Lenders Are Calling the Shots on Software Debt — and It's Changing the Risk-Reward for Public Credit

Thoma Bravo has a March 2027 deadline and nothing close to an exit.

The private equity firm that owns cybersecurity company Sophos needs to refinance roughly $2.1 billion in debt. Private credit funds — the same lenders who were competitive for software loans last year — are passing. Thoma Bravo's answer: offer a higher coupon, add amortization payments, tighten the covenants. More concessions. The same pattern they ran through with Proofpoint just months ago, where executives conceded 40 deal sweeteners to keep lenders from walking on a $5 billion refinancing.

This is not a company-by-company story. It is a structural shift in credit markets, and it flows through to publicly traded credit investors who hold similar software debt. The question is whether the market has priced the change or whether it is still catching up.

The maturity wall

The concessions have a mechanical cause. A wave of software and technology debt borrowed during the 2021 buyout frenzy is coming due at once. More than $330 billion in software and tech debt matures by 2028. About $142 billion comes due in 2028 alone — roughly three times what matures in any single earlier year.

The pressure is concentrated on private equity-backed companies. About $40 billion of the 2028 wall is from PE-sponsored software deals first financed in 2021, when interest rates were low and valuations were high. Those companies now need to refinance at a time when AI disruption has undermined investor confidence in parts of the software sector and lenders are unwilling to extend the same forgiving terms.

Thoma Bravo carries about $9 billion of portfolio company debt maturing by the end of 2028 — more than any of its private equity peers. Sophos is the next item on that list. But the firm's track record this cycle tells you what lenders expect: Medallia, acquired for $6.4 billion in 2021, was handed over to a consortium of lenders led by Blackstone, wiping out roughly $5.1 billion in equity. The $5 billion Proofpoint refinancing required a yield near 9.3 percent and a new "omni blocker" provision that prevents the company from moving assets out of lenders' reach.

In each case, the sponsor went in expecting smooth refinancing and came out writing concessions they never planned to offer. Time is no longer free, and when a maturity deadline approaches, the borrower who needs the extension is not the one setting the terms.

What this means for public credit

Software loans are not confined to private balance sheets. Publicly traded business development companies — the vehicles that let retail investors access private credit — hold substantial software exposure. Across the BDC sector, software averaged roughly 17 percent of portfolios in the first quarter of 2026, with some firms running well above that average. Approximately $5 billion in BDC-held software loans came up for refinancing between May 2026 and August 2027.

The sector has been punished for it. BDC stocks traded at an average of 0.83 times book value in early 2026, and software credit concerns triggered repeated selloffs. Twenty-eight of 53 BDCs swung to losses in the first quarter of 2026, up from profitability at the end of 2025. The sell-off was broad and indiscriminate — Blue Owl Capital, which had among the highest software concentrations, fell sharply enough that the company announced it would reduce its software exposure, selling $1.4 billion in direct lending loans at 99.7 percent of par to institutional investors.

But here is where the Thoma Bravo concessions matter for the public market. The same lender power that is forcing Thoma Bravo to offer higher coupons, amortization, and tighter covenants benefits the people on the other side of those deals. BDCs and other public credit funds are not the borrowers. They are the lenders. And when borrowers concede, lenders get better terms.

The higher yields Thoma Bravo is offering on Sophos and Proofpoint are the new floor for software credit pricing. The tighter covenants protect lenders if portfolio companies underperform. The omni blocker prevents asset stripping that leaves creditors holding a hollowed-out shell. These are not headline events — they are contract provisions that change the risk-reward for every dollar of software debt held by public credit funds.

The proof point

Ares Capital (ARCC), the largest publicly traded BDC, illustrates the mechanics. At $19.68 a share, it trades essentially at book value and pays a 9.7 percent dividend yield, supported by 21 consecutive years of dividend payments. Software represents roughly 22 percent of its portfolio, concentrated in second-lien loans — a senior position that gets repaid before equity holders if a company fails.

ARCC reported core earnings of $0.47 per share in the second quarter of 2026, with management describing portfolio quality as solid. An external valuation the firm commissioned showed negligible portfolio-level risk from the software credit selloff. The company's dividend coverage remained intact.

The numbers suggest something the broader BDC selloff may not have fully captured: the biggest, most diversified BDCs have the scale and underwriting discipline to hold software debt while benefiting from the pricing reset. They are not being forced to sell at discount like smaller, concentrated funds. They are the ones lenders like Blackstone, Blue Owl, and HPS approach when they need to place large software loans. The concessions Thoma Bravo is making are being priced into the loans these BDCs hold or are likely to hold next.

The market has not ignored the risk — BDC averages at 0.83 times book reflect real concern. But ARCC trading at book while offering a 9.7 percent yield, with a portfolio that has held up through the selloff, suggests the worst of the repricing may already be done for the quality end of the market.

The risk that stays live

The concessions are a lender's advantage only if the underlying software businesses keep generating enough cash to service their debt. A higher coupon does not fix a broken revenue trajectory. Tighter covenants help when a company is struggling but still viable — they do not resurrect one that isn't.

The Medallia outcome proves the limit of lender protections. Despite all the contractual advantages Blackstone and its partners held, the company was still worth handing over rather than supporting. Lenders with first-lien positions on a failed borrower can still face losses if the remaining assets are insufficient.

For BDC investors, the test is straightforward: watch portfolio-level performance, not headlines. Quarterly core earnings, dividend coverage ratios, and loan loss reserves tell you whether software exposure is translating into losses or into higher, sustainable yields. If those metrics hold steady while yields compress the discount to book, the power shift has worked in the lender's favor. If core earnings decline and coverage falls below 1.0 times, the concessions were not enough to offset deteriorating fundamentals.

The software debt maturity wall is real. The concessions are real. The question for investors is whether the companies on the borrower side can produce the cash flow their new, tighter contracts demand — and whether the public credit funds holding those loans are being properly compensated for finding out.