DAVOS, Switzerland, Jan 19, 2026, 14:03 CET
- Barclays economist Christian Keller noted that while the AI investment surge has cushioned markets against tariff shocks, the reliance on debt-fueled growth increases potential downside risks.
- Keller warned that tariff-driven price hikes might still affect U.S. core goods in 2026 as inventory levels decline.
- This month, KPMG’s Diane Swonk warned that leaning heavily on AI-driven investment strategies could leave the economy vulnerable to stock market corrections.
Barclays economist Christian Keller warned the AI investment surge fueling markets might be a bubble — “probably a good bubble” — yet the rising debt used to finance data centres ups the risk if it bursts. He made the remarks on the World Economic Forum’s Radio Davos podcast as the Swiss gathering kicked off. (World Economic Forum)
The episode dropped on Jan. 18, coinciding with the Forum’s release of its newest Chief Economists’ Outlook, a survey updated every four months, per the show notes. (Apple Podcasts)
For investors, the issue goes beyond whether a few tech stocks are overpriced. Keller said the jump in AI-driven valuations has morphed into a broader macro narrative — one powerful enough to influence spending, hiring, and set the mood for 2026.
He noted the latest Outlook reveals economists stuck between harsh realities — trade tensions and geopolitical pressures — and a market surge fueled by hopes for AI-driven productivity gains. He hinted that something’s bound to give, or at least recalibrate.
On tariffs, Keller noted that the initial economist worries haven’t unfolded neatly. Delays, exemptions, and companies rushing imports ahead of tariffs—hoarding stock before duties took effect—eased the immediate impact, he said.
That buffer won’t last. Keller warned that price increases for U.S. core goods might hit in 2026, once inventories run low and companies face higher tariffs on imports, leaving less wiggle room to absorb those costs.
He added that the AI boom has sparked a “wealth effect” — lifting spending as stock portfolios gain value. According to him, this helped cushion the blow from trade-policy shocks in 2025. But it also means any sudden drop in valuations could quickly drag down consumption.
Keller noted a shift in the financing mix. He highlighted “record debt issuance” earlier this year for projects linked to data centres and the energy and networks supporting them. But he warned that increased leverage raises risk when momentum slows.
He contrasted this with the 2007-09 global financial crisis, noting that back then leverage was widespread among households and banks. “This time, the leverage is more concentrated,” he said, adding that the spending targets “tech infrastructure,” not housing — which partly explains why some investors label it a “good bubble.” He pointed to the 1990s fibre-optic buildout as a case where investors took losses, even as the infrastructure eventually fueled growth.
Keller cautioned that few portfolios are genuinely shielded. Given the heavy influence of U.S. shares in global markets and the dominance of mega-cap tech in many funds, he noted that savers who believe they’re avoiding the sector can still be hit via pensions and index products.
The Motley Fool offered a take along the same lines, though from a different perspective. In a Jan. 17 column, Jon Quast forecast a stock market “correction” in 2026—a term investors use for a decline of at least 10%—citing a potential power supply bottleneck that might slow AI infrastructure investment. (The Motley Fool)
In its Jan. 12 Economic Compass essay, KPMG described the outlook as the “two faces of the economy,” caught between tariffs, immigration restrictions, and geopolitical strains on one hand, and AI-driven transformation on the other. The firm pinpointed the key challenge as figuring out how to “break out of jobless growth.” (KPMG)
Diane C. Swonk, chief economist at KPMG US, called inflation “a regressive tax” and cautioned that heavy reliance on AI has increased vulnerability to stock market corrections. She also pointed to a wildcard: speculators “using debt to finance data center construction,” amid tight chip supply and an electricity grid under strain. (KPMG)
Timing remains tricky. Keller suggested economists might have jumped the gun expecting tariffs to inflict their worst damage right away — yet he warned it could be equally mistaken to think the impact has already played out, with delayed price pass-through still “in the pipeline” for 2026. Should tighter inflation pressures coincide with a slump in AI-related investment, the very wealth effect that boosted spending could turn into a drag.
