Actor

Thoma Bravo

Thoma Bravo is a private equity firm that buys software companies and only software companies, running more than $172 billion of assets across roughly 80 portfolio companies as of 31 March 2026. Its model is to take a subscription business private using debt sized against recurring revenue rather than profit, cut costs, bolt on rivals, and refinance when the loan comes due. That last step is the one that has started failing: creditors took control of Medallia in June 2026, a $5bn Proofpoint refinancing needed sweetening in July after a lender revolt, and in August a $2bn Sophos refinancing could not find private credit backing at all. If you want a single name to watch for whether the software debt boom unwinds quietly or badly, this is it.

The software buyout machine

Thoma Bravo reported more than $172 billion in assets under management as of 31 March 2026, across about 80 portfolio companies. (A separate 2026 industry ranking puts the figure nearer $184 billion for the same date; the firm's own number is the conservative one, and it is the one we use.)

It arrived at software-only by subtraction. The firm began as Golder Thoma & Co, became Thoma Cressey Equity Partners, then Thoma Cressey Bravo, and finally Thoma Bravo after Bryan Cressey left. Orlando Bravo's account is that the software deals kept working better than everything else, so the partners spun healthcare into a separate group and, in his words, decided to "go forward and do software only." The firm dates its first software-only vehicle to Flagship Fund IX in 2008. Seventeen years later that specialisation is the largest of its kind.

The capital kept arriving right through the period when everyone else's did not. In June 2025 the firm completed a $34.4 billion fundraise, of which Fund XVI accounted for $24.3 billion. Fund XVI and the Europe Fund were oversubscribed and hit their hard caps; Discover V was oversubscribed by more than 30%. Nor has the buying stopped: in August 2026 Thoma Bravo agreed to take Accelerant private for more than $4 billion, and Bloomberg described the firm the same month as keeping up an "intense pace of dealmaking." So the story here is not a firm in retreat. It is a firm with an enormous new fund and an older portfolio whose financing arrangements are starting to argue back.

The ARR lending engine

The 2021–2022 vintage is where the model was pushed hardest. Five deals — Proofpoint at $12.3bn (announced April 2021), Anaplan at $10.7bn (March 2022), Coupa at $8.0bn (December 2022), SailPoint at $6.9bn (April 2022) and Medallia at $6.4bn (July 2021) — come to $44.3 billion of enterprise value across a two-year window. Thoma Bravo was not an outlier in this; it was the purest expression of it. Bain counted $284 billion of technology deals closed by private equity in 2021, 25% of total buyout value and 31% of deal count.

What made those cheques possible was a change in how lenders decide what a company can borrow. Traditional leveraged finance sizes debt against EBITDA, which is awkward when the borrower has no EBITDA. The recurring-revenue loan sizes it against annual recurring revenue instead, on the theory that a contracted subscription base with high retention and high gross margins is a more reliable source of repayment than a current profit number. Market descriptions put ARR facilities commonly at around 1.5x to 2.5x ARR, with some software loans described at 2x to 3x depending on structure. In place of an earnings test, the package ties to recurring revenue, liquidity, and profitability by a fixed date.

The equity return then comes from two levers. Cost cuts: because software gross margins are high, stripping out sales, general overhead and duplicated R&D converts revenue into EBITDA fast. And add-on roll-ups: buy smaller adjacent products, merge them into the platform, and the combined recurring revenue base supports still more debt while overhead is shared. The lender's downside protection and the 's upside case are the same asset — the renewal rate. Everything works if customers keep renewing. That is not a criticism of the structure. It is a description of it, and of what happens if the assumption fails.

Why it matters to this crash

The load-bearing assumption of a firm this size is not that any one company performs. It is that when the debt matures, somebody will refinance it. Thoma Bravo has run the world's largest test of that assumption, and in mid-2026 the answer came back qualified three times in three months.

In June 2026, Blackstone-led creditors took control of Medallia — a company bought for $6.4bn in 2021. In July, Thoma Bravo ceded to a lender revolt on a $5 billion Proofpoint loan, sweetening terms to get the refinancing done; Proofpoint, the $12.3bn deal, had already been reported as handed to lenders in 2025 amid a restructuring. Then in August, Sophos — bought in 2020 — spent months trying to place more than $2bn of loans with private credit funds, and the effort faltered. The company went back to its existing leveraged-loan holders offering a higher coupon, amortisation payments and a tighter . We wrote about that on 19 August, and again twice more the same day, because the direction of travel is backwards: private credit exists to be the buyer of last resort for paper the will not take, and here the paper went the other way.

Two details make this more than a bad month. First, Thoma Bravo told investors it would not inject fresh capital into Sophos, rebuffing a request from lenders. The sponsor's equity is the cushion under the debt, and nothing obliges a sponsor to top it up; when it declines, the loss stays with creditors and gets settled through terms rather than defaults. Nothing appears in a credit index. Second, the stated lender worry is AI — whether a security software product survives a technology shift that outlasts the loan.

Orlando Bravo has argued in public that AI valuations are in a bubble while enterprise software is not, and separately that some software names hit by AI deserve a valuation cut. His lenders appear to be applying the second view to his own portfolio. Set that against the median at the twenty largest listed reaching 2.8% of cost in Q2 2026, the highest since 2017, and the pattern is of repricing rather than rupture — so far.

The other side of the ledger is real. SailPoint, bought for $6.9bn in 2022, was relisted in February 2025 at a $12.8bn valuation, and Reuters reported in 2026 that Anaplan may be preparing an IPO. The playbook still produces winners. The question is what the losers cost, and who pays.

What would make this dangerous

The honest starting point is what nobody outside the firm knows: there is no public figure for aggregate debt across Thoma Bravo's roughly 80 portfolio companies, and no public maturity schedule. That absence is itself the risk — a maturity wall you cannot see is one you cannot price.

Five things would turn this from repricing into a problem. A fourth and fifth portfolio company going through the Sophos process in the same quarter, which would establish that the private credit bid for ARR paper is absent rather than selective. Another creditor takeover on the Medallia model at a materially larger company. A refinancing that fails outright rather than closing on worse terms, because worse terms are lenders working and no terms is a market clearing at zero. Thoma Bravo declining fresh equity a second and third time publicly, which teaches every other sponsor's lenders to demand amortisation up front. And lenders cutting the standard below the 1.5x-to-2.5x ARR range, which would reprice every unrealised software in the industry at once, not just this firm's.

Then the exit side. Bain's 2026 midyear work notes technology buyout deal value has fallen off sharply. If the Anaplan listing does not happen, or happens below cost, the model loses the one thing that makes heavy survivable: a bid at the end for the thing you levered.

As seen in

Every dispatch we have filed that touches this. Newest first.

  • Nobody wants to refinance Sophos

    When the buyer of last resort for LBO debt steps back and the sponsor refuses to add equity, losses get resolved through terms rather than defaults — which is why nothing shows up in spreads until it shows up all at once.

  • Private credit said no to Sophos

    The assumption that private credit will always refinance a sponsor's portfolio company is the load-bearing assumption of the whole asset class, and it just failed a $2bn test.

  • The lender of last resort said no

    Private credit's function in the system is to absorb debt nobody else will hold; when it starts declining, the marginal borrower has nowhere left to go.

Written 2026-08-21. Crashopedia entries are drafted from sourced evidence, fact-checked against it, and edited by hand. If something here is wrong, it is wrong in git and can be fixed there.