How a Public Software Crash Became a Private Equity Loss

Software stocks in the U.S. have been valued at lower multiples than the rest of the S&P 500. That has never happened before. Not in 2022 when interest rates spiked. Not in 2008. Not during the dot-com crash. Forward price-to-earnings for software companies dropped below the broader market multiple for the first time in modern history.

That was weird. But the weirder part is what happened next, five months later, in a world most retail investors don't usually see.

HarbourVest Partners — the firm that benchmarks what happens inside private equity funds — released its Q1 2026 buyout data on August 31. The global buyout benchmark was down 1.6% for the quarter. Negative. In the same quarter, much of the broader market was holding flat or rising. And the cause wasn't macro stress, rising rates, or an economy turning sideways. It was software.

Information Technology — the largest sector in the private equity buyout benchmark at roughly 29% of total value — fell 5.0% in the quarter. Within IT, Software & Services dropped 6.4%. A selloff that happened in public software stocks between January and March 2026 had somehow traveled into portfolios of private companies that don't trade on any exchange, don't publish daily prices, and are supposed to be insulated from the market's mood swings.

How does that travel happen? That's the plumbing question. And it turns out the answer reveals something useful about what private equity valuations actually are.

The public market shock came fast. On January 29, 2026, Anthropic announced Claude Cowork — an autonomous desktop AI tool that could execute multi-step business workflows without human input. Microsoft lost roughly $360 billion in market cap that day. Over the next two months, the S&P 500 Software & Services index dropped roughly 27%. The iShares Global Software ETF (IGV) fell more than 30% from its September 2025 peak, erasing about $2 trillion in market capitalization. By April, stocks like HubSpot, Atlassian, and Figma had crashed 70–80% from their 52-week highs.

The narrative was structural, not cyclical. AI agents were no longer just a productivity feature — they were perceived as replacing the licensed human seat that most enterprise software companies sell revenue on. If one AI agent does the work of ten employees, you don't need ten software licenses. Goldman Sachs called the shift "Results-as-a-Service" — enterprises paying for outcomes rather than access. The traditional SaaS growth model, which tied revenue directly to headcount expansion, faced something closer to a reversal than a slowdown.

But here's where the plumbing matters. Private equity funds don't trade daily. Their portfolio companies are privately held. So how does a public market event become a negative return line item in a private equity benchmark?

The mechanism is comparable-company valuation — the bridge that connects the two worlds.

Private equity funds value their portfolio companies quarterly under ASC 820, the accounting standard for fair value measurement. Most buyout portfolio companies are classified as Level 3 assets — meaning their fair value requires significant management judgment because there's no observable market price. The primary valuation technique is a comparable-companies approach: you look at what public companies in the same sector trade at, calculate a multiple (like enterprise value to EBITDA), and apply it to the private company's earnings.

This is where the bridge appears. The private company's valuation depends directly on the multiples of its public peers. When public software multiples collapsed from roughly 6.2x EV/revenue at the end of 2024 to around 3.0x by Q1 2026, the comparable multiple input for private software companies fell with it.

And it's not just one dial turning. Stout, a valuation advisory firm, described what they called a "valuation squeeze" — simultaneous pressure from lower comparable multiples and higher discount rates. The discount rate reflects the risk-free rate, equity risk premium, and company-specific risk factors. In Q1 2026, the Fed held rates steady despite market expectations for cuts. Higher rates push up the discount rate, which pushes down present value. So the private software portfolio was getting squeezed from both sides: lower multiples and higher discount rates.

Then there's leverage. In a leveraged buyout, equity value is the residual — what's left after you subtract debt from enterprise value. A one-turn decline in the EV/EBITDA multiple can wipe a meaningful chunk of equity value because leverage magnifies the effect. If a company was bought at 14x and the multiple compresses to 10x, the equity might go from positive to deeply negative, even if EBITDA didn't change at all. That's the arithmetic of leverage turning a multiple compression into an equity markdown.

HarbourVest's own data confirms the mechanism. U.S. buyout investors marked down their software holdings by 8.9% from Q4 2025 to Q1 2026. The firm's head of quantitative investment science, Sofia Gertsberg, confirmed that the public market SaaS repricing was directly reflected in buyout benchmark returns.

There's an important detail here about timing and lag. Private markets are supposed to lag public markets in their pricing. The whole selling point of private equity is that it smooths out daily volatility — no 3% intraday crashes, no margin calls triggered by overnight headlines. But the lag narrows when the shock is big enough to force valuation committees, auditors, and lenders to reflect reality faster.

In Q1 2026, the lag was noticeably shorter than usual. The public software selloff happened in January and February. Private marks for the quarter ended on March 31. That's a two-month transmission speed — unusually fast for a market built on quarterly reporting and management discretion.

And the lag wasn't just about marks. It was about liquidity too. The IPO window for software remained essentially closed in Q1 2026. So the markdown stayed a paper loss — an accounting mark, not a realized sale. That distinction matters for understanding what investors actually own. The private equity fund's NAV went down, but the underlying companies were still operating, still generating cash, and still privately held at their marked-down values. The question isn't whether the loss is permanent — it's whether the new marks are too conservative or not conservative enough.

So what does this mean for someone outside the institutional world?

First, it reveals that private equity returns are less insulated than the marketing suggests. The "smoother ride" comes from delayed visibility, not from immunity. When public market multiples shift structurally — not cyclically, but because the underlying business model is questioned — that shift flows through the valuation bridge and appears in private marks, perhaps two months later, but unmistakably.

Second, the concentration matters. Information Technology was 29% of the global buyout benchmark. Software represented roughly 18% of all U.S. private equity deal value in 2025, the highest concentration on record. That's a lot of exposure to a single valuation thesis — that software commands premium multiples because growth is predictable, recurring, and asset-light. When AI calls that thesis into question, a disproportionate share of private equity portfolios get marked down.

Third, there's a credit dimension most people don't see. Software makes up roughly 25% of private credit portfolios, and those loans are typically marked near par — close to their face value — because they're performing. But the same valuation compression that hit equity also thins the equity cushion underneath the debt. If the enterprise value drops and the debt stays fixed, the recovery rate in a distress scenario falls. The market term for this is "refinancing risk" — loans that are performing fine today may struggle to roll over when they mature if the underlying business model has been re-rated. Fewer than 10% of software loans in some BDC portfolios mature before 2028, so the pressure won't show up in defaults tomorrow. But it will show up somewhere between now and then.

The longer view adds a complication. Software & Services remains the strongest-performing buyout industry over the past decade, with 21.9% annualized returns. The trailing one-year figure slipped to roughly flat. A negative quarter in a decade of stellar returns doesn't automatically mean the thesis is broken — but it does mean the period when every SaaS company traded at a premium because growth was assumed to be permanent has ended. The market is now distinguishing between software companies with defensible data moats and embedded workflows, and companies that sell generic productivity tools any AI agent can replicate.

The private equity market will sort out which is which over the next couple of years, one portfolio company at a time. Until then, the marks carry the uncertainty — and they're the closest thing private equity has to a price signal.