AI Infrastructure Spending Set to Top $1 Trillion by 2027, UBS Says

According to a fresh report from UBS, the money that the world’s biggest tech firms are throwing at artificial‑intelligence hardware and software is on track to cross the $1 trillion mark by 2027. That’s a staggering figure – think of the entire U.S. defense budget in a single year.

The analysts say that AI‑related capital expenditure will already have topped $600 billion in 2026, and then sprint past the trillion‑dollar threshold the following year. "Demand has outpaced capacity ahead of schedule," the report notes, highlighting that data‑center space, chips and cloud power are being snapped up faster than anyone expected.

What does that mean for the wider economy? UBS estimates that the surge in AI spending added roughly 0.75 percentage points to U.S. GDP growth in the first half of 2026. On top of that, soaring equity prices of AI‑centric firms pumped up household wealth, nudging consumer spending a bit higher.

But the picture isn’t all sunshine and rainbows. The report stresses that there’s still little concrete proof that AI has dramatically lifted overall productivity, nor has it triggered a wave of job losses. Unemployment, according to UBS, is only nudged down by about 0.2 percentage points – essentially a statistical footnote.

Some sectors feel the impact more clearly. Employment in high‑tech fields like software design slipped 1.4 % year‑on‑year, while construction crews that build and maintain data‑centres actually saw jobs rise close to 3 %. The contrast underscores how AI is reshaping the labor map, but not demolishing it.

Financing, however, is beginning to feel the squeeze. As tech giants pour cash into AI gear, their free‑cash‑flow has weakened and appetite for their bonds has dulled. UBS projects that cash flow could turn positive again by 2028, but only if these companies can turn the massive outlay into recurring AI service revenue.

Prices are also creeping up in a few corners – software licences, digital products and even electricity are feeling the heat of higher demand. In the long run, though, the same AI tools could act as a brake on inflation by boosting efficiency.

Bottom line: the AI investment cycle will stay on life support as long as demand keeps outstripping the supply of compute power, and as long as firms can translate shiny new hardware into steady cash streams. Until then, investors and policymakers alike will be watching the numbers very closely.