The exit market has changed. The capital stack should too.
Record exits, very few winners
On the surface, 2026 is the year that long awaited exits came back. US venture-backed tech IPOs have raised nearly $90 billion[1] (the second-highest total on record), and venture-backed M&A is on track for a record year by value.
Look closer, however, and almost all of it went to a few names (at best). SpaceX accounts for 83% of the IPO capital, and its $60 billion acquisition of Cursor[2] makes up roughly half of US startup M&A spending. Private equity, the other traditional exit route, has largely stepped back from tech.
It seems the exit market has split in two: a handful of companies have achieved record outcomes, while many growth companies are just waiting (along with their early investors, employees and lenders). So what capital actually works for those left waiting? Debt is often an overlooked answer: it can buy these companies time, and it can fund liquidity for the shareholders waiting alongside them.
IPOs: record money, very few companies
Maybe surprisingly, about a quarter of this year’s tech IPOs came from the energy sector (Fervo, X-energy, Hadron Energy and Standard Nuclear), quantum (Quantinuum), defence and space (HawkEye 360, York Space Systems) and equipment rental (EquipmentShare) filled out much of the rest.
Enterprise software, the darling of venture-backed listings for two decades, has been almost entirely absent. This is not a shock: AI disruption, seat/user compression, budget shifts etc…at one point software traded at a discount to the S&P 500 for the first time in the modern era.
The bar for everyone else has risen sharply, with some suggesting the new threshold for a successful IPO is somewhere around $4 billion in revenues growing 50% or more. Few private software companies are anywhere near that… and Europe’s next hope for an (albeit Nasdaq) IPO, Oura, has pulled its listing due to uncertainty in the market.
Trade sales and private equity: the same pattern
Those that don’t list have traditionally turned to trade sales and private equity. In 2026, neither is doing that job for the middle of the market.
Strategic M&A looks strong on paper. Crunchbase cites at least 1,177 venture-backed acquisitions worth $182.7 billion announced by mid-June, against 1,132 deals worth $106.7 billion a year earlier[3].
However, the preserve of the elite is just as apparent under the hood – the same concentration applies as for IPOs. Roughly half of this year’s US startup M&A spending came from a single deal, SpaceX’s $60 billion purchase of Cursor. Acquirors are paying up for the few companies they really desire, they are not just clearing the queue or buying the revenues.
Private equity has pulled back further. Technology buyout deal value fell 70% between Q4 2025 and Q1 2026, and $1 billion-plus tech deals dropped from 15 to 4[4]. PE has long been the natural buyer of solid, mid-sized software businesses…so where do they turn to now?
Interestingly PE may also have its own exit problem. Buyout distributions have been at record lows relative to NAV for years, and tens of thousands of PE-backed companies remain unsold. More simply: the PE sponsors still trying to exit their own pre 2022 purchases are unlikely to be eager acquirors of more software in an increasingly uncertain market.
Why buyers have gone cold on software revenue
Public investors, strategic acquirors and PE buyers are all pricing the same problem: software revenue has become harder to trust. It affects established SaaS companies and AI-native companies in opposite ways.
Established SaaS is losing seats at renewal. Higher churn, weaker NRR, uncertain enterprise customers, seat compression due to AI agents…the list goes on.
AI-native companies have fast growth but fragile revenue. Is it committed spend or an experimentation budget? Is it about to be rapidly replaced by “better” agents or is the Tier 1 customer completing its own competing product in house? Gross revenue retention appears lower for AI-native companies even against the ever-weakening B2B SaaS ones. And average AI gross margins are also relatively low in comparison, at around ~50%, versus the 80% to 90% investors came to expect from traditional SaaS.
The result is that buyers pay up for the few companies whose revenue quality (and growth) is beyond question (“if I buy it today, will it have any revenue tomorrow?”), and wait on everyone else.
The stranded middle
Somewhere in all the noise sits a large group of growth companies with no clear route to an exit. Many raised in 2021 and 2022 at valuations set by SaaS multiples that no longer exist. Plenty of them are growing still and some are profitable, but not at the scale or speed public markets now reward.
For a company that can raise at will or become profitable, staying private is a very credible decision, but for many others waiting is not exactly free. The business faces a choice between a flat or down round or high liquidation preferences, while early investors in legacy funds may seek DPI, and employees hold options that may look less and less attractive (and thus retention becomes harder).
These companies need time and ideally a way to give their shareholders some liquidity while they wait.
What about Secondaries?
While the IPO market concentrated around the few, the secondary market did the opposite. Total secondary volume reached $124 billion in the first half of 2026[5], up 28% year on year, according to Lazard. Big, shiny secondary-purpose vehicles have followed (CVC, Stepstone etc).
A secondary sale still needs a buyer though, and buyers set the price. For a company that wants to offer its early backers liquidity, the question is who funds it and on what terms.
Where venture debt fits
For companies stranded in the middle, debt does two jobs that others maybe cannot do:
- It buys time (or other companies) without a new valuation or dilution. Whilst not the only use of our financing, this is something we have done at Claret many times since 2013. Companies with solid growth, efficient metrics, durable tech and robust customer bases can increase optionality by borrowing. Either they grow into their pre 2023 valuations, they improve the odds of timing their exit well, they wait for markets to settle or they show more proof points and value drivers in their model. Or of course they build scale and profits through M&A – increasing enterprise valuation by much more than the cost of the debt or the price of equity to fund the same acquisition.
- It can fund liquidity for shareholders directly, whereby the company borrows and uses the facility to buyback shares from early investors (and employees) at an agreed price. Shareholders get liquidity, no new equity investor need be found and the remaining cap table is not diluted.
Claret can provide both kinds of financing.
The record exit numbers are real, but they are the preserve of the elite. For the growth companies in between, venture debt is often the cheapest way to buy time, and is increasingly a way to fund the liquidity their early shareholders may seek now.
[1] A Hard Year For Software IPOs, Crunchbase News, 16 September 2026
[2] Cursor Deal Puts US On Track For Record Startup M&A Year, Crunchbase News, 15 June 2026
[3] SpaceX Acquires AI Coding Tool Cursor For $60B In Year’s Largest Startup M&A Deal, Crunchbase News, 16 June 2026
[4] What Bain’s Midyear Report Tells Us About Where the Market Really Stands, Private Capital Global, 10 June 2026
[5] #373: What’s Driving Venture DPI?, @TheFundCFO Newsletter, 15 Sept 2026
