Construction

Take a look at the chart from the census data below. Adjusted for construction inflation, it shows real U.S. construction spending through August 2026. There is a lot of information embedded in this chart. To start, it confirms that the level of concentration in AI is not only happening in market indices, it’s also happening in the real economy. It’s THE bet. That also applies to power, which is needed for data centers. Everything else is contracting, and most of them, if not all, are very sensible to interest rates.

That last one, manufacturing, deserves attention.

Onshoring and a factory-building renaissance were central promises of this administration. The spending data shows the opposite: manufacturing construction is falling faster than any other category. This line item in itself will weigh a lot in the midterms.

Meanwhile, capital is piling into a single asset class. Data centers are being financed with a growing amount of debt, built on the assumption that AI demand will grow fast enough to fill them and pay for them. The economics are far from proven. These assets depreciate quickly, the chips inside go obsolete in a few years, and returns depend on customers whose own profitability is still an open question.

When one sector absorbs this much of the economy's building capacity, the risk is no longer just to that sector. If the AI trade stumbles, construction has no other engine running. Housing is contracting, factories aren't being built, and the commercial pipeline is thin.

That's not a diversified investment cycle. It's a concentrated bet, made with borrowed money, in an economy at war, with a central bank that aims to rise borrowing cost to control inflation that is caused by too much government spending. We’re doubling down.

Want to know more? You can register for free at Fund@mental.

#iamfundamental #soyfundamental #wealthmanagement #familyoffice #financialadvisor #financialplanning #policymistake #ratecut #stagflation