Stacked SAFEs: How Multiple SAFE Notes Convert at Your Series A

You raised 250K on your first SAFE in 2023 at a 10 million dollar cap. Six months later you closed another 500K on a SAFE at 12 million. A year after that, you took 1 million at a 15 million dollar cap because the company looked stronger. Now you are closing a Series A at a 25 million pre-money valuation and your investor's term sheet says "standard 10% option pool, post-money." You open your spreadsheet and the founder ownership number does not match what you thought it would be. It is 8 points lower.

Each SAFE looks harmless on its own. Together they compound when they convert, and cap-table software was never built to model four or five overlapping rounds, an MFN clause on the oldest one, and a lead investor demanding a post-money option pool top-up.

Below: how stacked SAFE conversion works, where the math breaks in most founder spreadsheets, and when you need a deal-specific waterfall to negotiate from a real number.

What Is a Stacked SAFE Situation?

A stacked SAFE situation exists when a company has three or more outstanding SAFEs from different rounds (different caps, different discounts, potentially different MFN clauses) that will all convert at the next priced round. By the time founders are negotiating a Series A in 2026, four to six outstanding SAFEs is common, especially among AI-native startups that raised on rolling SAFEs through 2024 and 2025.

Each SAFE was a separate negotiation, often with different lawyers and sometimes with different SAFE versions (pre-money, post-money, with or without MFN). They all convert to preferred stock at the same priced round, but their legal terms differ, and those differences are what cap-table software flattens out.

How Do SAFEs Convert at a Priced Round?

A SAFE converts to preferred stock at the lower of the cap-based price or the discount-based price calculated against the priced round. The cap sets a maximum valuation at which the SAFE converts; the discount (if any) gives the SAFE holder a percentage off the priced-round price. Whichever produces more shares for the SAFE investor wins.

For a single SAFE at a 10 million dollar cap converting in a 25 million pre-money Series A:

• Cap-based price per share: 10M / total shares outstanding before SAFE conversion
• Series A price per share: 25M / same denominator
• Effective discount the SAFE gets: 60% (because 10M cap is 40% of the 25M priced-round valuation)

That math is simple. The SAFE investor effectively bought their shares at a 60% discount to the Series A price. The complication starts the moment you have a second SAFE outstanding.

What Changes With Multiple Valuation Caps?

When multiple SAFEs convert together, each one uses its own cap to calculate its share count, and the total dilution is the sum of every SAFE's conversion plus the new Series A round plus the option-pool top-up. The SAFEs do not "share" a cap or "blend" into an average. They each fire independently against the same denominator.

A worked example with three SAFEs (all post-money, no discount, no MFN):

• SAFE 1: 250K invested at a 10M cap
• SAFE 2: 500K invested at a 12M cap
• SAFE 3: 1M invested at a 15M cap

If the Series A is priced at 25M pre-money on a 12M+ raise, every SAFE converts at its own cap. SAFE 1 takes 2.5% of post-money (250K / 10M). SAFE 2 takes about 4.2% (500K / 12M). SAFE 3 takes about 6.7% (1M / 15M). Combined, the three SAFEs absorb roughly 13.4% of the post-money cap table before the Series A money lands.

Add the 8M Series A check at 25M pre-money: that buys another 24.2% post-money. Add the 10% option-pool top-up, which on a post-money pool comes out of the founders before conversion. Add common stock holders (founders, existing employees, advisors).

The total math is rarely intuitive without a deal-specific waterfall. The most common founder error: assuming the SAFEs collectively absorb their headline dollar amounts (1.75M total) at the priced-round valuation (25M), yielding 7% combined dilution. The real answer in this example is closer to 13.4%, almost double.

Post-Money vs Pre-Money SAFEs: Why the Difference Matters

Post-money SAFEs (Y Combinator's 2018 update, now the dominant version) calculate the investor's ownership against the post-money cap including all other SAFEs and convertible instruments. Pre-money SAFEs calculate ownership against the pre-money cap before other SAFEs. The difference is consequential: post-money SAFEs preserve investor ownership through subsequent rounds, while pre-money SAFEs dilute proportionally as later SAFEs land.

Practically, this means: if you raised your first SAFE in 2017 as a pre-money SAFE and your last three as post-money SAFEs, the pre-money holder absorbs additional dilution every time another SAFE was issued. Most founders do not realize they hold a mix until conversion time.

Two questions to answer before any Series A modeling:

• Which version of the SAFE template did each round use? (The legal name is on the document. Pre-money SAFEs typically have "Valuation Cap" only; post-money SAFEs have "Post-Money Valuation Cap.")
• If pre-money, what was the pre-money definition at the time each SAFE was issued, and how has it shifted with each subsequent issuance?

If your earliest SAFE was pre-money and you have raised three more SAFEs since, that early investor's effective ownership is now lower than the original SAFE-conversion table on Carta likely shows. Sometimes much lower. The investor may also have rights to be made whole. Check the document.

How Do MFN Clauses Compound the Dilution?

A Most Favored Nation (MFN) clause lets an early SAFE investor upgrade to the best terms offered to any later SAFE investor. If your first SAFE has MFN at a 10M cap, and you later raise a SAFE at an 8M cap with a 20% discount, the first SAFE investor can swap their 10M cap (no discount) for the 8M cap with the 20% discount. They keep whichever is better at conversion time.

MFN is rare on single-SAFE rounds and increasingly common on stacked SAFEs in 2024 to 2026, because savvy early investors have learned to protect themselves against later rounds at lower valuations.

The compounding effect: each new SAFE you issue at a lower cap or with a discount can re-rate every prior SAFE that holds MFN. By the time you hit your fifth SAFE round, the early investors who started at a 10M cap may have re-rated themselves to a 6M cap with a 25% discount. The dilution they take at Series A is now substantially higher than what your original spreadsheet had.

This is the math cap-table software almost universally misses. The software shows you the SAFE's stated terms, not the upgraded terms after MFN exercise.

What Does Stacked SAFE Dilution Actually Look Like? (Worked Example)

The cleanest way to see how stacked SAFEs compound is to walk through a complete Series A conversion side by side: founder mental math, cap-table software output, and the deal-specific waterfall.

Assume a company entering a Series A with:

• 10M founder common shares outstanding
• 2M shares in the existing option pool (15% pre-money)
• SAFE 1: 250K at a 10M post-money cap, with MFN
• SAFE 2: 500K at a 12M post-money cap
• SAFE 3: 750K at a 12M post-money cap, 15% discount
• SAFE 4: 1M at a 15M post-money cap, 20% discount
• SAFE 5: 1.5M at a 15M post-money cap, MFN
• Series A: 8M at 25M pre-money, with a top-up to 12% post-money option pool

The headline founder math: total raised on SAFEs is 4M, total raised at Series A is 8M, total capital in is 12M. At 25M pre / 33M post, that is 36% post-money for new money. Existing common holders should retain about 64%. The founders own approximately 80% of common stock pre-round, so founder ownership "should" be roughly 51% post-Series A.

The actual answer, when run through a proper waterfall:

• MFN triggers on SAFE 1: upgrades to 12M cap with 15% discount (matched to SAFE 3 terms)
• MFN triggers on SAFE 5: upgrades to 15M cap with 20% discount (matched to SAFE 4 terms)
• SAFE conversions absorb roughly 21% of post-money (not the 12% the founder estimated)
• Option-pool top-up to 12% post comes out of pre-Series A holders (founders take the brunt)
• Founder ownership post-Series A: approximately 41%

That is a 10-point swing from the founder's spreadsheet estimate. On a successful exit five years later, those 10 points can be eight or nine figures of personal value. The same arithmetic runs again at exit, where liquidation preferences decide who actually gets paid.

By the time a startup has 5+ outstanding SAFEs, founder ownership cannot be modeled from a cap-table dashboard. It has to be calculated from every SAFE's legal terms applied in the correct order against a deal-specific scenario.

What Negotiation Levers Still Exist After 5+ SAFEs?

Once SAFEs are signed, their economic terms are locked. But founders still have material levers in the Series A term sheet that determine how much of the resulting dilution falls on common stock versus the SAFE holders.

The four highest-leverage Series A levers when you walk in with stacked SAFEs:

1. Pre-money option pool sizing. The lead's standard ask is a post-money pool top-up, which dilutes pre-Series A holders (founders + SAFE investors). Pushing for some portion of the pool to come out of post-Series A reduces the founder bite. Hard to win but often partial wins are possible.
2. Pre-money valuation. Every 1M of incremental pre-money reduces SAFE conversion ownership proportionally. Worth fighting harder when stacked SAFEs are present because the math compounds.
3. Option pool top-up timing. Whether the pool refreshes pre-conversion or post-conversion matters when SAFEs convert at different caps.
4. Conversion mechanics negotiation. Some lead investors will negotiate the conversion convention (e.g., whether MFN counts shares issued in the priced round as part of the "later SAFE" trigger). Rarely flexible but worth raising.

Walk into the Series A with the actual waterfall and you negotiate from a real number. Walk in with the Carta dashboard figure and you will likely accept a term sheet 5 to 10 points worse than you think it is.

What Are the Most Common Stacked SAFE Modeling Mistakes?

The damaging errors fall into five categories: ignoring MFN upgrades, mixing pre-money and post-money SAFEs as if they convert the same, modeling the option pool incorrectly, missing discount-vs-cap interactions, and treating cap-table software output as the ground truth.

1. Ignoring MFN clauses entirely. The most common error on this list. Founders look at the original SAFE terms and forget that MFN means those terms re-rated every time a later SAFE landed at better terms.
2. Treating pre-money and post-money SAFEs as interchangeable. Post-money SAFE ownership is fixed; pre-money SAFE ownership dilutes with subsequent rounds. Mixing them in a spreadsheet without distinguishing produces a wrong number.
3. Modeling the option pool top-up out of the post-money cap. The lead investor's "12% post-money option pool" typically comes out of pre-Series A holders, not post. Founders modeling it the other way understate their own dilution by 3 to 6 points.
4. Not running cap vs discount independently per SAFE. Each SAFE picks whichever produces more shares. Some SAFEs convert on the cap, some on the discount. Cap-table software often defaults to one or the other.
5. Trusting Carta or Pulley to model the priced round. These tools are excellent at recording the cap table. They are not built to model a priced round with stacked SAFEs, MFN, option-pool top-ups, and shadow preferred all firing simultaneously. They will produce a number. The number is rarely right.

When Do You Need More Than a Spreadsheet?

If you have three or fewer SAFEs all from the same template version with no MFN clauses, a careful spreadsheet can handle the math. Beyond that, the interactions exceed what a one-shot spreadsheet can model without errors.

What matters is the number of interacting legal provisions, not the count of SAFEs. Three post-money SAFEs at three different caps with no other complications will model fine in a spreadsheet. Add five SAFEs, two of them MFN, one pre-money, one with a participation right, and an option-pool top-up demanded by the lead, and neither a spreadsheet nor Carta will get it right.

This is the gap TILT's Capital Waterfall Model (CapFall) was built to close. It loads every security's actual legal terms, simulates the priced round with each holder applying their rights in the correct order, and produces a side-by-side comparison of founder mental math vs. cap-table software vs. the deal-specific waterfall.

Frequently Asked Questions