SAFE Note Dilution: How Valuation Caps and Discounts Actually Hit Your Cap Table

Downloading a SAFE template solves the paperwork problem. It does nothing for the math problem, and the math problem is where founders actually get hurt.

A SAFE (Simple Agreement for Future Equity) isn't equity at the time you sign it - it's a promise to issue equity later, at your next priced round, at a price set by whichever is more favorable to the investor: a valuation cap or a discount off that round's price. Both terms sound simple in isolation. What's not simple is what happens when you raise several SAFEs, at different caps, from different investors, over 12-18 months - which is exactly how most pre-seed rounds actually get built.

Pre-Money vs. Post-Money SAFEs: The Distinction That Matters

Y Combinator introduced the post-money SAFE in 2018, and it's now the standard template most founders use - but the shift changed who bears dilution risk, and a lot of founders never registered that change.

• Pre-money SAFE (the original 2013 template): The investor's eventual ownership percentage is calculated against the company's valuation before the SAFE money comes in, and that percentage gets diluted by any other pre-money SAFEs issued afterward, plus the option pool created at the priced round. The investor doesn't know their real final ownership until the priced round actually closes.
• Post-money SAFE (the current standard): The investor's ownership percentage is fixed at signing - a $500,000 investment at a $5,000,000 post-money cap means that investor owns 10% of the company on a post-money basis, full stop, regardless of what else happens before the next round. The certainty that pre-money SAFEs took away from investors, post-money SAFEs restore - by shifting the risk squarely onto the founders.

That shift is the whole story of why stacking SAFEs is dangerous.

The Stacking Problem, With Real Numbers

Because each post-money SAFE locks in a fixed ownership percentage, multiple SAFEs stack additively against the founders' share of the company - not against each other.

Here's an illustrative example (your attorney should model your actual cap table, since the precise mechanics depend on your SAFEs' specific capitalization definitions):

• SAFE #1: $500,000 at a $5,000,000 post-money cap → locks in 10% ownership
• SAFE #2: $500,000 at a $5,000,000 post-money cap → locks in another 10%
• SAFE #3: $500,000 at a $4,000,000 post-money cap → locks in 12.5%

Total raised: $1.5 million. Total dilution to founders before a single priced round happens: roughly 32.5% - not 10%, which is what a founder skimming any one term sheet in isolation might assume. Add a standard 10-15% option pool created at the Series A, and founders can be looking at 45%+ combined dilution before the new Series A investor's own check is even factored in.

This is the single most common surprise in pre-seed fundraising: founders raise a "small" SAFE round in three or four tranches, each one individually reasonable, and only discover the cumulative dilution when their attorney or a lead investor builds the actual pro forma cap table ahead of the priced round.

Valuation Cap vs. Discount: Which One Actually Controls

Most SAFEs carry both a cap and a discount (commonly 15-20%), and the SAFE converts at whichever gives the investor more shares - meaning whichever produces the lower effective price per share:

• If your Series A prices well above your SAFE caps, the caps control conversion, and early investors get a steep effective discount relative to the new round's price.
• If your Series A prices near or below your caps, the discount is what ends up mattering instead, since the cap isn't doing any work.

A "most favored nation" (MFN) clause, common in early pre-seed SAFEs, adds another layer: it lets an earlier investor automatically upgrade to match the best terms given to a later investor. If you give SAFE #3 a lower cap than SAFE #1 carried an MFN clause, SAFE #1 may now convert at SAFE #3's better terms too - which means a single generous term sheet can retroactively increase dilution from every prior SAFE with an MFN provision.

What to Actually Track Before You Raise Another SAFE

• Build a live, fully-diluted pro forma cap table that includes every SAFE issued to date, at its actual cap and discount - not just the headline amount raised. A spreadsheet that only tracks dollars raised, not resulting ownership percentage, will not show you the stacking problem until it's too late to renegotiate.
• Check every SAFE for an MFN clause before issuing a new one at better terms - a low cap on your next SAFE can silently reprice every earlier MFN-linked SAFE.
• Model the option pool separately from SAFE dilution. The pool created at your priced round dilutes founders and existing SAFE holders together, on top of whatever the SAFEs already convert to - it's not a separate, smaller bucket.
• Know which capitalization definition each SAFE uses. Post-money SAFEs vary in whether they count other outstanding SAFEs and the future option pool inside "Company Capitalization" - small differences in this definition materially change the conversion math.
• Loop in counsel before your third SAFE, not after your Series A term sheet arrives. Stacking problems are cheap to fix when you still control the terms of the next SAFE, and expensive to fix once a priced round is negotiating against a cap table you didn't fully understand.

The Bottom Line

A SAFE's simplicity is real for the paperwork and misleading for the math. Each individual SAFE looks like a small, founder-friendly instrument. The cumulative effect of several of them, especially with MFN clauses in play, is where founders lose track of what they've actually sold - and by the time a Series A term sheet forces a full cap table into view, the terms of the SAFEs are already locked.

---

Raising a SAFE round or preparing for a priced round? Talking Tree offers AI-powered contract review and connects founders with experienced startup attorneys through Find Counsel. See our guides on what a SAFE note actually is, negotiating your funding round, and splitting co-founder equity.