The Real Risk to 2026’s Job Market Isn’t a Recession. It’s an AI Bubble.

Economists have genuinely become less worried about a conventional downturn. In Bankrate's most recent quarterly survey, the perceived probability of a US recession fell to its lowest level in a year, with 95 percent of surveyed economists placing the odds below 50 percent, down meaningfully from the prior quarter.

On the surface, that reads as good news.

The reason behind it is less comforting. Growth in 2025 and into 2026 has been propped up disproportionately by a handful of forces: AI-related capital investment, tax policy changes, and lower interest rates.

Strip out AI investment specifically, and the picture underneath looks considerably weaker. Hiring has stagnated in several sectors, the housing market has stayed frozen, and consumer sentiment has soured even as consumers continue spending.

Deloitte's 2026 Global Economic Outlook is unusually direct about this risk. The report notes that much of the economy's expected growth is concentrated in the first half of 2026, with domestic demand slowing in the second half as both AI-related consumer spending and business investment growth shift lower.

Critically, Deloitte states plainly that a drop in AI-related spending alone "could be enough to push the economy into a recession," because the rest of the economy isn't currently strong enough to absorb that loss if it happens.

This isn't a fringe view. ING's list of the top ten risks to the global economy in 2026 opens with a strikingly similar scenario: US tech companies fail to adequately monetize their AI investments, questioning the logic behind the scale of hardware and software spending involved.

Tech stock valuations fall, disproportionately affecting the top 20 percent of American earners who hold the majority of domestically-owned equities. That group has been an outsized driver of consumer spending growth over the past several years, even as the bottom 60 percent of earners have struggled.

To understand why this risk carries so much weight, it helps to look at the scale of AI-related capital expenditure driving current growth.

Combined AI infrastructure spending across major hyperscalers, the companies building the data centres and computing capacity behind the current AI boom, is now estimated in the range of $600 billion.

Business fixed investment, a key component of GDP growth, has shown real strength over the past year, driven substantially by this AI infrastructure buildout.

This creates a genuine structural vulnerability. When a meaningful share of measured economic growth depends on continued confidence that this level of investment will eventually generate commensurate returns, any meaningful pullback in that confidence doesn't stay contained to the technology sector.

It moves through construction, through the industries supplying data centre buildouts, through consumer spending tied to equity wealth, and eventually into broader hiring decisions across sectors that have nothing directly to do with AI.

A conventional recession tends to build gradually and telegraph itself through slowing consumer demand, rising layoffs, and softening business investment across the board.

The risk described here is different in character. It's narrower, tied specifically to AI-related valuations and spending, and potentially far faster moving.

If confidence in AI monetization shifts quickly, whether through disappointing earnings, a high-profile setback, or simply a broader reassessment of return timelines, the resulting hit to markets and business investment could arrive with considerably less warning than a traditional demand-driven slowdown.

This distinction matters for workforce planning specifically. A conventional recession gives organizations time to see it coming and adjust hiring plans gradually.

A sharp correction in AI-related valuations wouldn't necessarily offer that same runway, particularly for organizations whose growth plans, capital budgets, or hiring pipelines are more exposed to AI-adjacent spending than they might realize.

Canadian organizations don't need to be directly involved in AI infrastructure to be exposed to this risk.

Global GDP growth, business confidence, and hiring activity are all sensitive to shifts in US market conditions, and a sharp correction tied to AI valuations would likely affect Canadian capital availability, corporate investment decisions, and hiring confidence even for organizations with no direct AI exposure themselves.

The practical takeaway isn't to predict when or whether this correction happens. Nobody, including the economists making these arguments, claims to know that with confidence.

The more useful move is to build workforce and hiring plans that don't assume the current growth environment is guaranteed to continue at its present pace.

Organizations that have spent the past year assuming AI-driven growth simply continues indefinitely are making a bet, whether they've articulated it that way or not.

Recognizing that bet, and consciously deciding how much of the organization's workforce strategy should depend on it, is a more useful exercise than most 2026 planning conversations have made room for so far.