Signal & Noise Filter
One prominent feature of this cycle is how the evolution of the economy has rendered the business of traditional economic cycle forecasting obsolete. Sacred cows have been slaughtered.
Yield curve inversion is now another noisy data print that fewer people care about.
Cyclical sensitivity to the housing market, which has led the economy into eight out of eleven post-war recessions, has diminished.
The triple punch of high policy and high mortgage rates and a freezing of the housing market was supposed to land the economy into recession since 2022. Any day now….
Trump’s tariff dislocation was supposed to deliver a shallow recession right around now and yet, the Atlanta Fed GDPnow nowcast is tracking a monster 3.8% for Q3 2025 and La Rue is busy raising year-end price targets to match the price action and avoid looking like fools, again.
It wasn’t supposed to end this way; textbooks and models be damned. The latest explanatory cope revolves around the narrowness of economic support coming from the AI spend and data center/power construction.
There is in fact a wedge between booming GDP growth and a weakening labor market. While GDP and the labor market do diverge from time to time, they rarely do so to this degree. 3.8% GDP growth rates are typically associated with non-farm payroll prints in the hundreds of thousands; not negative revisions as have recently occurred.
The disbelief is manifesting itself with comparisons to the 1990’s internet/telecom buildout, bubble-like valuations, and wondering out loud when do things really get max stupid enough to stop this train.
This Bloomberg quote of Morgan Stanley’s Lisa Shallet sums up the current angst:
“Given the GenAI capex story’s centrality to both markets and the economy, the most important question from investors in the next coming year is: What inning are in? When it comes to market-discounting scenarios, we believe we are closer to the seventh inning than the first” (emphasis mine).
As the business cycle is increasingly tied to the fate of the tech boom, the risk appetite and financial health must be followed to gauge sectoral tail risks and subsequent broader recessionary risks.
Since cash flows are the building blocks of financial capitalism, Morgan Stanley’s thesis is best tested by examining the cash flows associated with their claim.
Below is a table that compares the capex of the largest (public) AI hyperscalers with their free operating cash flow.