A $150m Valuation Does Not Mean the Founders' 35% Is Worth $52m
A startup raises at a $150 million post-money valuation. The founders own 35%. What is their stake worth?
The instinctive answer is $52 million. The honest answer is: it depends entirely on the exit, and on terms most people skim.
A headline valuation is the price paid for one share of the newest, most senior, most protected security in the company. Founders and employees hold a different security — common stock, with no protection at all. So common is worth less than the headline price per share, and at modest exits it can be worth nothing.
I built a workbench to make that visible. It prices the round with the methods that fit the company's stage, then runs the cap table and the full liquidation waterfall so you can see what the price actually pays each holder.
Open the private company valuation workbench — four stage presets, seven methods, everything live.
The number that matters
On the Series B preset — $45 million raised with a participating preference capped at 3×, senior to the earlier rounds:
| Exit value | Founders own | Cap table implies | Waterfall pays | Share of implied |
|---|---|---|---|---|
| 0.6× post-money | 35.2% | $32m | $13m | 40% |
| 1.0× post-money | 35.2% | $54m | $38m | 70% |
| 2.0× post-money | 35.2% | $107m | $91m | 85% |
An exit exactly at the headline valuation leaves the founders with seven-tenths of what their percentage suggests. Below about $63 million they receive nothing.
Then flip one checkbox — make the preference non-participating — and at 0.6× post the founders' share jumps from 40% to 61%. One term, half again the founder outcome. That is why the term sheet deserves as much attention as the price.
Five things the build made explicit
1. Choose methods by stage, not by preference
Before revenue, there is nothing to multiply and nothing to discount. Berkus and the scorecard method are structured opinion: they turn judgements about team, product and market into a number that can be argued about consistently. Once revenue is real, a revenue multiple starts to carry information. Once there are earnings, EBITDA multiples and a DCF take over and the venture methods fall away.
The spread between methods is the finding. On the seed preset the midpoints run across more than a factor of two. A blended number is a negotiating anchor, not a measurement.
2. The VC method and First Chicago double-count if you let them
The VC method demands a 20× return at seed because most investments fail. First Chicago puts failure in explicitly, as a scenario with a 55–60% probability. Apply a venture-style 50% discount rate on top of that and the same risk gets charged twice. The workbench uses 22–30% on First Chicago for exactly that reason.
3. A growth-stage DCF is mostly terminal value — and sometimes negative
On the Series B preset the terminal value is more than twice the entire DCF, because the ten forecast years are worth less than nothing combined. Growth is being bought with reinvestment.
Drop the sales-to-capital ratio from 3.0 to 1.5 and the DCF goes negative outright: growing 90% a year at that capital intensity destroys value at an 18% discount rate. That is the most useful thing a DCF says about a startup — growth is only worth paying for if each reinvested dollar buys enough revenue.
4. The option pool shuffle
Investors usually require the option pool to be topped up to a target after the round — and the top-up is created inside the pre-money, before their money arrives. The investor's percentage is untouched. The existing holders absorb the whole cost. The workbench shows the effective pre-money, the price the founders are really selling at, beside the headline.
5. Conversion is a decision, and the model has to solve for it
A non-participating preferred holder chooses at exit: take the preference, or convert to common. Each series' best choice depends on what the others do. The engine iterates until no series would gain by switching, and the test suite checks that equilibrium at 80 exit values per preset, along with two other invariants: payouts always sum exactly to the exit value, and no holder's payout ever falls as the exit rises.
What it leaves out
Options are treated as exercised with the strike ignored. No SAFEs or convertible notes, no anti-dilution on down rounds, no pay-to-play, no accruing dividends, no management carve-outs. All of these exist in real term sheets, and most of them move money away from common. The presets are illustrative, anchored loosely on published medians — per Carta's Q1 2026 data, a non-AI Series A sits near $55 million against roughly $300 million for foundation-model companies at the same stage. Not investment advice.
Part two covers public companies: what today's share price already assumes about the future, for retail investors, and how a leveraged buyout actually makes its return.