TL;DR
- Each SAFE converts independently at its own price — check size and ownership outcome are not proportional.
- Stacked SAFEs create far more dilution than founders expect, often making SAFE holders the second-largest ownership group after founders.
- Model your full SAFE stack before the next round — common mistakes consistently understate founder dilution.
- The time to fix cap table problems is between SAFEs, not after the term sheet arrives.
Quick navigation
If you have issued more than one SAFE, your cap table is more complex than any single SAFE would suggest. The dilution you will experience at your Series A is almost certainly larger than you have modeled.
That is not a hypothetical. It is how the math works. Each SAFE you issue sits as a separate convertible-securities line item until a priced round triggers conversion. When several SAFEs convert at once, they do not blend, average, or negotiate a group rate. Each one converts on its own terms. Each produces its own per-share price. Each produces its own ownership percentage. The result is a post-conversion cap table that almost always surprises founders—and rarely in a favorable way. New to these instruments? Start with our guide:
What is a SAFE: A Founder’s Guide to Simple Agreements For Future Equity.
Each SAFE Converts on Its Own Terms. There Is No Blended Rate.
When multiple SAFEs convert at a priced round, there is no averaging, no group rate, and no negotiated blend. Each SAFE converts at its own governing price — derived from its own valuation cap, its own discount (if any), and whether it is structured as a pre-money or post-money instrument. Every SAFE converts to preferred stock simultaneously, but each at a different per-share price. The investor with the lowest effective conversion price receives the most shares per dollar invested.
This is why the stack is uneven, not uniform. Three SAFEs issued at the same dollar amount but at different caps will produce three different ownership percentages on your post-conversion cap table — sometimes differing by several percentage points each. The check size and the ownership outcome are not proportional. The cap does most of the work. For how a single cap sets the conversion price, see our guide to the SAFE valuation cap.
The pre-money or post-money structure of each instrument also matters. Post-money SAFEs — the Y Combinator standard since 2018 — lock in the holder’s ownership percentage against later SAFEs and other converting securities. Pre-money SAFEs share dilution from later SAFEs proportionally with founders. These two forms behave differently within the same stack. Mixing them, as happens frequently when a company issues SAFEs across multiple fundraising cycles, produces a hybrid that neither form’s intuition captures and must be modeled line by line. Post-money caps create their own later-round dilution trap — see our guide to the post-money SAFE anti-dilution trap.
A Worked Example: Three SAFEs Converting at a Series A
Consider a company with 10 million founder common shares and three outstanding post-money SAFEs:
SAFE #1: $750,000 at an $8 million post-money cap, no discount (friends-and-family round)
SAFE #2: $500,000 at a $20 million post-money cap, 20% discount (small angel)
SAFE #3: $1,250,000 at a $12 million post-money cap, no discount (lead angel)
Total raised on SAFEs: $2.5 million.
The Series A: $4 million in new money at a $20 million pre-money / $24 million post-money valuation. To keep the focus on the stacking mechanics, this example assumes no existing option pool and no increase in the option pool at the Series A. A real Series A model requires both layers; the pool refresh would reduce founder ownership further and is addressed below.
| Holder | Instrument | Amount | Cap | Discount |
|---|---|---|---|---|
| Founders | Common stock | — | — | — |
| SAFE #1 | Post-money SAFE | $750,000 | $8M | — |
| SAFE #2 | Post-money SAFE | $500,000 | $20M | 20% |
| SAFE #3 | Post-money SAFE | $1,250,000 | $12M | — |
| Holder | Cap-Implied Price | Discount-Implied Price | Governing Price | Shares Issued | % Pre-Series-A FD |
|---|---|---|---|---|---|
| SAFE #1 | $0.6167 | n/a | $0.6167 (cap) | 1,216,216 | 9.375% |
| SAFE #2 | $1.5417 | $1.2333 | $1.2333 (discount) | 405,405 | 3.125% |
| SAFE #3 | $0.9250 | n/a | $0.9250 (cap) | 1,351,351 | 10.417% |
| Holder | Shares | % Fully Diluted |
|---|---|---|
| Founders | 10,000,000 | 64.24% |
| SAFE #1 | 1,216,216 | 7.81% |
| SAFE #2 | 405,405 | 2.60% |
| SAFE #3 | 1,351,351 | 8.68% |
| Series A | 2,594,595 | 16.67% |
| Total | 15,567,567 | 100.00% |
What the Numbers Tell You
SAFE #1 and #3: Cap, Not Check Size
Not the largest check in the stack, but at the lowest cap — ends with 7.81% of the company. SAFE #3, with more than 50% more capital deployed but at a $12 million cap, ends only modestly higher at 8.68%. The check size and the ownership outcome are not proportional. The cap governs.
SAFE #2
This illustrates how discounts and caps interact within a single instrument. Its $20 million cap was set at the Series A pre-money valuation, meaning it provided no advantage beyond what a Series A investor would have received at the round price. The 20% discount is what gave that investor a better deal — it dropped the conversion price from approximately $1.5417 to approximately $1.2333 and produced roughly 81,081 additional shares compared to a cap-only outcome. A discount only changes the result when the discount-implied price is lower than the cap-implied price. That relationship depends on the specific cap, the discount percentage, and the Series A valuation — it is not automatic.
The Stack in Aggregate
Read in aggregate, the picture is stark. You raised $2.5 million across three SAFEs. Before the Series A money is counted, that stack converts into approximately 22.9% of the pre-Series-A fully diluted cap table. After closing, the SAFE holders collectively hold approximately 19.1% of the company. Viewed in isolation, each instrument looked manageable. Stacked, they collectively form the largest ownership category on your post-Series-A cap table other than the founders themselves. Put in dollar terms: at the Series A’s $24 million post-money valuation, that 19.1% is worth about $4.6 million — and at a $100 million exit, every percentage point of founder ownership you failed to model is worth roughly $1 million.
How to Model Your SAFE Stack Before the Priced Round
The procedure for building an accurate, fully diluted capitalization table from a multi-SAFE stack is not complicated, but it must be executed instrument by instrument — not at the aggregate level:
- List every outstanding SAFE: amount, cap, discount, pre-money or post-money form, any side-letter rights, and any MFN clauses.
- Confirm which definition of “company capitalization” each SAFE uses. Post-money SAFEs lock in the holder’s ownership against later SAFEs and other converting securities. Pre-money SAFEs do not provide that protection.
- Select a realistically priced round scenario — pre-money valuation, new money in, and any expected option-pool increase. Model an optimistic case and a conservative case.
- For each SAFE, calculate two conversion prices: the cap-based price and the discount-based price (if the instrument carries a discount). The lower price governs.
- Compute the shares issued to each SAFE holder at its governing price.
- Add the Series A new shares and any option-pool top-up, then re-percent the full table.
- Compare founder ownership across your scenarios. If the spread exceeds a few percentage points, the SAFE stack is the source.
- Most off-the-shelf cap table calculators cannot handle a stack that mixes pre-money and post-money SAFEs, or instruments with side-lettered MFN rights. If your stack is large or mixed, the analysis is more reliably done by hand or in a custom spreadsheet built to your specific facts.
Four Cap Table Mistakes That Show Up in Practice
Mixed pre-money and post-money SAFEs in the same stack.
A company that issued one pre-money SAFE early and post-money SAFEs in subsequent rounds ends up with a hybrid stack that does not behave like either form. The modeling must be done line by line. Running it at the stack level produces the wrong answer — and the error compounds at the priced round.
An MFN clause that was never mapped.
MFN rights are not built into YC’s standard valuation-cap or discount SAFE forms by default. YC’s current SAFE library offers a separate uncapped MFN SAFE alongside cap-only and discount-only forms; MFN rights can also be added via a side letter or custom-drafted terms.
Where an MFN clause exists, the holder may be entitled to amend into the more favorable terms of a later SAFE — subject to the scope of the clause and any carve-outs. Founders must identify which holders have MFN rights before issuing the next instrument. Do not assume that side letters or pro rata rights from one round carry forward automatically.
Modeling the SAFE conversion without modeling the option pool.
Series A investors typically require a refreshed option pool sized at the post-money valuation. If the top-up is included in the pre-money valuation — the standard but negotiated structure — it dilutes every party in the cap table, founders and SAFE holders alike. Post-money SAFE holders are not insulated from the option-pool top-up. Omitting the pool refresh from your model consistently understates founder dilution by several percentage points. The worked example above isolates the stacking math by excluding the pool; a complete pre-Series-A model includes both.
Running the conversion math against a stale cap table.
The output is only as accurate as the input. The most common upstream errors are SAFEs that closed but were never formally added to the cap table, MFN amendments to earlier SAFEs that were not documented after a later SAFE was issued, and option-pool increases counted twice — once in the pre-money pool and again in the post-money refresh. Reconcile the cap table from primary documents before running the conversion model. Do not work from the last snapshot saved at the prior SAFE close.
When the Stack Is Large Enough to Require Counsel
There is no universally accepted threshold. In our practice, the following circumstances consistently warrant careful modeling before the next instrument is signed — and a conversation with startup financing counsel before the next term sheet arrives:
- Total SAFE dollars approach or exceed roughly 15 to 25% of the likely Series A pre-money valuation
- Three or more SAFEs are outstanding with materially different terms
- One or more SAFEs contain an MFN clause, and a later SAFE has since been issued
- The stack mixes pre-money and post-money SAFE forms
- A priced round is expected within roughly six months, and the conversion has not yet been modeled
Any one of these is sufficient reason to model carefully before signing the next instrument. Two or more is reason to engage counsel before the next term sheet arrives.
Ready to talk through your funding options?
Frequently Asked Questions
Do all my SAFEs convert at the same price?
No. Each SAFE converts at its own governing price, calculated from its own cap and discount. Two SAFEs issued to the same investor on the same day at different caps will produce different per-share conversion prices at the priced round. There is no blending and no averaging.
What happens if one SAFE has a discount and the others do not?
If a post-money SAFE includes both a valuation cap and a discount, the conversion price is the lesser of the cap-based and discount-based prices. The discount only yields a better outcome for the investor when the discount-implied price falls below the cap-implied price—a relationship that depends on the specific cap, the discount rate, and the Series A valuation. YC discontinued the combined cap-and-discount form from its standard library in 2021. Its current library offers cap-only, discount-only, and uncapped MFN forms. The cap-and-discount structure nonetheless remains common in pre-2021 SAFEs and in custom-drafted instruments. The other SAFEs in the stack convert at their own governing prices regardless.
Does an LLC need a cap table?
Yes — in substance, if not in form. LLCs track member units and economic interests rather than shares, and ownership modeling is no less important than it is for a corporation. Corporate SAFE math does not carry over cleanly, however. If your company is organized as an LLC and you are issuing convertible instruments, the analysis requires separate treatment. For the entity-specific issues, see our guide to SAFEs vs. convertible notes for LLCs.
Can I avoid stacking by using a single SAFE for all investors?
Sometimes — with side letters or a single-document close — but the same investor group will frequently end up with separate caps or discounts depending on when each investor signed. The cleanest way to avoid stacking is to limit the number of SAFE rounds and move to a priced round earlier. That is a fundraising-strategy decision, not a documents decision, and the right answer depends on your company’s stage, investor relationships, and timeline.
Model Your Cap Table Before the Next Round, Not After
The cap table is knowable today. Founders who model their SAFE stack before the priced round closes negotiate from a fundamentally different position than founders who do not. You know which terms move the math and which do not. You know where your leverage is before it expires.
Cap and discount terms are contractual. Once signed, they are difficult to reopen — possible in some circumstances, with the required holder consent, but not something to rely on. The decisions that protect founder ownership are made between SAFEs, before any of them lock in. After the term sheet arrives, the conversation shifts to valuation, round size, and option-pool mechanics. The SAFE terms are largely done.
If your stack is large, mixed, or has been issued across more than one fundraising cycle, that modeling is worth doing with counsel, both to verify the arithmetic and to identify which terms on the next instrument still have room to move. Our startup financing practice handles this work as part of pre-Series-A cap-table cleanup. Contact us to discuss your cap table before the next term sheet is on your desk.
