Dagmara Michalczuk, Creditflux, March 2026

In the early 1940s, economist Joseph Schumpeter articulated his theory of ‘creative destruction’, a dynamic process through which innovations disrupt and sometimes destroy old industries and economic structures, thereby driving long-term growth and improving living standards. He viewed the costs of destruction — and the recessions necessary to weed out inefficient firms and create room for more productive ones — as a core feature of capitalism.

He further characterised this process as one of overlapping innovation cycles of varying duration: long waves driven by major technological changes, medium waves shaped by fixed capital investment, and short waves tied to inventory fluctuations.

Many economists argue we are now in the sixth long wave, which is defined by the convergence of new digital, biological and physical technologies. These include generative artificial intelligence, machine learning, genomics, renewable energy, robotics, quantum computing and space systems.

For investors, all options are risky

Financial markets are the engine of every innovation cycle. They allocate capital to fund transformations and face the consequences. As in prior regime shifts, this wave will produce winners and losers. Backing disruptive challengers is inherently risky, particularly early in the cycle. Yet financing incumbents does not guarantee safety, as dominant firms can see profits erode or be rendered obsolete if they fail to adapt.

The challenge for investors is therefore twofold: identify the likely winners and losers — and assess the timing and magnitude of change.

This is where AI may prove different. Unlike prior technological transitions, AI has the potential to be both faster and broader in its disruption. The fifth innovation wave, driven by the internet and advances in communications, was constrained by physical infrastructure build-out, behavioural change and the evolution of digital payment systems. AI, by contrast, taps into that existing infrastructure. While development requires significant investment in data centres and energy, adoption at the end-user level faces fewer barriers.

AI tools can plug directly into established workflows and software ecosystems, allowing rapid scaling. As a result, disruption timelines may compress relative to previous cycles.

The breadth of impact may also exceed earlier waves. AI is not merely a new channel of distribution. It also alters how products and services are designed and produced. It applies across sectors, not just a handful of industries. Once trained, models can improve through iteration and data at relatively low marginal cost, accelerating diffusion. Crucially, AI also reshapes decision making, increasing speed, reducing costs and potentially shifting competitive dynamics across business models.

Therefore, investing through the AI-driven transition requires a new approach. Industry classifications alone are likely insufficient. Understanding AI sensitivity demands a deeper examination of business models, cost structures, advantages and adaptability.

For CLO investors, this has important implications. One key risk is that AI becomes a catalyst for increased credit migration and default correlations across industries. CLO portfolios are diversified across sectors and issuers, but not by business model. If AI simultaneously pressures similar value propositions, earnings compression could emerge across sectors. As a result, traditional sector labels and diversification measures may mask AI vulnerability.

Struggling companies could be boosted

At the same time, AI is not only a threat. Operational efficiencies may expand margins for many issuers. Cost savings in areas such as customer service, underwriting, logistics and coding could strengthen credit profiles, even in sectors currently perceived as structurally challenged, such as software.

Recent “whack-a-mole” market reactions, with rapid sell-offs moving from software to insurance and beyond, suggest investors may still be relying on blunt sector proxies rather than fundamental analysis of AI exposure. In an environment where disruption cuts across industries, passive categorisation risks missing both emerging vulnerabilities and hidden resilience. In a world where business-model sensitivity matters more than sector labels, alpha will be generated by CLO managers willing to look deep under the hood.