Thoma Bravo's 2021 Medallia deal has ended with creditors taking control, and the useful lesson isn't that software suddenly stopped working. It's that expensive capital turns brutal when growth slows and rates stop cooperating.
The Medallia handover landed on June 17, 2026, and the numbers are ugly. Thoma Bravo took the San Francisco customer experience software company private in 2021 for $6.4 billion. Now a lender group led by Blackstone, Apollo and KKR is taking ownership, while Thoma Bravo loses about $5 billion of equity, according to the Financial Times. That is not a soft miss. It is one of the largest private equity losses on record.
The restructuring brings in $150 million of fresh capital and is meant to cut Medallia's debt to a level the business can carry. Mark Bishof, Medallia's chief executive, is pitching the new balance sheet as room to invest in AI products and the company's broader transformation. Fine. A cleaner balance sheet helps. But it doesn't change what just happened: a software buyout priced for one market had to be rescued in another.
This is the part founders should pay attention to, because the collapse wasn't driven by a dead product. Medallia still sells software used by large companies to collect customer feedback and manage customer experience programs. Barron's reported in March that Blackstone provided $1.5 billion of the original $1.8 billion loan financing, and that FS KKR Capital held about $230 million of the debt. The problem was the structure sitting on top of the company. As floating rates rose and performance lagged, lenders marked down their loans and Thoma Bravo had little equity left to defend.
Orlando Bravo had already conceded the basic point. In a March 2026 CNBC interview cited by The Wall Street Journal, he said Medallia was a fine company but that Thoma Bravo had overestimated its growth rate and paid too much. You don't often get a cleaner diagnosis from the buyer itself. The price was too high, the debt was too heavy, and the market that made the deal look sensible in 2021 didn't last.
There is a lazy way to tell this story as a morality play about private equity greed. Don't bother. Leverage is not automatically reckless, and plenty of software companies have used private equity ownership to tighten operations, buy competitors and grow faster. The sharper point is narrower: if your return depends on cheap debt, high software multiples and uninterrupted growth all arriving together, you haven't bought margin for error. You've bought a three-part bet.
Founders know this better than anyone, even if they don't always say it when the offer price is large enough. In 2021, high valuations made many deals feel like validation. A sponsor could tell you that the debt was manageable, the market was still expanding, and the next product cycle would carry the business through. Then rates moved, public SaaS multiples compressed, AI changed buyer expectations, and procurement teams got much more careful about vendors with distressed balance sheets. That is a lot to ask one capital structure to absorb.
The AI pitch now has to earn belief
Medallia's next story is AI, because almost every enterprise software company now needs one. The company has been talking about a move beyond traditional experience management into prediction, automation and faster analysis of customer signals. In March 2025, Medallia announced seven AI-powered capabilities, part of the same repositioning Bishof is now trying to push with new ownership behind him.
Frankly, the timing is hard. Qualtrics and other customer experience rivals have spent the past two years putting AI features in front of the same enterprise buyers. Medallia is trying to tell customers that its product roadmap is moving forward just as those customers are reading about a debt-for-equity swap. Enterprise buyers notice that. Before they sign a multiyear contract, they check whether the vendor will still be stable enough to support the work.
The optimistic version is simple enough: creditors as owners may be better than creditors waiting for a missed payment. Blackstone, Apollo and KKR now have reason to back the product instead of just protect their loans. The $150 million capital infusion gives Medallia more room than it had while the old debt burden hung over the business. Blackstone's Brad Marshall told the Financial Times that Medallia is profitable and serves many of the world's largest companies. That is the argument Medallia will take to customers.
But the product has to do the work. AI language is cheap in enterprise software, and buyers have become quicker at spotting the difference between a real workflow improvement and a feature label added to an old dashboard. If Medallia can use AI to shorten analysis, automate responses and make customer feedback more useful inside large organizations, the restructuring may give it a second chance. If not, the company will simply have swapped one owner group for another while the market moved on.
The wider warning is still sitting in the background. The FT reported that Thoma Bravo's loss is second only to the TXU collapse in private equity history, according to Daniel Rasmussen of Verdad Advisers. Barron's had already flagged Medallia as one of the higher-profile problems in private credit. You should read that as a signal, not an isolated accident. The 2021 software buyout class was priced in a market where debt was cheap and growth seemed more durable than it was. Medallia is the loud example because the number is so large.
For founders, the practical test is plain. When someone offers capital, ask what happens if rates rise, if growth misses plan, and if the next buyer values revenue at half the old multiple. If the answer is still survivable, you may have a deal. If the answer depends on everything going right, Medallia has just shown you the bill.
Also read: India's largest stock exchange files for an IPO that is really a payout for its existing owners • Nuclear power is becoming the defining infrastructure bet behind the AI buildout • Allbirds sells its sneakers for $39 million and bets what's left on AI infrastructure
