AI Just Repriced the Loan Book

AI wiped over $1 trillion off software stock valuations in the first week of February. Salesforce, ServiceNow, and Adobe each dropped 25-30% year to date. The iShares Software ETF fell 20%. Bloomberg called it the "SaaSpocalypse."

That was the equity story. Most founders noticed it.

The credit story is quieter, slower, and arguably more consequential for anyone with private market allocations. And it accelerated sharply this week.

From Equity to Credit

When AI repriced software stocks, the logic was straightforward: if AI agents can replace workflows that companies currently pay subscription fees for, the growth assumptions behind SaaS valuations break down. Public markets repriced quickly. Enterprise software price-to-sales ratios compressed from roughly 9x to 6x by mid-February, levels not seen since the mid-2010s.

But software companies don't just issue equity. They borrow. Heavily. Private equity sponsors spent the last five years acquiring software businesses with leveraged buyouts financed through private credit. The thesis was compelling: recurring revenues, fat margins, low churn, predictable cash flows that service debt reliably. Software became private credit's favourite sector.

S&P data shows software and technology accounting for roughly 25% of the private credit market through year-end 2025. UBS estimates 25% to 35% of the broader market is exposed to AI disruption risk. Those loans were originated during a period when no one priced AI displacement into credit models. Five-to-seven-year maturities mean businesses that look insulated today could face competitive threats well before their loans come due.

The question equity investors answered in February — "what are these companies worth?" — is now being asked by credit investors: "can these companies service their debt?"

JPMorgan Just Answered First

This week, JPMorgan marked down the value of software loans held as collateral by private credit funds and began restricting lending against those assets.

That's worth unpacking. Wall Street banks don't just compete with private credit — they finance it. Banks lend money to private credit funds using the funds' loan portfolios as collateral, a practice known as back-leverage. According to a Moody's Ratings report based on Federal Reserve data, Wall Street lenders had provided roughly $300 billion in such financing as of mid-2025. JPMorgan alone had $22.2 billion of exposure.

When JPMorgan marks down collateral, it directly reduces the amount of private credit funds that can borrow. In some cases, funds may need to post additional collateral. The effect ripples: less leverage means less lending capacity, tighter terms for borrowers, and reduced liquidity buffers for funds already facing elevated redemptions. A source close to the bank told CNBC the move was about financial discipline — acting on market valuations before actual loan losses force your hand. JPMorgan previously pulled back leverage during the early months of COVID.

Jamie Dimon reinforced the signal at the bank's leveraged finance conference last week, saying JPMorgan was becoming more cautious when lending against software assets. Deutsche Bank separately flagged €26 billion in private credit exposure in its annual report this week.

The banks are moving before the defaults arrive. That's the signal.

The Scale of What's Exposed

UBS analyst Matthew Mish laid out the numbers in a February research note. His baseline scenario: $75 billion to $120 billion in fresh defaults across leveraged loans and private credit by the end of 2026. Those figures assume default rates increase by up to 2.5% for leveraged loans and up to 4% for private credit.

The mechanisms are specific. Software companies acquired by private equity during the 2020-2024 boom were underwritten on the assumption that recurring revenues and high margins would persist indefinitely. AI challenges both. If enterprise customers can automate workflows through AI agents instead of paying for SaaS seats, renewal rates compress. If margins narrow because incumbents need to invest heavily in AI capabilities to stay competitive, debt service becomes harder. Payment-in-kind loans — where borrowers defer interest payments — are concentrated in software. Those structures work while cash flow is growing. They become credit problems when growth stalls.

The real exposure may be larger than headline allocations suggest. Companies classified as "business services," healthcare segments, and financial services firms are often fundamentally software-driven businesses. True technology exposure in private credit portfolios is likely understated by traditional industry classifications.

Goldman Sachs analysts warned that the software sector could follow the pattern of newspapers facing internet disruption: share prices (and by extension, enterprise values) only stabilising after years of earnings decline. If that comparison holds even partially for leveraged software companies, the credit implications extend well beyond 2026.

Where It's Already Showing Up

The AI-through-credit repricing isn't theoretical. It's triggering real liquidity events.

Morgan Stanley, BlackRock, and Cliffwater all hit redemption caps on their largest private credit funds this quarter. Cliffwater's $33 billion interval fund saw 14% redemption requests against a 7% cap — half the investors who wanted out are now in the queue. Morgan Stanley's North Haven fund had nearly 11% of shares tendered against a 5% ceiling. Jefferies data shows that private wealth flows into alternative products fell 19% quarter over quarter.

Boaz Weinstein of Saba Capital, who launched tender offers for Blue Owl's non-traded BDC at a roughly 35% discount to NAV, framed the dynamic bluntly: private credit's problems are "multiplying by the quarter," driven by what he called the "financial alchemy of promising liquidity that isn't there."

Geopolitics added fuel. Oil briefly hit $120 as the U.S.-Iran conflict disrupted shipping through the Strait of Hormuz. The S&P 500 posted its lowest weekly close of 2026. New Section 301 trade investigations targeting 16 partners added another layer of uncertainty. When public markets sell off and volatility spikes, investors pull cash from whatever they can access. For many wealth-channel clients, that was their "semi-liquid" credit fund.

What Founders Specifically Should Watch

This story has two sides for founders, depending on which side of the software economy you're on.

If you built a software company — or still hold equity in one — the credit repricing affects your sector's valuation multiples, acquisition appetite, and refinancing environment. Private equity sponsors who would have been buyers for your business are now managing stressed loan portfolios. Leveraged buyout financing for software companies is tightening. That doesn't mean exits are closed, but the pricing and terms are shifting. Founders pre-exit should factor in a more conservative buyer universe and potentially longer deal timelines.

If you're allocated to private credit — and post-exit wealth management portfolios frequently steer 15-25% into credit strategies — the AI repricing reaches your portfolio whether or not your manager calls it out. The questions worth asking: what's the fund's software and technology concentration, and does that include companies classified under other industry labels? What vintage are the loans, and were they originated before AI disruption was a pricing factor? How does the manager distinguish between software businesses that AI will enhance versus those it will displace?

Both sides share a common thread. The AI disruption that's repricing software credit wasn't in anyone's underwriting model two years ago. Five- to seven-year loans originated in 2021-2023 carry assumptions that may not hold through maturity. The institutions financing those loans are already adjusting. Individual investors, particularly those in wealth-channel vehicles, tend to be the last to see the markets move.

Dispersion Trade

Not everyone reads this as a crisis. iCapital's analysis makes a structural argument: private credit is income-producing, underlying loans average a three-year life, and roughly a third of any portfolio turns over annually. That natural cash generation sits well above the 20% annual redemption ceiling. If credit quality holds, redemption queues should normalise within three to five quarters.

Weinstein himself isn't bearish on the category. He's reportedly buying equity in Ares, Apollo, and Blackstone while simultaneously launching discount tender offers against weaker vehicles. That's the dispersion trade: the best managers with direct origination, diversified loan books, and genuine workout capability will absorb this and emerge with more market share. Managers with concentrated software exposure, layered fund-of-fund structures, and bank-leverage dependency face continued pressure.

McKinsey's latest data confirms the broader pattern. Secondaries volume hit $240 billion in 2025, up 48% year-over-year. GP-led continuation vehicles reached $115 billion. Average buyout pricing declined 200 basis points to 92% of NAV. Quality assets still clear. Everything else discounts.

For founders building or restructuring investment frameworks: this is ultimately an architecture question, not a market-timing question. The founders who handle stress well are the ones who built genuine liquidity separation from the start — cash and liquid securities for 18-24 months of hard commitments, semi-liquid allocations sized only against capital with no near-term claim, and locked-up positions only with truly patient money.

AI disrupted software equity in weeks. The credit repricing will take quarters. But it started, and the institutions are already moving. Founders should understand where they sit in that sequence — as builders, as allocators, or as both.