David Hay — Time For Some Appreciation of Depreciation (Haymaker, May 2026)

One-line take: A short macro caution daily — no stock picks. Per Vincent Deluard / StoneX: academic studies show heavy-capex companies lag the market long term, yet we're in one of the greatest capex sprees ever — from the famously "capital-light" superstars. The four leading hyperscalers — Microsoft, Meta, Google, Amazon ("the Fab Four") — are on track to spend ~$700B this year, mostly on AI, setting up an extraordinary spike in depreciation over the decade (especially using a realistic ~4-yr asset life given chip obsolescence; some say ~7 — call it ~5). Despite the drag, S&P 500 earnings estimates stay very optimistic and returns are uncertain — a dramatic divergence between profit margins (reported earnings) and free-cash-flow margins (which Buffett and others weight more). The binge may boost US productivity (more sales/earnings with lower headcount), and a further irony: AI may benefit the less-glamorous slices of corporate America more than the Mag 7 themselves. Feeds the master macro viewpoints.

1. Stocks & names mentioned

No individual stock picks in this piece — it's a macro caution daily on AI capex and depreciation. Referenced only as the spending cohort: the "Fab Four" hyperscalers Microsoft, Meta, Google (Alphabet), Amazon (~$700B 2026 capex), with Warren Buffett cited on FCF margins and Vincent Deluard / StoneX the source. The substance is in the talking points below and the master macro viewpoints.

2. Talking points

Heavy capex lags long term (Deluard / StoneX)

Academic studies show heavy-capex companies lag the market over the long term — yet we're seeing one of the greatest capex sprees ever, from the famously "capital-light" superstars.

The Fab Four — ~$700B this year

The four leading hyperscalers — Microsoft, Meta, Google, Amazon — are on track to spend ~$700B this year, mostly on AI. That sets up an extraordinary spike in depreciation over the decade, especially using a realistic ~4-year asset life given chip obsolescence (some say ~7 years — call it ~5).

The profit-vs-FCF divergence

Despite this accounting drag, S&P 500 earnings estimates remain very optimistic, and returns on the outlays are uncertain — creating a dramatic divergence between profit margins (reported earnings) and free-cash-flow margins, which Buffett and others consider more important.

The ironies — productivity, and who actually benefits

The binge may boost US productivity (growing sales/earnings with lower headcount). A further irony: AI may benefit the less-glamorous slices of corporate America more than the Mag 7 / the "Fab Four" themselves.

Summary derived from the paid Haymaker Substack post (text in transcript.txt) for personal study. Not investment advice. © Haymaker / David Hay for source material.