Exits are back: what the record quarter really means for your next round

For the past three years, "exit" has been the word UK founders learned to say quietly. IPO windows were shut. M&A mostly meant selling at prices nobody put in a press release. And every DeepTech founder came to dread the question that surfaces in the earliest commercial conversations - who, exactly, is ever going to buy this?

Then came the second quarter of 2026. 32 companies went public at valuations above $1 billion. Twenty-four more were acquired at or above $1 billion, worth a combined $113 billion, the highest quarterly total on record. SpaceX delivered the largest venture-backed IPO in history, listing at $1.77 trillion and raising $75 billion.

These are global figures, per Crunchbase, dominated by American exchanges. And one company distorts them badly: strip out SpaceX and the record loses much of its shine. Not everyone got out. But the window, unmistakably, reopened.

So, should you care? You are raising a Seed round in Cambridge or a Series A in Bristol. Nothing about a trillion-dollar rocket listing changes your term sheet.

Except it does. Here is the thing to understand about an exit window: it moves at two speeds. Belief travels at the speed of a headline. Cash travels on a ten-year clock. The two look identical from a distance and confusing them is one of the more expensive mistakes a founder can make this year. The first will change how investors talk to you within weeks. The second determines whether they can actually write the cheque.

The quarter the queue moved

The Q2 numbers describe the crest of a wave, not a bolt from the blue. Sentiment had been turning underneath for a year. By mid-2025, Carta reported that the share of US venture funds beginning to return capital to their investors had jumped across every vintage from 2017 through 2023, as fund managers showed a new urgency about generating distributions.

That urgency was not altruism. It was survival. Fund investors had grown loudly impatient with paper gains, and managers knew their next fundraise depended on returning real cash. So, when the public market finally offered a door, there was a queue already pressing against it. Q2's record is what pent-up supply looks like when the door opens.

For founders, the sentiment shift is the fast dividend. An investor who last quarter modelled no plausible liquidity for a hardware business now models one. The mood in pitch meetings changes long before any money does.

Hardware at the front of the queue

Look at who led that queue. After SpaceX, the two largest Q2 listings were not software companies. Cerebras, the AI chipmaker, rose almost 70% on its debut in May. Quantinuum, the quantum computing company, raised $1.68 billion in an upsized offering and, per CNBC, closed its first day worth $15.7 billion.

Now set that against public software, where funds tracking listed software companies have fallen by 30% since the start of 2026 as investors question what survives AI commoditisation, a rout the industry has nicknamed the "SaaSpocalypse".

We argued in February that hardware is becoming the defining moat of the AI era. The 2026 European Deep Tech Report makes the same case from the capital side: DeepTech investment overwhelmingly builds defensible IP and products that cannot be recreated overnight with generative AI tools. The public market has now voted. Engineering depth is being repriced upward at the precise moment software-only defensibility is being marked down.

For a DeepTech founder, the practical effect is simple: the Series A objection, that businesses like yours have no exit path, now has names attached.

Priced on belief, not earnings

Look closer at Quantinuum, though, because the most important detail is not in the headline. Its own SEC prospectus discloses that first-quarter revenue fell 73% year on year to $5.2 million, with a quarterly net loss of $136.6 million. The market looked at those numbers and still paid a remarkable $15.7 billion.

That is not a valuation of earnings. It is a valuation of conviction: a bet that trapped-ion quantum computing matures into infrastructure, and that this company is among those left standing when it does. For a decade, public markets punished pre-profit hardware. They are now, visibly, prepared to underwrite long-duration engineering categories on trajectory.

But conviction cuts both ways. Analysts expect Quantinuum's performance to influence valuations across the whole quantum sector, where stocks tend to move in tandem. A repricing that arrives on belief can leave on belief. A category re-rated upward in a fortnight can be re-rated downward just as fast, and every private valuation in that category moves with it.

Belief is the fast speed. Use it. But do not lean your whole weight on it.

The ten-year clock

Now the slow speed. If funding records kept being set these past two years, why did your Seed round stay so brutally hard? Because the people at the top of the capital chain stopped getting cash back. Carta's fund data, covering US funds, shows that five years in, 59% of funds raised in 2017 had begun returning any capital at all to their investors. For the 2019 vintage, the figure was just 39%. Three of every five funds from that year had distributed nothing at the five-year mark.

The mechanism runs in a loop. Fund investors recycle distributions into new funds. When distributions stall, new allocations stall. When allocations stall, cheques thin out, and they thin out at the earliest stages first. That loop is the market UK founders have been living inside for three years.

A record exit quarter starts to refill the pool. But it refills through fund lifecycles measured in vintages, not quarters. Cash from June's listings must clear lockups, waterfalls and reallocation decisions before any of it reaches a new Seed fund. The clock has started. But it has only just started.

The metric that decides who can pay you

Where the two speeds meet is a number most founders never think to ask about: DPI, or distributions to paid-in capital, the measure of cash a fund has actually returned to its investors. It has quietly become the gatekeeper of the fundraising cycle. With cash returns scarce, LPs grew hesitant to re-commit to venture at all, and the funnel narrowed brutally: roughly 1,300 new VC funds closed globally in 2024, down from 4,000 at the 2021 peak, with 81% of new capital flowing to established firms.

The implication for founders is awkward. The question is not whether "the market" has money. It is whether the specific fund across the table has a recent close, live deployment pace, and a distribution record that lets it raise again. A partner at a firm that cannot raise its next fund can take your meeting, run your full process, and never be capable of leading your round.

And that data, the fund closes, the pacing, the distribution track records, is largely non-public. It sits inside institutional research platforms priced far beyond any startup's budget. Most founders are therefore pitching blind to the single variable that decides whether a term sheet can physically be written.

The window opened in New York

One asterisk, and it is a familiar one. Per KPMG's Venture Pulse, European exit conditions stayed subdued in early 2026, with IPO activity soft and M&A the more viable route to liquidity. The record quarter happened overwhelmingly on American exchanges.

Quantinuum makes the point painfully well. The company was formed in part from Cambridge Quantum, one of Britain's genuine quantum pioneers. It listed on Nasdaq. The pipeline behind it, with Anthropic having filed confidentially and OpenAI preparing to follow, per CNBC, is American too.

Regular readers will recognise the pattern. We wrote in September about the great transatlantic divergence, and about why most late-stage funding for UK scale-ups arrives from overseas. Q2 extends the same asymmetry to exits: the comps cross the Atlantic instantly, the listing venues do not.

For a UK DeepTech founder, that is uncomfortable but usable. The belief created in New York reprices your Cambridge cap table. The liquidity, for now, stays mostly on the other side. Build your investor map, and in time your exit map, accordingly.

Two speeds, two disciplines

So, hold both truths at once. Use the fast speed now. The exit objection that has shadowed every DeepTech raise for three years can be retired with named, public counterexamples, and the repricing of engineering depth will do quiet work for your narrative in every meeting this autumn.

Refuse to depend on the slow speed. Size the round to the milestone, not the mood. The cash behind the record has not reached the Seed market - and will not for some time. Build your runway on the market you are in, not the one the headlines describe.

And keep one habit as a watching brief: the slow money will arrive one fund announcement at a time, each new close the ten-year clock striking. The window is open. Just be clear about what has actually arrived: the belief is here now, but the money is still years behind it.

 
Let's talk.