Founders and founding engineers take the same risk for similar cash. The equity is not the same — one side gets QSBS-eligible restricted stock, the other gets a tax bill.
May 26, 2026
There’s a guide on startup equity I recommend reading first: Team Equity. It’s a good overview of how the different instruments work — RSA, ESOP, RSU — and why they’re not interchangeable.
Once you understand the mechanics, something becomes obvious: founders and founding engineers are not in the same boat. The downside is roughly the same. The upside is not.
A seed-stage founder pays themselves about $130–150K, though $200–250K is not uncommon either once a real round closes. A mid-level engineer at the same company makes $100–145K in the Bay Area. A senior one makes $180–235K. So if you join as a founding engineer, you’re not taking a meaningful pay cut relative to the founder.
The risk of failure is also the same. You both spent years on something that returned zero.
So far, symmetric.
A founder gets the best equity instrument that exists. Restricted stock, granted at near-zero strike price, with the QSBS clock starting on day one. Five years later, the first $15M of gains are federal-tax-free. Anything above that is long-term capital gains, ~20%.
A founding engineer joining after the seed round gets none of this. The options are:
Two co-founders incorporate a company. Each gets 5,000,000 shares of restricted stock at $0.0001/share — $500 each. They file 83(b). The QSBS clock starts.
Seed, month 6. They raise $4M at $24M post-money. Investors take ~17%. A 10% ESOP pool is carved out pre-money. Founders end with ~36% each.
Series A, year 2. They raise $15M at $60M post-money. Investors take 25%. ESOP refreshed back to 10% post-money. Founders end with ~25% each.
The first founding engineer is hired right after the Series A. They get a 1% ISO grant — 250,000 shares at a $0.50 strike (the 409A FMV). Cost to exercise the full grant: $125K, for illiquid stock. They don’t early-exercise. 4-year vesting, 1-year cliff.
By this point the founder has been holding restricted stock for two years. The QSBS clock is running. Three more years and the first $15M of their gains is federal-tax-free.
The engineer’s QSBS clock hasn’t started. It can’t, until they exercise. And now they face a choice that the founder never has to make.
Path 1: they leave. Maybe at year 3, vested at 50%. 125K shares of ISOs. They have 90 days to exercise or lose everything. Exercise cost: $62,500 out of pocket. On top of that, AMT on the spread between strike and current FMV — if the company has raised a Series B in the meantime, the FMV could be $3–4/share, so AMT on $400–500K of paper gain, easily another $100K of tax due that year. Total cash required to keep the equity alive: $150–200K. For shares that may still never become liquid.
The other option is to walk away from the equity entirely. Many people do.
Path 2: they stay. The typical hold from Series A to exit is 7–10 years. Tender offers, when they happen at all, usually arrive at Series C or D — year 5 or 6 — and let employees sell a small fraction. The rest stays locked up until the company exits or dies.
So the engineer’s equity question is really: are you willing to commit the next 7–10 years of your career to this one company, on the chance that it exits and on the chance that the exit is large enough to matter after the tax haircut? Because the alternative is writing a check the day you leave for paper that may never become real.
Meanwhile the founder’s QSBS clock turned five at year 5 either way.
Joining a startup as an engineer is a VC investment where you put in time instead of money. Multi-year time, at a salary discount that could have been earned in big tech, in exchange for a lottery ticket with worse tax treatment than the founder’s.
VC works because of the law of large numbers. A fund invests in 30 companies, expects 25 to fail, and one or two to return the whole fund. You can’t do that with your career. You get maybe two or three founding-engineer bets across your working life. The variance doesn’t average out.
For a skilled engineer with options, big tech is the better expected-value play. The salary is a multiple higher. RSUs vest on a schedule and are liquid. There’s no exercise problem, no QSBS clock, no acquisition that turns your options into a tax bill. The only reason to take the founding-engineer seat instead is experience — building from zero, owning a whole system, the optionality of becoming a founder yourself later. Those are real. But they are not monetary upside. Calling them monetary upside is the trick.
The exception is when you’re 100% sure the company will be exceptional. But “100% sure” is not a thing in startups. The founders themselves aren’t 100% sure.
A few things would actually help.
Extended exercise windows. A growing minority of companies (Pinterest, Quora, others) give employees 5–10 years post-departure to exercise instead of 90 days. This is a partial fix — QSBS still doesn’t apply — but it breaks the worst dilemma in ESOP: leave and write a six-figure check on the way out, or stay locked into the company for years to avoid it.
Larger grants to compensate. If the instrument is going to take a 50% haircut, the grant size should reflect that. A 2% grant treated as ordinary income is equivalent, after tax, to a ~1% grant of QSBS-eligible restricted stock. Founders and boards should price this honestly.
Tax reform. Under §1202, the QSBS clock starts when you actually acquire the stock, and for an option that means exercise. That’s not really ambiguous — it’s what the IRS treats as the date of issuance, and it’s what people file. The real problem is that exercising is expensive. For a founder, “acquiring the stock” means writing a $500 check at incorporation when the FMV equals the strike. For an engineer joining at Series A, the same act means writing a six-figure check and paying AMT on the spread between strike and current FMV. So the clock-start rule looks fair on paper and is brutal in practice — only people who can afford to exercise early get the QSBS benefit.
The targeted reform isn’t changing when the clock starts. It’s making early exercise feasible: allow ISO grants to be treated as QSBS-qualifying from the grant date if the holder files an 83(b)-equivalent election, without requiring the upfront cash and AMT. That would let engineers start the clock the same day the founder did, at no real cost to the founder.
None of this fully equalizes things. The founder still gets the best deal. But the gap could be a lot smaller than it is, and the people who could close it are the founders themselves.
