I raised our seed round over fourteen months, not fourteen days. Three SAFEs, three different valuation caps, and by the time I finally sat down to model our post-money cap table before the priced round, I found out I'd given away almost twice what I thought I had.
That gap is not a rounding error. It is the single most common way early-stage founders under-price their own dilution, and almost nobody catches it until the priced round forces the math.
Why SAFEs feel cheap until they aren't
SAFEs took over pre-seed and seed fundraising for good reasons. Roughly 88-92% of pre-seed rounds today are structured as unpriced instruments rather than priced equity, because a SAFE is fast and cheap: a standard template runs a few thousand dollars in legal fees and closes in days, skipping the valuation negotiation entirely. Compare that to $15,000-$30,000 in company-side legal fees for a standard priced seed round, plus weeks of term sheet back-and-forth, and it's obvious why founders default to SAFEs for anything raised in pieces.
The problem isn't the SAFE itself. It's what happens when you raise more than one, at more than one cap, over more than a few months.
The mechanic almost nobody explains clearly
Since Y Combinator moved its standard template from pre-money to post-money SAFEs in 2018, each SAFE dilutes only the founders and existing shareholders, not the earlier SAFE holders. That single change shifted real dilution outcomes on seed cap tables by something like ten percentage points, and it means every additional SAFE you stack on top of an earlier one comes directly out of your ownership, not out of a shared pool.
Here's the part that catches founders off guard: because post-money SAFEs don't interact with each other, you can't just add up the percentages each investor thinks they're buying and assume that's what you're giving away. A $500K SAFE at a $6M cap feels like roughly 8%. Layer a second $500K SAFE at an $8M cap on top of it three months later, and that second SAFE doesn't dilute the first, it dilutes you. By the time you've closed a third tranche, you're often 3-6 percentage points more diluted than the sum of the individual cap math suggested, before a single priced round has happened.
This is exactly what happened to me. Three SAFEs I'd mentally priced at roughly 22% combined dilution actually converted at just under 31% once I modeled them together at the priced round. Nobody had lied to me. I had just never modeled it as a stack.
The data backs up how common this is
Median founding teams own about 56.2% of their company right after a seed round, and that number drops to roughly 36.1% by Series A. Some of that is the priced round itself. A meaningful chunk of it, in my experience and in conversations with a dozen other founders who raised in tranches, is SAFE stacking that nobody modeled properly until it was too late to renegotiate.
The three-question test before you close SAFE number two
Before you accept a second or third SAFE, run this test.
- What's my fully-diluted ownership if every open SAFE converts today, at its own cap, simultaneously? Not sequentially, not on a napkin. Build a real fully-diluted cap table spreadsheet, free templates from Carta or YC work fine, and convert every outstanding SAFE at once, the way your priced round will.
- Are my caps trending up fast enough to offset the additional dilution? A rising cap protects you somewhat, but only if it's rising faster than your raise size is growing. A $500K SAFE at a $10M cap after a $500K SAFE at a $6M cap isn't automatically better; model both scenarios before assuming the higher cap saved you anything.
- Would a small priced round right now actually be cheaper than a third SAFE? Past a certain raise size, often once you're past $1.5-2M in cumulative SAFEs, the legal cost gap between a SAFE and a priced round shrinks relative to how much clarity you get. A priced round dilutes every existing shareholder proportionally, including your earlier SAFE and note holders, instead of concentrating all of that dilution onto the founders alone.
What I'd tell myself before that first SAFE
Model the stack before you sign the second one, not after. Ask your lawyer for the actual conversion math at your specific caps, not the generic "SAFEs are founder-friendly" explanation everyone gives on social media. And if you're raising in more than two tranches, put a hard ceiling on how many SAFEs you'll layer before you force a priced round, because the instrument that felt cheap and fast at $500K starts quietly compounding against you by the third close.
The SAFE didn't cost me equity. Not modeling it as a stack did.