SAFEs are designed to be easy to sign. That is their best feature and their sharpest edge: because each one takes an afternoon, founders sign them one at a time over months, $100k here, $200k there, and nobody ever sits down to compute what the whole stack converts to. Then the priced round arrives, the conversion math runs, and the founders discover they sold more of the company than they thought.
This post walks one realistic stack through conversion, step by step, with every number shown. If you have SAFEs outstanding and haven't done this exercise, do it this week.
The stack: four SAFEs in 18 months
Meet a typical seed-stage company. Over 18 months it signed four post-money SAFEs, each at the valuation cap that felt right at the time:
| SAFE | Investment | Post-money cap | Locked ownership |
|---|---|---|---|
| #1: angel | $100,000 | $4,000,000 | 2.50% |
| #2: pre-seed fund | $200,000 | $6,000,000 | 3.33% |
| #3: seed fund | $400,000 | $8,000,000 | 5.00% |
| #4: strategic | $500,000 | $10,000,000 | 5.00% |
| Total | $1,200,000 | — | 15.83% |
The fourth column is the whole story, and most founders never total it. A post-money SAFE locks the investor's ownership at signing: ownership = investment ÷ post-money cap. $100k on a $4M cap is 2.5%, fixed on the day you sign, no matter what you raise afterwards. Add the four up and this company has already sold 15.83% of itself. It just doesn't show on any cap table yet, because SAFEs sit off to the side until a priced round converts them.
Each new post-money SAFE dilutes only the founders and existing shareholders, never the earlier SAFE holders, whose percentages are locked. Signing SAFE #4 didn't touch SAFEs #1–3. It came entirely out of the founders.
Then the round happens
Eighteen months in, a lead offers a Series Seed: $3M on a $12M pre-money valuation, with the usual condition that a 10% option pool be in place post-round, topped up before the money comes in. Post-money valuation: $15M. The investor's share is $3M ÷ $15M = 20%.
Now everything converts at once, and the founders' final number assembles from three separate bites:
Worked example: where the company goes
1. New investor: $3M ÷ $15M post-money = 20.00%.
2. Option pool: negotiated to 10.00% of the post-round company, carved out before the new money prices in, so it comes from the existing holders, not the investor.
3. SAFE holders: their locked 15.83% is measured against the pre-round capitalization, then diluted by the new money like everyone else: 15.83% × (1 − 20%) = 12.67%.
4. Founders: what's left. 100% − 20% − 10% − 12.67% = 57.33%.
The founders walked into the round believing they owned the whole company minus "some SAFEs." They walk out with 57.3%. Total dilution from the round event: 42.7 points, of which the four quietly-signed SAFEs account for 12.67 points, more than the option pool and almost two-thirds of what the $3M lead investor took for actual new money.
And notice what the SAFE holders paid per point: $1.2M for 12.67% is an effective $9.5M post-money valuation on money that arrived when the company was worth far less. That's the deal working as designed: early risk, better price. The problem isn't the SAFEs. The problem is discovering the total at conversion instead of at signing.
Why the number sneaks up on founders
Three reasons this keeps happening:
- SAFEs don't appear on the cap table until they convert. The spreadsheet says the founders own 100% because it only tracks issued shares. The 15.83% overhang lives in four PDFs in four different folders.
- Each individual SAFE looks small. 2.5% here, 3.3% there. No single signature felt like a big decision, so no single signature triggered a modeling session.
- Different caps make mental math impossible. A stack at one cap is easy to total. A stack across $4M/$6M/$8M/$10M caps, some with discounts, some with MFN clauses, is a computation, not an estimate. (For how each mechanism works, see our SAFEs & convertibles lesson.)
Three rules of thumb
1. Model at signing, not at conversion. Before every SAFE, recompute the whole stack's combined conversion against your next expected round. The SAFE conversion calculator does one instrument; for a stack, sum the locked percentages of every post-money SAFE. That total is equity you've already sold.
2. Track total SAFE overhang as a standing number. Treat "sum of investment ÷ cap across all outstanding SAFEs" as a first-class metric, reviewed every time you consider new paper. Past roughly 20%, experienced leads start asking hard questions, and founders start getting surprised.
3. Know your conversion denominator. Post-money SAFEs convert against the pre-round capitalization including the pool top-up but excluding the new money; pre-2018 pre-money SAFEs and many convertible notes behave differently, and mixing forms makes hand math genuinely dangerous. If your stack mixes instruments, model it properly. The dilution math lesson walks the full layered computation.
This modeling is exactly what Vquity's rounds module exists for: it holds every outstanding SAFE next to the live cap table, shows the combined overhang continuously, and when the round comes, converts the whole stack with a tested YC-style post-money solver, pool top-up included, in one atomic close, so the number you modeled is the number that lands.
Before you sign the next SAFE
- Total the locked ownership of every outstanding post-money SAFE.
- Add the new SAFE's investment ÷ cap to that total.
- Run the combined stack against your next realistic round (valuation, new money, pool ask).
- Look at the founder percentage that comes out. If it surprises you now, it will hurt later.
- Watch the pool ask too, it's the other quiet lever. See the option pool shuffle.
SAFEs are a fine instrument. Unmodeled SAFEs are an expensive habit.