capex plans; a buildout slowdown could reverse the near-term benefit even if long-run AI demand remains intact.
• Watch for a widening funding-cost gap between established platforms and highly leveraged AI infrastructure providers. If spreads widen further, favor well-funded cloud platforms over weaker borrowers; treat this as an alert rather than a direct short until identifiable issuers and financing terms are verified.
• Keep power and grid-equipment exposure on a catalyst watchlist, not as an automatic long. Add only if project backlogs, utility interconnection data, or local approvals confirm that constrained power is shifting spend toward those suppliers; thesis weakens if data-center cancellations materially reduce order pipelines.