SAFE Note Stacking: How Multiple SAFEs Actually Dilute Founders (The Cap Table Math Most Guides Skip)

Multiple post-money SAFEs don't dilute each other — each one locks in a fixed percentage of the company relative to the money that existed when it was signed, and every SAFE you stack comes entirely out of the founders' side. But "post-money" doesn't mean "permanently protected": when your next priced round creates an option pool and brings in new investor money, your SAFE holders get diluted pro-rata right along with you, which is the part most explainers leave out.

If you've signed two or three SAFEs on your way to a priced round, you already have a real number sitting on your cap table even though no shares have been issued — most founders just haven't calculated it, or they've calculated it wrong by assuming their earlier SAFEs are untouchable once "fixed."

The two SAFE types, and why post-money won by default

A SAFE is a promise: cash now, equity later, at a price determined when the promise converts — usually at your next priced round. There are two ways to size that promise.

Pre-money SAFEs calculate the investor's ownership as a percentage of the company's valuation before the round. If you raise $1M at a $9M pre-money cap, you sold roughly 10% — but if you'd already sold another SAFE before that one, the earlier investor's percentage gets recalculated down, because the pre-money base they were measured against just got smaller. Pre-money SAFEs dilute each other. That compounding was one of the reasons Y Combinator retired this structure as its default template.

Post-money SAFEs, standard since YC's 2018 template revision, calculate ownership as a percentage of the company after the SAFE money is added — and that percentage doesn't move when you sign the next SAFE. Multiple post-money SAFEs sit side by side instead of stacking on top of each other. Every major international accelerator and most seed investors outside the US now default to post-money paper, which is why the rest of this article uses it as the base case.

The tradeoff: post-money is easier for an investor to model, and easier for a founder to track, but it means 100% of the dilution from every SAFE you sign lands on founders and existing shareholders. Nobody else absorbs any of it — not even the SAFE holders who came before.

The overhang: what your SAFEs already cost you, before conversion

Say you raise three post-money SAFEs over 18 months, structured with valuation caps only, no discount:

• SAFE A — $150,000 at a $3,000,000 post-money cap → 5.0%
• SAFE B — $250,000 at a $5,000,000 post-money cap → 5.0%
• SAFE C — $400,000 at an $8,000,000 post-money cap → 5.0%

Because each cap is well below where the company will eventually price its priced round, and because post-money SAFEs don't dilute each other, these three simply add: 15% combined SAFE overhang. That number exists the moment SAFE C is signed — before a single share has been issued, before you've even set a date for your Series A. It's the single most useful number a founder with multiple SAFEs outstanding can track, and it's the one most spreadsheets don't surface until conversion forces the question.

The "just add the percentages" shortcut is genuinely valid here — but only because every cap in this example sits below the eventual round price. That assumption breaks in one specific case, covered below.

What actually happens at the priced round

This is the part generic SAFE explainers skip, and it's the part that catches founders who think "post-money" means "locked forever." A post-money SAFE's percentage is fixed only relative to the capital that existed when it was signed. It is not fixed relative to capital that shows up later. Two things get added to the cap table immediately before your new investor's money comes in, and both dilute your already-converted SAFE holders exactly like they dilute you:

1. The SAFEs convert into real shares, based on the company's capitalization immediately prior to the round.
2. A new or topped-up option pool gets created, sized as a percentage of the post-round cap table but carved out pre-money — a mechanic we've covered in detail separately, because it has its own founder-unfriendly wrinkle.

Walk the numbers through. Before the round, treat the cap table as 100%: founders hold 85%, the three converted SAFEs hold the 15% overhang calculated above.

Step 1 — the pool. The Series A term sheet asks for a 10% post-money option pool, created pre-money. That pool doesn't come only from founders — it's carved pro-rata from everyone already on the table, founders and converted SAFE holders alike:

• Founders: 85% × 0.90 = 76.5%
• SAFE holders (combined): 15% × 0.90 = 13.5%
• New pool: 10%

Step 2 — the new money. The new investor buys 20% of the post-money cap table. That dilutes everyone else, including the pool and the SAFE holders, by the same factor:

• Founders: 76.5% × 0.80 = 61.2%
• SAFE holders (combined): 13.5% × 0.80 = 10.8%
• Pool: 10% × 0.80 = 8%
• New investor: 20%

The SAFE holders who signed at a fixed 15% combined now hold 10.8% — diluted by the exact same mechanics that diluted the founders, not exempted from them. If you pitched your early SAFE investors on "your percentage is locked in," you pitched them something that was only ever true relative to the round that didn't exist yet. Model this before you sign your next SAFE or negotiate your next term sheet, not after — an investor who assumed permanent protection and gets a smaller number than they expected is a harder conversation to have post-close.

The one case where "just add the percentages" breaks

Everything above assumes each SAFE's cap sits below the eventual round price, so every SAFE converts at its own cap and the shortcut of summing percentages holds. If your Series A prices below one of your SAFE caps, that specific SAFE converts at the round's actual share price instead — because SAFE conversion terms always give the investor the better of the two. When that happens, that SAFE's percentage stops being the fixed number you calculated at signing and starts depending on the round price, which means your overhang math needs to be rebuilt around actual round terms rather than the caps alone. This is rare for a company whose valuation is climbing between SAFEs, but it's exactly the scenario a down round or a slower-than-expected raise creates — worth checking explicitly rather than assuming it won't apply to you.

Where this gets sharper outside the US

Everything above assumes you're issuing an actual SAFE — a specific security that only cleanly exists under Delaware, Cayman, or Singapore corporate law, the three jurisdictions YC itself publishes templates for. If your company isn't incorporated in one of those three, what you're signing is often a locally adapted instrument that behaves like a SAFE but isn't one — sometimes a convertible note with interest and a maturity date bolted on, sometimes a jurisdiction-specific hybrid. That changes the stacking math, because notes accrue interest that adds to the conversion amount and carry a maturity date that can force a conversion or repayment event a SAFE never has. We've written separately about how SAFEs and convertible notes diverge outside the US — read that first if you're not sure which instrument you actually signed, because the stacking rules in this article only hold cleanly for true SAFEs.

Track the overhang, not just the individual agreements

The mistake isn't signing multiple SAFEs — it's tracking each one as a standalone agreement in a folder instead of a running percentage against your cap table. A founder who knows their combined SAFE overhang is 15% going into a term sheet negotiation is negotiating from a real number. A founder who's tracking three PDFs and adding up caps in their head, mid-negotiation, is not.

Govy models SAFE and convertible note overhang on the same ledger as your issued shares, so the combined dilution shows up before you sign the next agreement, not after your Series A lawyer builds the conversion spreadsheet for you. See how it tracks your SAFE stack at govy.tech.

FAQ

Do multiple post-money SAFEs dilute each other?

No, not directly. Each post-money SAFE's percentage is fixed relative to the company as it stood the moment that SAFE was signed, so a later SAFE doesn't shrink an earlier SAFE's slice. What it does shrink is the founders' side — every post-money SAFE you stack comes out of founder and existing-shareholder ownership, not out of other SAFE holders.

What is SAFE overhang?

SAFE overhang is the combined percentage of the company that all outstanding, unconverted SAFEs represent on a fully diluted basis. If you've signed three post-money SAFEs worth 5% each, your SAFE overhang is 15% — a number that exists the day you sign the paperwork, even though no shares have actually been issued yet.

Does a post-money SAFE protect an investor from all future dilution?

No. It only fixes that investor's percentage relative to the capital that existed when they signed. It does not protect them from the option pool or new investor money added at the next priced round — those dilute the SAFE holder pro-rata along with the founders, exactly like common stock. "Post-money" describes how the percentage was calculated at signing, not a permanent floor.

Should a founder outside the US use pre-money or post-money SAFEs?

Post-money is the default for a reason — it's what Y Combinator standardized in 2018, it's what most international investors expect, and it's easier to track because each SAFE's dilution doesn't depend on what you sign next. Pre-money SAFEs are rarer today and mainly show up in older templates or investor-specific paper; if you're offered one, model it separately since it compounds differently than a post-money stack.

What happens if my priced round prices below one of my SAFE caps?

That SAFE converts at the round's actual price per share instead of its cap, because SAFE conversion terms always give the investor whichever is better for them. This breaks the simple "just add up the post-money percentages" shortcut, because that SAFE's share count now depends on the round price rather than being fixed at signing — model it explicitly rather than assuming your overhang math still holds.