The Bond Market Is Also Asking a Bigger Question About Productivity

In Kevin Gaynor’s last post (see Crowding Out) he argued that fears of public borrowing crowding out private investment—particularly the surge in AI-driven capex—were likely misplaced. The real challenge is not whether savings will adjust—they must, by definition—but rather what shifts in prices, policy, or expectations will be required to expand household savings sufficiently to finance both government and corporate borrowing. In a world where global capital no longer flows as freely, that adjustment process becomes central.

There is also a deeper and more consequential issue lurking beneath these concerns: will the recent wave of AI-related investment actually deliver the productivity renaissance that markets now seem to be pricing in? The recent rally in equity markets—led by tech and infrastructure names—has been powered as much by narrative as by numbers, with investors embracing a vision of AI that promises to transform corporate efficiency and unlock a new era of structurally higher trend growth.

But that optimism carries risks. If the payoff from this capex boom proves slower or smaller than expected—if productivity gains remain elusive or uneven—it could puncture the market’s confidence in both the technology itself and the vast infrastructure build-out it demands. In that sense, AI is shaping up as a high-stakes macro bet: one that either justifies the new valuation regime, or exposes it as another case of over-extrapolated hype.

The good news, as we illustrate below, is that latest US productivity data suggest a renaissance may be underway. The bad news is that this momentum is far from global. In Europe and much of the developed world, productivity growth remains anaemic—held back by high real energy prices, slower AI adoption, and more rigid labour and regulatory frameworks. And even in the US, the scale and durability of recent gains remain uncertain—early signals are noisy, and capex cycles take time to translate into broad output effects.

Furthermore, if the surge in AI-related spending continues to push real yields higher without delivering a commensurate lift in productivity, the result could be a more painful repricing of both equity and credit markets. In that sense, the stakes are not just financial—they are macroeconomic. AI could yet prove to be the next great global growth engine. But it could also turn out to be a capital-intensive detour in a world economy still constrained by deeper structural imbalances.

US Renaissance?

Let’s start, though, with some positives. Recent US data offer genuine encouragement. Non-financial corporate productivity—measured as output per hour—rose by 4.2% year-on-year in 2024, marking one of the strongest gains in around 15 years. One year doesn’t make a trend, but the timing is notable: the rebound has coincided with a sharp acceleration in AI-related investment, particularly in cloud computing, automation, and data infrastructure (see charts below.

However, as the second chart above shows, this capex resurgence has been heavily concentrated in the US—and even there, it appears to have come at the expense of other types of investment, such as structures, plant, and industrial equipment. That matters. It highlights both the opportunities and the limitations of the current cycle: much of the investment is digital, asset-light, and energy-intensive. It is not—at least not yet—a broad-based retooling of the real economy.

Moreover, recent productivity gains may not be solely attributable to AI. A sharp decline in real US energy prices—especially electricity—has probably improved margins and reduced input costs, effectively boosting measured productivity. If those gains prove transitory, the headline improvements may fade.

Beyond the US

In addition to this and reflecting those non-US capex observations above the picture remains much less promising in Europe. In the region as a whole productivity growth continues to flatline, with no obvious signs of an AI-led inflection. Structural headwinds in the meantime abound: elevated energy costs continue to drag on industrial margins and deter capital formation; labour market rigidities slow down reallocation; and persistent fiscal constraints hinder public co-investment in digital infrastructure.

The story is much the same elsewhere. Japan remains constrained by demographic drag and an underdeveloped AI investment ecosystem. Canada and Australia arguably remain more cyclically tied to commodities than to high-tech capex. And across much of the developing world, energy constraints, financing costs, and institutional limitations all act as brakes on productivity-enhancing transformation.

This growing divergence, moreover, has meaningful macro consequences. If only a handful of economies—chiefly the US—are experiencing tangible productivity gains, then a global rise in real yields may prove more damaging than expected. For countries with stagnant productivity but rising capital costs, the risk is a squeeze: financial tightening without a growth offset, undermining both investment and fiscal sustainability. It’s a familiar story in The Age of Constraints—a world where energy, capital, and even knowledge diffusion are no longer abundant or frictionless.

The Bottom Line?

The AI investment boom could continue to support US productivity-linked assets, especially in tech and digital infrastructure. But the broader global investment outlook remains fragile—clouded by rising capital costs, trade frictions, and a widening divergence in productivity performance. While US real yields may reflect stronger growth potential, the sustainability of that narrative still hinge on whether productivity gains persist—and whether US households are willing and able to save enough to fund the investment surge without triggering a broader tightening of financial conditions.

Crucially and perhaps obviously, US trade protectionism adds another layer of risk. By raising input costs, splintering supply chains, and curbing global capital mobility, tariff policy may ultimately weaken the very investment cycle it seeks to shield.

So yes, markets may be right to price a more productive future. But in the age of constraints, capital is no longer cheap, global isn’t frictionless, and growth has to be earned.