Count the deals and UK startup M&A looks only a little slower. About 110 startups that had raised £1m or more were sold in each of 2021 and 2022; since then it has been about 90 a year. The trouble with counting deals is that the number of companies waiting to be sold has grown by more than a quarter over the same period.
We took every startup in Sonde that had raised £1m or more, had outside investors on its share register and was less than 15 years old, and followed each one year by year: was it sold, did it fail, or was it still independent at the end of the year. There were about 3,600 such companies at the start of 2021 and about 5,200 at the start of 2025.
Sonde analysis of Companies House confirmation statements, PSC registers and SH01 filings, corrected for reporting lag and validation precision.
In 2021 and 2022, close to 3% of these companies were sold in a year. Since 2023 it has been a little under 2%, which is roughly where it was before the boom. In short, what this means is that the deals have not dried up, but a startup's chance of being one of them in any given year has fallen by about a third.
Failures have risen much faster
The other way out is the one nobody plans for. The share of these companies entering administration, liquidation or being struck off has gone from 1.8% in 2021 to 4.5% in 2025.
Same data. Insolvency and strike-off filings at Companies House.
Put simply, a £1m-backed startup was about two and a half times as likely to fail in 2025 as in 2021 (2020 and 2021 were flattered by the pandemic support schemes, which held insolvencies down across the whole economy, so 2019 is the fairer comparison, and the rate has still risen by nearly half since then).
Same data.
Put the two together and the balance has flipped. In 2021 and 2022 there were fewer failures than sales. In 2025 there were more than two failures for every sale.
The deals have not dried up, but a startup's chance of being one of them in any given year has fallen by about a third.
Five years on
Yearly rates are abstract, so we also followed each company from the year it first passed a raised-capital mark. The 2019 group is the most recent with five full years behind it.
Five years after first passing the mark
| First passed | £1m: companies | Sold | Failed | £500k: companies | Sold | Failed |
|---|---|---|---|---|---|---|
| 2017 | 135 | 7.6% | 2.7% | 240 | 8.2% | 6.0% |
| 2018 | 252 | 8.4% | 7.5% | 446 | 6.6% | 8.7% |
| 2019 | 419 | 9.2% | 9.7% | 579 | 8.1% | 10.1% |
For the companies that passed £1m in 2019, sales and failures over the following five years were level, at 9.2% and 9.7%. At the £500k mark, 8.1% of the 2019 group had been sold and 10.1% had failed five years on. Roughly four in five had done neither. The 2017 and 2018 groups are small, a few hundred companies at most, so I would not read much into the differences between them.
Most of each group is still privately held five years on. Among the companies that did reach an outcome, 2019 is the first group in which failures outnumber sales at both marks.
Who is buying, and when
Trade buyers dominate. Of the 732 sales we found from 2019 to 2025, about 15% went to a financial buyer (a private equity fund or its acquisition vehicle). At least 28% went to a buyer registered outside the UK (that is a floor, because many overseas groups buy through a UK subsidiary).
Companies are also older when they sell. The median company sold in 2023 to 2025 was 8.5 years past incorporation, against 7.3 years for sales in 2019 to 2022.
Why I think this is happening
I suspect two things are going on. The companies that raised in 2020 and 2021 did so at prices that buyers have been slow to meet, so sales that would have happened are being deferred rather than cancelled. And the same boom created a large cohort that is now reaching the age at which a startup usually either sells or runs out of money, which is exactly where the failures are concentrated.
Method
How we measured this
We started from every company carrying Sonde's venture tag (at least one clean priced equity round of £100k or more, excluding PLCs, holding vehicles and intra-group capitalisations). A company enters the analysis from 1 January of the year after its clean equity raised reaches £1m (or £500k for the second series), as long as its share register shows outside investors (five or more holders, an investment fund, or a minority corporate investor) and it was incorporated within the previous 15 years. The age limit matters: without it a third of the 2025 population was more than ten years old, established businesses with private equity or family money rather than startups (the headline figures barely move either way, but the word would be wrong). A company leaves the population when it is sold, fails or passes 15 years.
A sale is a new owner taking all the shares, or a majority by buying existing shares, read from consecutive confirmation statements and the PSC register, and dated by the PSC register where it records the new owner. We exclude group reorganisations (a new holding company over the same owners, or a move to an overseas parent), sales to founders' own vehicles or to employee trusts, existing controllers buying out minorities, and register misreads. A company sold after a reorganisation is followed to its new UK parent. A failure is an administration, a creditors' or compulsory liquidation, a voluntary arrangement or a strike-off before any sale; solvent wind-downs are counted separately.
We then checked our work. We sent a random sample of 125 sales, weighted towards recent years, to be verified against Companies House records and press coverage. Of the sales dated 2018 to 2022, 42 of 45 were genuine changes of control. Of those dated 2023 to 2026, 60 of 80 were, because the signals that expose a reorganisation or a misread take time to appear. We corrected each year's count for that, and for sales that have happened but not yet been filed. And 18 of 20 sampled failures were genuine losses for shareholders (the other two were solvent closures).
Three limits are worth stating. The correction matters: without it, the share sold falls from 3.0% to 2.3% rather than from 2.8% to 1.8%, a fall of about a fifth rather than a third, and the two precision figures each carry a margin of about ten points, so treat the true fall as somewhere between a fifth and a half. Our filings start ten years back and our company set over-represents survivors among companies added recently, so failure rates are if anything understated. And a company that moved to an overseas parent and was later sold abroad is invisible to us.
See the numbers
| Year | Startups at 1 January | Sold | Failed | Share sold | Share failed |
|---|---|---|---|---|---|
| 2019 | 2,669 | 56 | 84 | 2.1% | 3.1% |
| 2020 | 3,177 | 54 | 56 | 1.7% | 1.8% |
| 2021 | 3,602 | 103 | 65 | 2.9% | 1.8% |
| 2022 | 4,229 | 116 | 94 | 2.7% | 2.2% |
| 2023 | 4,732 | 79 | 132 | 1.7% | 2.8% |
| 2024 | 5,087 | 87 | 201 | 1.7% | 4.0% |
| 2025 | 5,228 | 98 | 236 | 1.9% | 4.5% |
What's next
This piece counts sales of whole companies. The other route to liquidity is a secondary sale, where an existing investor sells shares to a new one while the company stays independent, and that is the next thing we want to measure. We also want to see which companies are more likely to be sold, which, from a first look, has more to do with how long since they last raised than how much they raised. If you're seeing the same pattern in your own portfolio, I'd be interested to hear: get in touch.