The OG of Private Credit: The Dot Calm

For over a decade, software has been the darling of investors. The reasons were evident long before the pandemic made the sector impossible to ignore. Recurring revenue, high retention rates, and sticky enterprise contracts offered lenders a business model that was both predictable and perceived to be resilient through any economic downturn. The pandemic accelerated that dynamic.

The forces driving the boom were not solely a product of post-crisis regulation. Mega-equity funds raised capital at an unprecedented pace, competed aggressively for software assets. Direct lenders matched this fundraising sprint, armed by robust retail inflows. They offered sponsors speed, scale, and certainty of execution needed to close large deals. 

Cheap credit and a near-zero rate environment supported sponsors in paying ever-higher valuations as digital transformation drove growth rates to historic highs. As multiples expanded, software became the dominant investment theme across the asset class. The largest credit platforms launched dedicated tech funds to capitalize on the opportunity.

Multiple disruptions followed, none expected or fully stress tested. Covid’s downturn was short-lived, but we are still living through its aftereffects. Fed rate hikes pushed us into a higher-for-longer interest rate environment and exposed a fundamental weakness in the software thesis.

The flaw was that many high-growth software firms operated with an annual recurring revenue (ARR) model. This assumed such confidence in revenue growth that cash flows could be fully dedicated to growth. The PIK (payment-in-kind) toggle provided room for this, allowing more cash interest to be accrued. Theoretically, if growth slowed more cash could be diverted to debt payments.

The problem now is that if AI disruption causes a more dramatic revenue slide, even a lower spend environment won’t be enough to stave off defaults. Today’s AI “crisis” could be the 2026 version of the dot-com bubble. The prospect of low-cost AI agents replacing entire categories of SaaS products forced lenders to take a hard look at existing portfolios. The question is less whether a company generates recurring revenue, but whether a non-mission-critical business will generate any revenue at all.

Bad loans emerge in every credit cycle. But the more important distinction is that stress in software credit is largely an upper middle market phenomenon. The large check sizes and concentrated software exposure now being scrutinized with BDC valuations has not translated to the same degree in the core middle market (see Chart of the Week).

The lesson is not that software is broken. Indeed, some would argue today’s best performers are in the software space. But what will tomorrow bring? The crush of retail cash in-flows can erode manager discipline, and out-flows can slash manager capacity. 

The good news is that AI risk is dispersed in core middle market portfolios among prosaic businesses in low-tech industries. Add to that better covenant protections, more conservative leverage, and fewer ARR loans, and the AI narrative softens. Experienced managers who never chased the story are the ones who never have to explain it.