The nature of work is being redesigned by artificial intelligence. We are already seeing early signals. Block, the parent company of Square and Cash App, recently announced plans to reduce its workforce significantly as part of a restructuring driven by AI adoption. As CEO Jack Dorsey noted, a smaller team equipped with the right AI tools can now accomplish more and do it better. Markets responded positively, with the company’s stock rising sharply on the news.

This is a reminder that AI is not just another technology layer; it is changing how organizations think about productivity, scale, and the role of human labour. Tasks that once required large teams are increasingly being automated, streamlined, or augmented by intelligent systems. Investors tend to reward these transitions because they see AI lowering operational costs and increasing efficiency, effectively shifting parts of human work toward a more commoditized layer of execution.

We have seen similar inflection points before. Technology repeatedly resets the structure of industries, creating new opportunities even as it makes some roles redundant. As AI becomes more embedded in the tools companies build and use, it will both eliminate certain functions currently done by humans at scale.

There is an analogy here from a political ad. Recall the Obama campaign ad often referred to as “The Stage”, one of the more pointed political ads during the U.S. presidential race against Mitt Romney. In that narrative, workers assembled a stage for a town hall meeting, only for the candidate to step onto the very platform they built and announce layoffs. The symbolism was powerful: the system people help construct can sometimes render them expendable.

Artificial intelligence carries a similar paradox. As companies invest in building smarter tools, those tools increase productivity, but they also reduce the need for certain roles. The same engineers, analysts, and operators who train and refine AI systems may eventually see parts of their functions automated and phased out by AI.

The reality is not malicious; it is structural. Technology improves, efficiency rises, and organizations recalibrate. The lesson is not fear but awareness. As AI advances, the key is to continuously reposition oneself.

A revised worst-case forecast from UBS is intensifying scrutiny of the fast-growing private credit market, with strategists warning that stress tied to AI disruption in software could cascade beyond direct lenders and into broader bond markets.

Earlier this month, UBS’s credit team outlined a tail-risk scenario that included a spike in private credit defaults. Two weeks later, the bank returned with a more negative update, citing mounting sector-specific pressure — particularly in software — and the potential for knock-on effects if valuations fall sharply.

The timing suggests a structural shift is underway in capital markets. As generative AI tools challenge traditional enterprise software models, investors are reassessing revenue durability across SaaS and tech-enabled service firms. Given private credit’s heavy exposure to sponsor-backed software companies, the implications extend well beyond equity markets.

Private credit has grown rapidly over the past decade, filling a lending gap as banks retreated from riskier corporate loans after post-2008 regulatory tightening. Direct lenders stepped in, often financing private-equity-backed companies with floating-rate loans and relatively light covenants.

According to an analysis from Bloomberg, roughly 40% of all private-equity-backed loans are tied to software businesses. That concentration makes the sector especially sensitive to shifts in revenue expectations or competitive disruption from AI-native platforms.

Many of these borrowers were underwritten during periods of high valuation multiples and cheap capital. With interest rates elevated, debt servicing costs have risen. If AI compresses pricing power or slows growth for legacy software providers, cash flow coverage ratios could deteriorate quickly.

UBS warned that in a downturn marked by higher defaults and collapsing valuations, capital adequacy and loss-absorption capacity could come under strain. Unlike public credit markets, where price discovery is continuous, private credit valuations are typically model-based and updated periodically, potentially delaying recognition of losses.

Liquidity mismatch and systemic spillover

A central risk lies in structural liquidity mismatches. Some private credit funds — particularly those marketed to retail investors — offer periodic redemption windows while holding illiquid loans that cannot be quickly sold without steep discounts.

Last week, alternative asset manager Blue Owl said it would restrict withdrawals from one of its retail-focused private credit funds. Former PIMCO CEO Mohamed El-Erian described the move as a potential “canary-in-the-coal-mine” moment, comparing it to early warning signals preceding the global financial crisis.

If redemptions accelerate in stressed conditions, funds may be forced to gate withdrawals or reprice assets more aggressively. That, in turn, could tighten credit availability for mid-market borrowers reliant on direct lending.

UBS’s scenario also highlights interconnectedness. Private equity sponsors, direct lenders, and structured credit vehicles often share overlapping exposure to the same portfolio companies. Stress in one layer of capital structure can transmit through others, especially where leverage is layered.

The risks entered the mainstream last September with the bankruptcy of First Brands, an automotive-parts supplier backed by private equity. Post-bankruptcy reviews revealed aggressive leverage and raised questions about underwriting standards and covenant flexibility in private credit deals.

Industry leaders argued that exposure was limited and that the case was idiosyncratic. Still, the episode challenged the narrative that private credit portfolios are uniformly insulated by conservative structuring.

AI as accelerant

Artificial intelligence is not the sole cause of stress, but it is acting as an accelerant. Software stocks have faced volatility as investors debate whether generative AI will augment or displace incumbent products. If enterprise customers migrate toward AI-native tools or demand pricing concessions, leveraged software firms may face margin compression.

That risk compounds an already fragile setup: elevated rates, thinner covenant protections, and heavy sponsor-driven leverage. In such an environment, even modest revenue misses can trigger refinancing challenges.

Even as caution grows, private credit continues to expand. Asset managers are launching interval funds and other vehicles to broaden retail access. Policymakers have debated whether private credit could be included more widely in retirement accounts such as 401(k) plans.

Greater retail participation raises regulatory and reputational stakes. Retail investors may be less familiar with liquidity constraints and valuation opacity inherent in private lending structures.

From a macroeconomic perspective, widespread private credit stress could dampen capital formation in the mid-market segment, where direct lenders are dominant. A pullback in lending would affect private-equity dealmaking, hiring, and expansion plans.

Unlike the pre-2008 mortgage market, private credit is less directly linked to household balance sheets. However, its integration with institutional portfolios — including pensions and insurance companies — means losses could influence asset allocation decisions and risk appetite more broadly.

However, there is a constructive counterargument: heightened awareness may limit excess. As warnings multiply, lenders are under pressure to tighten underwriting standards, reduce leverage multiples, and demand stronger covenants.

The unresolved question is legacy exposure. Loans originated at peak valuations, and optimistic growth assumptions remain on balance sheets. If AI-driven disruption reshapes the economics of software more quickly than expected, some of that debt may prove difficult to refinance.