Early this decade, I saw a lot of activity with PE firms seeming to rush to invest in the latest tech buzzwords.

This year, PE firm Thoma Bravo reported a 5 billion dollar loss related to this activity. What do you think about this loss and what it will mean for the future of PE and their investments?

https://ca.finance.yahoo.com/news/thoma-bravo-5-1b-medallia-170011858.html

I doubt it means much of anything. Thoma Bravo is a $170B+ PE firm. This sucks for that fund but it happens. This article is from April and recent headlines show similar investments by them across the software space in a variety of industries. It's who they are. The only way something changes in this space is if the banks stop lending or raise rates so high they can't cover the cost and that's unlikely to happen.

The 2021 and 2022 vintage SaaS take-privates are all generally in a bad spot. EV/Sales ratios were 10-15x+, and debt was cheap. Now debt is more expensive, growth has slowed, and multiples have collapsed. A deal that was done at 15x sales would struggle to exit at 15x EBITDA in this market. All of these businesses are cash flow positive, so as long as they continue to service and restructure their debt the sponsors can avoid the write offs.

Quote:

---

The 2021 and 2022 vintage SaaS take-privates are all generally in a bad spot. EV/Sales ratios were 10-15x+, and debt was cheap. Now debt is more expensive, growth has slowed, and multiples have collapsed. A deal that was done at 15x sales would struggle to exit at 15x EBITDA in this market. All of these businesses are cash flow positive, so as long as they continue to service and restructure their debt the sponsors can avoid the write offs.

---

Debt vs equity is going to be a huge part of this. But investors are itchy too, what's the exit strategy? Every SaaS company out there is trying to figure out how do we shift our revenue composition to get 30%+ from inference/compute, because that's at least what the market believes creates sustainable GRR. Without tha AI moat (and specifically how data funnels the AI moat competitive advantage), markets are dropping those valuations to 3x (airstory) to 6-8x (workday buyout offer from silverlake) based on risk adjusted discounted future cash flows. However, hit those numbers and you're sitting at a 20-40x multiple on ARR, which changes the entire portfolio risk comp. Interesting times to be in PE if you're there. No questions asked, whatever people say about SaaSpolcalypse, there's a continued big time filter the industry is going through. Not everyone is going to crater, but some will crash and burn hard.

Private money, VC money, under performs public stock indexes.

It's gamblers. Chasing the 1-2% of funds that beat the index. And for the winners, there is no convincing evidence that it's anything more than survivorship bias. You cannot predict the winners before they've won.

When a fund does well and becomes a winner, they get all of the money and connections flooding in. And then they get their choice of companies to invest in so it becomes a self fulfilling prophecy. If you're not in the very top funds that get first pick of investment opportunities, then you're probably going to be better off in a public index.

It's a feature not a bug . Every one of these alt managers will say something like "we love PC and where we are in the stack. Worst case, they default and then our PC becomes PE and we can run that playbook"

I disagree on the link between monetized inference and GRR. Inference costs generally scale linearly while subsequent outputs show diminishing marginal value. The long-term value sits in the system of record. As much as everyone hates them, ERP and CRM systems have defensible moats that will be hard to displace via a vibe-coded app due to the built in controls and governance within the data layer.

CRM/ERP and other SORs have a data moat for sure. But, they have to capitalize on it. For example, I'm not bullish on SAP's ability to execute because they're too german and too GDPR bound. For everyone else in SaaS (niche enterprise or single industry vertical companies), while it's true that compute/inference is a variable cost, the markets are objectively pricing in, at least in the short to midterm, ability to sustain GRR based on revenue composition. Presumably, the thought is longer term cash flows are more predictable in these companies. For non-AI native companies, expectation in the market is that deflationary headwinds are going to crush their future cashflow.

I'd push back on lumping private equity and VC together. They're very different. The hurdle for investing in VC should be pretty high. A generic VC fund probably doesn't justify the fees, illiquidity, capital calls and complexity versus just owning a low-cost public index.

To me, the case for VC depends a lot on whether you have access to really good managers and enough diversification to catch the handful of investments that drive most of the returns. Simply matching the S&P or Nasdaq isn't good enough. I think it's much harder to make that same argument about private equity as a whole.