First thing this week, in “The Numbers Behind America’s AI Dominance,” I cited capex and R&D figures which underscore the extent to which the mega-cap names we all know and love (in some cases to hate) have come to monopolize (figuratively and literally) the AI arms race.
You can read the linked article for yourself (it’s short), but the gist of it is just that the Magnificent 7 are (is) blowing through hundreds of billions of dollars in a bid to stay at the forefront of what Jensen Huang swears is a new industrial revolution.
What about the rest of the market — the so-called “S&P 493”? How’s capex holding up outside the vaunted septet whose share of index market cap is rivaled only by its share of market headlines?
The answer, as it turns out, is not well. Or anyway not as well.
The figures, from SocGen’s Andrew Lapthorne, are remarkable, even as they aren’t surprising.
“There is a lot of talk of a capex wave, with corporates keen to invest in AI as they seek growth, [b]ut evidence from US reports and accounts data suggests the capex surge is very much a mega-cap Magnificent 7 story,” Lapthorne wrote.
Capex growth ex-Mag7 has slowed sharply in recent quarters, and it’s no secret why. As Lapthorne explained, “the simple reason” is that outside of the big names, cashflow’s declining.
I wonder how much of the sp497 decline in cashflow is as a result of share buybacks.
The cash flow comes first. That is the supply. The uses include buybacks, R&D, Capex and the cost of growth. The faster a firm, especially a large one grows, the more capital is required to support that growth. If a firm is doubling in size it will need to double its working capital, plant capacity, labor, support and other expenses. If the growth in some of that support can grow at a slightly lower rate as a result of short-term productivity increases, that will help, but high rates of growth, say 30% or more, lots of cash will be required. Increasing competition, market forces, input prices and other factors can compress operating margins and lower free cash flow. IMO, buybacks are the least defensible uses of free cash.
My guess is if you were to screen and exclude for ai thematic and government subsidy adjacent capex, e.g. ira, you are deeply negative.
I never bought in to the storyline about buying the Russell 2000 because, collectively, it was supposed to “catch up” in performance with SP500. However, I am looking at a few smaller companies that don’t have a debt problem, have increasing cash flow, and have cap ex budgets for solid and sustainable business opportunities (not solely for AI).
This also helps to explain the tepid growth in robotics & automation gear where we are supposed to be seeing a spending boom as “nearsourcing” gains more traction.