What this covers
The conversation traces Thoma Bravo from its origins to its current underwriting habits, with A.J. doing most of the explaining alongside Rick and Paul. It starts in 1980, when Stan Golder and Carl Thoma left a First Chicago venture group and, in 1982, pioneered buy-and-build consolidation. The PageNet spectrum-license story shows how little equity such a deal could need: roughly $6 to $7 million became about $985 million. Thoma's Don't Lose Money Heuristic carried into software, where Marcel Bernard brought Six Sigma discipline and the Profit 21 playbook split companies into R&D and sales centers.
The middle stretch is mechanics. Underwriting rests on gross retention plus consumption expansion plus CPI-level inflation. Customer connectivity has moved into the top three diligence priorities. Holding periods have lengthened from about 3.5 years to a Year-10 Underwriting horizon, which changes what is asked of CEOs. A.J. also covers headcount savings of 10% to 20%, a single annual revenue-and-EBITDA target, and why large, institutionalized buyout firms with layered committees decide more slowly.
The contested material is the market call. A.J. reads the SaaSpocalypse as a Duration Trade: Guidewire's growth assumption falls from 12% for 20 years to negative by years 8 to 11, which he estimates cuts valuation about 30%. He calls software's discount to the S&P median, despite roughly twice the growth and 80% gross margins, absurd. His case against "vibe-coded" replacements rests on Downstream Interdependency, the workflows and decisions around a system of record, while AI-native tools such as Harvey win at the edges. The talk closes on culture, including philanthropy, music and racket sports as pattern-recognition inputs.
A.J. Rohde argues Thoma Bravo's durable edge comes from a fixed playbook of backing existing management, operational discipline, and deep customer connectivity, combined with a cohesive, low-ego team culture. He contends that AI and the SaaS selloff have not made mission-critical, customer-connected software obsolete, and that the market has mispriced the sector.
- Thoma Bravo has kept the same core model since the 2008 Fund 9 split: back existing management, teach operational discipline, and avoid formal hierarchical approvals.
- Incumbent software embedded in enterprise workflows has downstream interdependencies that vibe-coded or AI-native challengers underestimate.
- Public markets priced software as if growth decays to zero by year 10-20, which A.J. considers a mispricing that should revert.
Customer trust and weekly operating data, not premium AI spending or hype, should drive software competitiveness.
- Public enterprise software has run EBITDA margins near zero for about 20 years, and stock-based compensation abuse was a bad behavior the public markets punished, though Thoma Bravo's companies do not follow it.
“Why is the industry for 20 years running at zero margin? That's absurd.”
- Thoma Bravo does not take AI-native challengers for granted and responds by offering the same capabilities through its own teams, layering AI talent in a capital-efficient bootstrap model rather than paying premium salaries for frontier talent.
“We don't take the AI risk on the AI-native apps for granted. We say we can offer the same thing.”
- There is growing populist sentiment against AI companies that study their customers' businesses in order to become their customers, which damages trust, and Thoma Bravo's AI ambition is to help customers be more efficient rather than to replace them.
“there's some companies that want to study their customers and they want to become their customers. That does not make people feel good.”
- Over time, AI talent will become more pervasive and commoditized, creating labor deflation with tiers of skill that companies can match to the salary a customer's need requires.
“I'm more optimistic that there'll be labor deflation in that as that skill becomes more pervasive and more of a commodity”
- Many AI-native and SaaS startups fail because they do not appreciate capital efficiency, which was the downfall of many SaaS companies in 2021.
“'Cause a lot of these startup companies, they just don't. And that was the downfall of a lot of SaaS companies in 2021.”
- Portfolio companies should build AI capability by retraining existing engineers or hiring AI overlays within a thoughtful bootstrap model, rather than paying millions to frontier AI talent for capabilities customers do not yet need.
“it doesn't mean necessarily going and paying some 24-year-old kid a $5 million guarantee because they know language models”
- Software companies took their excellent business model for granted and lost connectivity with customers, not knowing how customers used the product, how their org designs were evolving, or what adjacent tech was doing.
“Software companies for years, I think, took advantage of the fact that the business model was so good, and therefore lost connectivity with their customers”
- Most people go through an emotional oscillation about AI between optimism, pessimism, and doom, and Thoma Bravo counters this by anchoring decisions in weekly operational data such as pipeline, retention, and customer feedback.
“every person has gone through this sine function of optimism, pessimism, doom and gloom.”
Backing few long-horizon leaders through long holds, not quick sales, defines Thoma Bravo's model.
- Thoma Bravo historically sold companies after about 3.5 years, often because buyers came calling, but now underwrites a year-10 to year-15 horizon because buyers are less frequent and arbitrage has compressed, requiring CEOs who want to stay ten years.
“Historically, we used to hold for 3.5 years. That wasn't by design.”
- Thoma Bravo's model is to back existing management and teach them to be better operators, whereas the early 1980s buy-and-build firms found an industry executive first and gave them capital to aggregate a fragmented market.
“We back existing management and teach them how to be better operators.”
- A rare attribute to underwrite is a management team that has stayed together since COVID and hit its numbers every quarter through the buyout spree, the hangover, GPT-3, geopolitics, and tariffs.
“has that team been together since the beginning of COVID and have they hit their numbers every quarter since that time?”
- Thoma Bravo's operating plans center on one simple annual target combining revenue and EBITDA growth, which management teams must be excited to hit, and which screens for management quality.
“And it usually manifests itself in one number every year. That number becomes written in Sanskrit on the walls of the company”
- Long holds require finding CEOs and teams who see they have a tiger by the tail and want to stay together for ten years, a trait that is rarer now because tech talent changes jobs frequently.
“That does require a little bit more screening of the leadership, the CEO and his or her team, that they appreciate that they have a tiger by the tail”
- Thoma Bravo appeals to ambitious founders who want to take their business up three levels and then make material acquisitions, which is how the firm built its franchise.
“we tend to appeal to the ambitious business builder type who wants to take their game up three levels, not one level”
- Holding companies for seven to ten years requires finding leaders who view the tiger-by-the-tail opportunity as a long-term commitment, because a buyer in year seven will care about the year-ten trajectory.
“Someone who's going to buy from us is going to care about year 10 because they're going to buy from us in year 7.”
- Commentators erred by lumping cybersecurity, developer tools, and sticky applications into one software category, when sticky, deeply embedded applications face different disruption risk from other software.
“Picking on software, first of all, it was all lumped together, which is ridiculous.”
- Enterprise software systems are embedded in downstream workflows used by people who do not hold a license, so vibe-coding a narrow function like a CRM database misses the many processes that depend on its decisions, and CIOs therefore replace it only in pieces.
“there's this shadow economy in the customers that is a derivative effect of the system that is put in place.”
- Thoma Bravo's filter for investment ideas is wide, but the pool of leaders who will sustain a 7-to-10-year effort through COVID, GPT-3, and other disruptions is small, so the firm needs to buy only about three companies per fund per year and must be increasingly judicious.
“We have to buy three a year in each fund. That's it.”
Customer-close incumbents resist AI challengers because core systems and downstream workflows are hard to replace.
- Customer connectivity has moved from one important factor to the top three priorities in Thoma Bravo's diligence, because SaaS switching costs have fallen and AI makes customers rethink their architecture.
“That always used to be important. Now it's number one, two, and three.”
- Thoma Bravo's core underwriting metric is gross retention plus consumption expansion from seats, servers, or payments plus CPI-level inflation, which is required for any acquisition and has not changed over time.
“gross plus consumption expansion that naturally happens in our customer base, whether it's seats or servers or payments or whatever measure of the business model, plus a little bit of CPI inflation. That is an inviolable underwriting metric”
- AI has made customer treatment binary for software incumbents: companies either keep customer connectivity and benefit, or lose it and risk losing their franchise, whereas previously the outcome was a sliding scale.
“And it's much more binary than it used to be. It used to be a bit of a sliding scale there. Now I think it's existential in how you treat the customers.”
- Incumbent software companies that have taken customer relationships for granted are rightfully on a path to obsolescence, while those that stay obsessively close to customers and deliver what they want are well positioned in an AI environment.
“They're rightfully on their path to obsolescence. Absolutely.”
- AI-native applications like Harvey are succeeding at the edge of enterprise systems, while core systems of record are harder to attack because customers will not rewrite the core without knowing downstream effects.
“if you see the AI-native successful applications, they tend to be on the edge of the system. Like Harvey in legal.”
- Declining inference costs and open-source model price competition let incumbent software firms lower cost of goods sold, route queries to the right model, and maintain their execution advantages.
“the incumbents have this beautiful opportunity to leverage decreasingly low-cost token.”
- Critics who say AI can vibe-code a replacement for enterprise software such as Salesforce focus on the core database and miss the downstream workflows, decisions, and shadow processes that depend on the system, so ripping out incumbents is harder than it appears.
“I don't think people realize the downstream processes that people who don't have a Salesforce seat license do because of the information that came out of that database”
- Under perpetual licenses, customers who had built infrastructure around systems like Oracle or SAP were captive and could not switch, but SaaS lowered switching costs and AI-based architectures may be the next step in that shift.
“They were captive. It's like Oracle or SAP. You can never get off.”
- A.J. expects some of the AI capital bubble to release and flow back into application software, and believes private valuations have normalized relative to public markets since early 2026.
“I think some of a release of the AI capital bubble will flow back into application software.”
Good people and small-group decisions beat institutional process, preserving small-team culture as the firm grows.
- Thoma Bravo selects well-intentioned, capable, creative leaders with an honest character, which explains the firm's concentration of Midwestern companies, and it believes it picks people better than it picks stocks.
“I actually think we pick the people in charge of those enterprises really well.”
- Many large private equity firms have become institutionalized, with multi-layer investment committees and LP allocation targets that slow decisions, whereas Thoma Bravo can decide quickly with a small group using gut instinct and data.
“There is not any formal hierarchical decision-making approval acceptance. There's a lot of data and a lot of gut instinct on people”
- Thoma Bravo's core investment committee is a small group, roughly seven partners, that can make decisions in ten minutes when prepared, complementing one another's blind spots without designing it that way.
“We have seven of us that run the investment committee, and I would say if you looked at it, everybody complements someone else's blind spot.”
- Thoma Bravo's Discover funds, which A.J. led and grew over about 12 years, are a middle-market, vertical-focused extension of the main fund that keeps the firm scrappy and creative while maintaining five-year mentor-mentee bonds.
“That was sort of the Discover Fund. It was the baby sister fund to the main fund.”
- A.J. says people wired with intense competitiveness tend to bond and do deals together, and his wife only understood his tennis-like intensity after watching the Finance Cup, which is evidence that the community is cut from the same cloth.
“everyone's cut from the same intensity cloth.”
- Orlando Bravo's tennis background attracted other competitive tennis players to Thoma Bravo because people wired to compete intensely tend to gravitate toward one another, and the firm later found systematic patterns linking tennis backgrounds to strong investment results.
“And so when you're wired that way, you tend to want to be around other people who are equally as crazy and wired the same way.”
- A.J. describes himself as the firm's family-style little brother who adds to the system by pushing the team and keeping it deep in verticals.
“My role, I think there was an article recently that called me Mr. Fix-It Guy.”
- A.J. makes sure junior partners join every podcast and deal discussion and gives them real decision-making agency, believing that supported risk-taking in the trenches is how the culture transfers.
“I think it's important that you give young people a lot of agency to make those decisions as long as you're there with 'em in the trenches”
- Successful investing depends on an interdependent infrastructure of partners, and a seven-person investment committee works because each member covers another's blind spot; this composition was not designed but emerged, and it persists when egos are checked at the door.
“everybody complements someone else's blind spot. And that was not by design.”
- Thoma Bravo's investment committee and operating culture have stayed structurally similar since 2010 and even 2003, with a small team able to make decisions in ten minutes, a shared obsession with the business, and a single wolfpack identity that has persisted across funds.
“There is an obsession, OCD about the business and driving it forward that there always has been.”
- A.J.'s role as the 'little brother' fund leader is to keep Thoma Bravo middle-market, scrappy, humble, and deep in verticals, and to resist evolution away from the 2010 operating model, with Discover Fund as a sibling fund to the flagship.
“So I always think about it like my job is to keep Thoma Bravo the way Thoma Bravo was in 2010.”
Buy-and-build roll-ups and value-investing discipline shaped the firm's origins and software strategy.
- Thoma Bravo began as the software team inside Thoma Cressey, which split in 2008 over investment strategy differences, with the first dedicated fund traced as Fund 9.
“So Thoma Bravo was created as the software team inside Thoma Cressey. That was 2008. We call that Fund 9”
- Carl Thoma brought a value-investing mindset rooted in Ben Graham and Warren Buffett, with the rule of not losing money first, which made good unit economics a precondition even in a venture-style roll-up business.
“Classic Ben Graham, Warren Buffett, disciple — don't lose money, number one. Number two, don't lose money.”
- Golder Thoma invested roughly $6 to $7 million to acquire spectrum licenses after 1984 deregulation and built PageNet, which eventually returned about $985 million, showing how regulatory change can create outsized returns with little equity when banks finance the aggregation.
“I think their capital invested was $6 million or $7 million to buy spectrum, start a company, and the eventual return was $985 million, something like that.”
- Golder Thoma was credited in 1982 as originating the buy-and-build strategy in private equity, consolidating fragmented industries where value could be created by combining many small businesses.
“That firm in 1982 was credited at the time for being the originator of the buy and build strategy in private equity.”
- Stan Golder and Carl Thoma spun out from a First Chicago Bank venture group in 1980 to form Golder Thoma, which in 1982 pioneered the buy-and-build strategy of consolidating fragmented industries, the precursor to modern roll-ups.
“That firm in 1982 was credited at the time for being the originator of the buy and build strategy in private equity.”
- Thoma Bravo's first software playbook, developed at Profit 21, separated product P&Ls and split companies into an R&D center and a sales center, then used the freed capacity to bootstrap growth and finance M&A.
“Focus on breaking the company into a manufacturing center, which is R&D, and a sales center, which is your distributor.”
- Early buy-and-build roll-ups could aggregate assets like spectrum with very little equity because banks financed the acquisitions, and the firm then sold before cell phones made the asset less valuable.
“you could buy most, continue to aggregate spectrum financed by the banks at that time, so they could do that with very little equity.”
- Orlando Bravo's early investments in IT services and Y2K readiness companies, many of which went to zero, pushed him to look for a better model, which he found by pairing an operator with lean and Six Sigma principles applied to software.
“Those investments, many of them went to zero.”
Headcount cuts and operating discipline lift margins and free capital for acquisitions.
- In a headcount-driven software business, a buyer can remove about 10% of headcount costs upfront without disturbing operations, and up to 20% if done thoughtfully, freeing capacity to fund growth and acquisitions.
“the thing Marcel taught was in a headcount business, you could take out 10% of the headcount costs of a company upfront without disturbing the operations one bit. And you could go up to 20% upfront if you're thoughtful about it.”
- Thoma Bravo's portfolio companies now run at roughly 50% margins while growing about 20%, compared with earlier targets of 25% margin and 10% growth that the firm celebrated.
“Back then we would high-five at 25, and we would high-five at 10% growth. Now it's 20% growth, 50% margin.”
- Marcel Bernard applied lean manufacturing and Six Sigma principles from Motorola to building, shipping, and distributing software, which Orlando Bravo saw as a repeatable operating lever for application software companies.
“So he had taken these lean manufacturing Six Sigma operating principles and started to apply 'em to building and shipping and distributing software.”
- The Profit 21 take-private, in which Thoma Bravo cut costs, shelved speculative projects, and used the freed capacity to fund acquisitions, led to the realization that existing management plus product P&L metrics could bootstrap software companies.
“And the bell goes off at that point. It's like, wow, obviously they're great business models, but there is a process we could run”
Acquisitions and AI transitions should be managed case by case using weekly operating data.
- Post-acquisition integration should be tailored case by case, with decisions about which people, data schemas, and applications to keep, rather than following the old tiered migration playbook, because rigid migrations can kill a competitor's customer base.
“So no more playbooks, none of that stuff.”
- Bolt-on acquisitions should pass the test of standing on their own without distribution or development synergies, since the firm has overintegrated and destroyed a well-running asset before.
“We made that mistake too, several times where we had a great asset that was running well on its own.”
- Managing AI-driven change is best handled by focusing on weekly operating data such as pipeline, retention, and customer feedback rather than the Silicon Valley hype cycle that promised and scared people at the same time.
“I try to look at what's happening week to week, the pipeline, things coming in and out, the retention rate, what customers are saying.”
Outside pursuits and tennis peer networks build sharper, failure-tolerant investors.
- The Finance Cup, started in 2015 by Bill Ackman and Jeff Appel, connects younger tennis players with private equity, hedge fund, and banking leaders, and it has built a long-term mentorship network.
“Bill and Jeff hatched the idea to start to connect some of the younger folks in the New York tennis scene to some of the private equity hedge fund investment banking leaders in New York City.”
- Tennis players tend to be resilient, a little bit wired differently, and their 'play the next ball' mindset carries into investing, producing good results and helping the firm build a pipeline of talent.
“Especially with — play the next ball, play the next ball. You got smacked on that point and you got hit with an overhead, like next ball. And there was a resilience to the investment business that comes from that.”
- Thoma Bravo's culture treats failure as acceptable, so people are not afraid to take risk, while still requiring strict risk management and stock selection discipline.
“failure is totally fine here. We have to be really smart about risk management”
- Partnerships often fracture because partners tire of the job and let outside activities and egos creep in, and Thoma Bravo has avoided this by treating outside pursuits as additive rather than dilutive.
“One of the reasons the partnerships fracture is because people get a little bit tired of the job sometimes and what the job requires.”
- Talented people who come up competing alongside each other, such as a cohort of tennis players or the private equity business, feed on one another and make each other better over time.
“I always find these ecosystems feed on each other.”
- Engaging in philanthropy and activities like music and racket sports makes an investor better at the day job by building empathy, humility, and an understanding of different business models, which feed a decision tree used for pattern recognition.
“Now with philanthropy, it actually makes you a way better investor.”
- The top 100 players in Division I college tennis are basically professional athletes, so the firm must adapt its recruiting to account for players who may pursue pro careers after college.
“frankly, the top 100 guys in Division I college tennis are basically professional athletes.”
- A.J. used drumming and multiple sports as outlets for his ADHD energy, and says the school system's push for instruments helped him discharge energy in a productive way.
“I was just looking for ways to fill my ADHD. Drumming, especially the way I drum, I drum like Keith Moon.”
Software valuations compressed on duration fears and now look undervalued.
- Vertical market application software traded around 1.5 times revenue and 2.5 times maintenance in the early 2000s, when companies sold perpetual licenses and maintenance, and banks lent against hard assets rather than intellectual property.
“which at the time, by the way, was trading at 1.5 times revenue, 2.5 times maintenance.”
- The SaaS selloff reflected a duration trade in which public markets cut long-term growth assumptions, for example taking a company like Guidewire from 12% growth for 20 years to 4%, 2%, 0%, and negative growth by years 8 to 11, which A.J. estimates reduces valuation by about 30%.
“I think what happened in the SaaS apocalypse is people would start to put the growth rate of Guidewire in year 8, year 9, year 10, year 11, down to 4% and 2% and 0% and -2%.”
- Software trades at a discount to the S&P 500 median on a PE basis despite roughly twice the growth rate and 80% gross margins, a gap A.J. considers absurd and expects to revert.
“software on a PE basis trades at a discount to the S&P median for twice the growth rate and 80% gross margins is so absurd that that has to revert itself over time.”
A.J.'s competitive, people-focused temperament, shaped by his upbringing, drives how he bonds with others.
- A.J. started a dad band after a 40th birthday party where he drummed on a pillow, leading to shows and a studio in his house.
“So I get out a pillow and I'm drumming and my buddies are like, you know how to play?”
- A.J. was drawn to the personalities and storytelling of 1980s dealmakers (RJR Nabisco, corporate raiders, Delaware law) and combined that interest with his mother's behavioral psychology background, which made him attentive to people's motivations.
“I was very keenly aware of different people's motivations.”
- Public school music programs that once steered kids toward orchestra and saxophone have shifted toward rock band curricula, which has made band participation pervasive and brought live rock bands back to college parties.
“there's a hard push for rock band. Rock band as a class, rock band as a curriculum track.”
- Early exposure to drumming and band culture gave A.J. an outlet for high energy and a way to relate to people across social groups, which later helped him bond with counterparties and understand different ways of thinking in deal-making.
“It accesses a part of your brain you just don't get to access.”
- A.J. chose Villanova largely for the move to the East Coast, a blue-collar Philadelphia culture, a mid-sized school of about 1,800 students per class, and a strong college tennis program with mentors such as coach Bob Bateman.
“My coach was amazing. Most people, Rick, back then, didn't have great college tennis coaching or cultures.”