Skip to content

Exit Proof

VC rounds mostly buy dilution, and founders pay for it at exit

Petri Lehmuskoski ·

Across 10,000+ startups reviewed, my read is simple. In most cases a VC round buys dilution, not company value growth. The founder ends up owning less of a company that is not worth proportionally more.

A paper from Wharton and the University of Zurich now puts a number on where that road ends. Eduardo Azevedo, Florian Scheuer, Kent Smetters and Min Yang tracked 248,454 founders at 157,669 US VC-backed startups founded between 1987 and 2025. In the July 2026 version of their paper, Dilution vs. Risk Taking: Capital Gains Taxes and Entrepreneurship, 88% of those founders received nothing at exit.

Most VC-backed founders leave with nothing, even when the company sells

The zero comes from three places.

In 81% of cases the founders are wiped out as a group. Either the company fails, or the investors take all the proceeds of a weak exit. Only 31% of the exits in the data carried any value at all.

Selling the company does not save the founders either. Among IPOs and acquisitions, 19% left the founders' combined stake at zero.

The third source is softer. A founder who left before the exit is counted as zero. In the IPO filings the authors used, 39% of the people tagged as founders were missing from the cap table. The authors say openly that some of them were paid through private sales before the IPO. That part of the 88% overstates the problem.

For the 12% who do get paid, the median payout is $4.8M after an assumed 28% tax. The top is brutal: the top 1% of founders take 79% of all founder exit value.

Read the paper for what it is. It is a tax study, written to test what taxing unrealised capital gains would do to founders. It covers US companies only, VC-backed companies only, and it ignores secondary sales. It makes no recommendation on whether to raise. It describes the payout distribution on the VC path, and that is enough.

The zero is written into the preference stack long before the sale

The authors trace the zero in sold companies to investor-protection terms. Liquidation preferences pay investors first. Anti-dilution clauses hand investors extra shares after a down round. Their model assumes a 1.25x preference.

The arithmetic is worth doing once, slowly. A company is sold for €30M. It has raised €20M. The founders hold 40%, the late-stage investors 30%, the early investors the rest. The preference sits only on the latest round and does not participate. The late-stage investors compare a €20M preference with the €9M their 30% would return, take the preference, and the remaining €10M is divided among everyone else. The founders receive €5.7M. A 40% holding collects 19% of the price.

Now change one number. The same €20M at a 2x preference is a €40M claim against a €30M sale. The late-stage investors take everything and the founders receive nothing. Nobody has behaved badly and nobody has broken the agreement. To get back to €5.7M, the founders would now need a €50M sale.

A term-sheet multiple leaves ownership alone. It moves the finish line.

Preferences are rarely one layer. Each round adds its own, and the stack grows whether or not the valuation does. Every round celebrated as progress raises the price at which the founders receive their first euro.

Employees' options go to zero alongside the founders

The paper counts founders only. Employees stand in the same place or lower. Option holders sit together with the founders at the bottom of the waterfall, or beneath them. When the founders get zero, options and RSUs go to zero too.

A great exit for the first angel was a bad deal for the founders

One of my early investments ended differently for different people in the same room. For me, as the first angel in, it was a strong exit. For the founders, it was not.

Between finding the right segment and the eventual sale, the company raised more capital than its own evidence required. Much of it went to opening offices in several countries and diversifying into adjacent products before either move had earned proof. Most of it did not work. What it did produce was a series of bridge rounds on hard terms, each one taking a larger share of the founders' ownership than the round before.

The exit was real. For the founders it arrived heavily diluted by rounds that had bought time and geography instead of proof. At a later presentation, one of them said what by then needed no explaining: they could have had a better outcome without the VCs at all.

An exit can be proof the company mattered, and still be evidence that someone else decided how to get there.

An €8M exit without VC can pay a founder what the VC path rarely does

A founder who grows the camel way, financed mostly by customers, does not need substantial dilution. Keeping 60 to 80% of the company is realistic. At an €8M exit, that founder takes €4.8M to €6.4M before tax.

That is the same order of magnitude as the paper's $4.8M median for the 12% of VC-backed founders who got paid. It is not an exact match: mine is in euros and before tax, the paper's is in dollars and after tax. The point is the size of exit needed to reach it. On the VC path, the founder needs a far larger exit to clear the stack and still be left holding the same amount.

The probabilities differ too, but they need careful reading. Of the camel companies we have backed, 25 to 30% have failed. That is a company failure rate. The paper's 81% measures something else: founders wiped out, including in companies that sold. The two numbers do not share a denominator, and I will not pretend they do.

Time is the third cost. The paper's model implies an average of around 16 years from founding to exit. My read is that the camel path gets a founder to the same money with much less risk and in much less time, and keeps the options open: sell, keep going, or raise.

A VC will say the 88% is the price of the jackpot. Across all founders, zeros included, the average payout in the paper is $5.6M, and the top of the distribution runs to hundreds of millions. On expected value, a rational founder takes the lottery.

The jackpots are real. They are also rare. The median founder in the paper gets zero.

Without a local exit market, the VC path needs an even bigger exit

The paper's 88% comes from the US, the deepest exit market in the world. In less developed markets it is worse.

Where there are no local buyers for a large company, the VC path has to aim at a foreign exit, usually in the US. That exit has to be bigger, it takes more rounds to reach, and every round adds a layer to the stack above the founder. Many companies move to the US for that reason: to raise there for an exit that only exists there. Small trade sales still find foreign industrial buyers. The large exit is what goes missing at home.

Raising VC is the right call when development or go-to-market needs heavy capital

Some companies have no camel path. Deep tech, medtech and medical companies, anything with a long development cycle, and industries where going to market requires substantial capital cannot avoid multiple rounds.

For those founders the question changes. It stops being whether to raise and becomes what each round does to the waterfall. Run your own waterfall after every round rather than in the month before a sale. Ask, at term sheet stage, whether later preferences will be stacked or pari passu. Know your break-even exit, and choose investors whose required outcome you can actually deliver.

Fair question: why is a VC writing this? For 10+ years, Gorilla Capital has invested in camel companies, where founders keep control, on sensible terms and with less dilution. We studied the market data and facts before choosing that strategy, and the market has not changed much since.

For everyone else, the zeros in sold companies were written into term sheets signed years before the sale.

You choose your exit strategy the day you choose your investors.

More scars than trophies.