A SAFE is a simple agreement for future equity. It is a short document, usually one page, that says an investor gives a founder cash today in exchange for shares at a later priced round. No interest, no maturity date, no covenants. The trade is speed and simplicity now in exchange for a fixed claim on the cap table later. In 2026, SAFEs are the default instrument for pre-seed fundraising in the United States, used in 92% of all pre-priced rounds in Q3 2025 according to Carta. The mechanics behind that one-page document, however, are not simple. This guide explains how SAFEs actually work, where the math hides, and the four mistakes that turn a casual SAFE round into a Series A surprise.
What a SAFE Actually Is
Y Combinator introduced the SAFE in late 2013 as a replacement for the convertible note. The original goal was to remove the parts of a convertible note that did not fit early-stage startups: interest rates, maturity dates, repayment risk if no priced round happens. A SAFE has none of those. It is a contractual right to receive shares when a priced round (a Series A or Seed equity round) is closed. Until that round, the SAFE just sits on the cap table as an obligation.
In 2018, Y Combinator released a major update: the post-money SAFE. The change matters more than the name suggests. Under the original pre-money SAFE, every SAFE holder in a round diluted every other SAFE holder along with the founders. Under the post-money SAFE, the SAFE holder's ownership percentage is locked in at the moment of signing. Future SAFEs and the priced round dilute the founders, not the SAFE holder. Carta data shows that over 85% of SAFEs signed in 2024 were post-money, with that share continuing to rise through 2025.
This is the single most important fact a first-time founder needs to internalize. A post-money SAFE is not a IOU. It is a fixed slice of your company sold today, with the math hidden inside a one-page form.
How the Math Actually Works
Every post-money SAFE converts using one equation: divide the SAFE investment amount by the post-money valuation cap. The result is the SAFE holder's ownership percentage on a fully diluted basis. A $500,000 SAFE on a $5 million post-money cap locks in exactly 10% ownership. A $1 million SAFE on a $10 million cap locks in exactly 10%. The investor knows their stake at signing. So can the founder, but most first-time founders never run the math.
Here is what that math looks like in a real Series A scenario. A founder raises $2 million on a single post-money SAFE with a $10 million cap. SAFE ownership is locked at 20%. Founders pre-Series A own the other 80%. The Series A lead invests $5 million on a $25 million post-money valuation, taking 20%. The term sheet also requires a 10% option pool to be created post-closing. New money plus the pool consume 30% of the post-money cap table. The remaining 70% is split between existing holders. The SAFE holder gets 70% of their original 20% slice (14%). The founders get 70% of their original 80% slice. That is 56%. Going from 100% pre-SAFE to 56% post-Series A is normal. Going from 100% to 35% is what happens when SAFEs are stacked without modeling.
In 2025, median valuation caps on post-money SAFEs hovered around $10 million for rounds in the $250,000 to $1 million range, and $15 million for rounds in the $1 million to $2.5 million range. For sub-$250,000 rounds the median cap was $7.5 million in Q2 2025, up from $6.5 million in Q1 2025. The numbers are rising, but the structure is what matters. A higher cap protects founders. A lower cap, multiplied across multiple SAFE rounds, compounds dilution faster than founders expect.
Cap, Discount, MFN: The Three Levers
A SAFE has three economic terms that change how it converts. Most pre-seed SAFEs use only one of them. The combinations matter.
The valuation cap is the ceiling on how the company can be valued at conversion. If the priced round happens at a $20 million valuation but the SAFE has a $10 million cap, the SAFE converts as if the company were valued at $10 million. This means the SAFE holder gets twice as many shares as they would at the actual round valuation. The cap is the protection an early investor gets for taking on more risk than the priced-round investor.
The discount is a percentage reduction on the per-share price at the priced round. A 20% discount means the SAFE holder pays $0.80 per share when the priced-round investor pays $1.00. Discounts are typically 10 to 25%. They reward early investors for committing before terms are set.
The MFN clause, or Most Favored Nation, says that if the founder later issues another SAFE on better terms, the original investor automatically gets those better terms. If a $10 million cap SAFE is followed by a $5 million cap SAFE, the original investor's cap drops to $5 million. MFN-only SAFEs (no cap, no discount) are common in very early angel checks before a round is priced.
Cap-only is the dominant structure by a wide margin. Carta data summarized in a Crunchbase News analysis by Kong and Michelman (news.crunchbase.com/venture/startup-funding-safes-kong-michelman) puts the breakdown at roughly 62% cap-only, 29% cap-and-discount, 9% discount-only, and 1% uncapped. Cap-with-no-discount is the cleanest math: the investor knows their exact ownership at signing. A $1 million raise on a $6.7 million post-money cap gives the investor exactly 14.9%, with no further calculation needed at conversion.
SAFE vs Convertible Note: Which One You Actually Have
Most first-time founders ask about SAFEs and convertible notes as if they were interchangeable. They are not. The difference is in three contractual mechanics that change what happens if a priced round never closes, or closes much later than expected.
Term | SAFE | Convertible Note |
|---|---|---|
Legal form | Equity-style claim. Not a debt instrument. | Debt instrument with a promise to repay. |
Interest rate | None. | Typically 5 to 8% annually, accruing until conversion. |
Maturity date | None. SAFE sits indefinitely until a priced round. | Yes, usually 18 to 36 months. If no priced round closes, the note is technically due. |
Repayment risk | Investor cannot demand cash repayment. | Investor can demand repayment at maturity. In practice this rarely happens, but legally it is on the table. |
Conversion mechanics | Fixed at the cap (post-money) or pro-rata to the round (pre-money). | Converts at cap or discount in qualified financing. May convert into priced round, or roll. |
Document length | One to two pages. | 6 to 15 pages, with covenants. |
Founder's downside if no priced round | Dilutes at the next equity event, whenever that happens. | Default risk, investor pressure for repayment, potential personal exposure depending on covenants. |
SAFEs comprised 90% of pre-seed deals on Carta in Q1 2025 and rose to 92% in Q3 2025, up from 88% in Q3 2024. Convertible notes at the pre-seed stage have correspondingly collapsed to 9% in 2025 by Carta count. They have not disappeared from the wider startup ecosystem. Outside pre-seed, convertible notes still appear in seed-stage bridge rounds and in larger growth-stage financings, but at pre-seed the SAFE has clearly won.
For a first-time founder raising under $2 million for the first time, the SAFE is almost always the right instrument. It is faster, cheaper to issue, has no maturity risk, and aligns with what investors expect. The math problem is what the founder fails to model, not the legal structure.
The Four Mistakes First-Time Founders Make
Most first-time founders sign SAFEs without modeling. The following four mistakes account for the majority of cap-table cleanup conversations at Series A.
Mistake 1: Treating the cap as a future floor. Founders often interpret a $10 million post-money cap as "my company will be worth at least $10 million next round." It is not. The cap is the ceiling for the SAFE holder's conversion, nothing more. If the next priced round happens at a $7 million pre-money valuation, the SAFE still converts at the cap. The cap protects the investor on the upside, not the founder on the downside.
Mistake 2: Stacking SAFEs at different caps without tracking the math. Each SAFE feels small. Three $500,000 SAFEs at a $5 million cap, a $7 million cap, and a $10 million cap all look reasonable in isolation. At conversion, each one locks in a different ownership percentage and the founder bears all dilution. The cumulative effect can be 30 to 50% of the company sold before the priced round even arrives. Median dilution from large SAFE rounds in 2025 hovered around 20%, but stacked rounds can easily double that.
Mistake 3: Underestimating the option pool top-up. Series A term sheets almost always require an option pool refresh of 10 to 15% before the priced round. That pool is created by diluting existing shareholders, including the founders, but not the new money. Carta's Founder Ownership Report finds that median founders hold around 56% of their company at the seed stage but only 36% by the time the Series A closes. That 20-point drop is the sum of three layers hitting on the same day: SAFE conversion, new money, and the pool refresh. Stacked SAFEs at low caps drag the post-Series A number well below the 36% median.
Mistake 4: Ignoring the post-money anti-dilution mechanic. Post-money SAFEs have built-in anti-dilution: any SAFE issued after the original SAFE dilutes only the common stock (founders and employees), not the existing SAFE holders. This becomes dangerous when a founder takes a follow-on SAFE at a lower cap because cash ran out. The new SAFE dilutes the founder twice: once for the new SAFE itself, and once again because the original SAFEs are insulated from the new round. Founders forced to take a down SAFE in a tight market can lose more equity than they would in a clean priced down round.
Why "Sign and Move On" Fails Investors Too
The reason most first-time founders sign SAFEs without modeling is that the document presents itself as a handshake. The investor wires cash. The founder signs. Both move on. The math sits dormant for 18 months. At Series A, both sides discover the cap table looks different than they assumed.
This is not just a founder problem. Series A investors do not want to walk into a deal where the founders own less than 30% and feel betrayed by their own SAFE round. Misaligned cap tables create friction at exactly the moment a startup needs to be moving fast. A Crunchbase News analysis by Ling Kong and Sanford Michelman (news.crunchbase.com/venture/startup-funding-safes-kong-michelman) warns that founders often underestimate dilution with post-money structures and that stacking SAFEs leads to cleanup provisions, legal rework, and investor friction at later rounds.
The fix is to model every SAFE before it is signed, share that model with the SAFE investor, and update it as more SAFEs come in. The founder who walks into Series A with a current cap table, a clean SAFE conversion model, and a clear understanding of what each percentage point means earns time back in negotiation. The founder who shows up surprised loses leverage at the worst possible moment.
What Investors Are Actually Testing at Conversion
The SAFE itself is just paper. What investors test at conversion is whether the founder built a real business on top of it. Three things matter.
The first is whether the founder ran the math. A founder who can pull up a cap table and walk an investor through every SAFE conversion, every option pool top-up, and every dilution event signals operational maturity. A founder who fumbles the math signals a future problem.
The second is whether the founder used the SAFE money to create proof. A SAFE is bridge capital for the moment between idea and traction. The expectation is that by Series A, the founder has converted that capital into evidence: customers, revenue, retention, defensible product surface. If the SAFE money funded burn without proof, the cap table is a secondary concern.
The third is whether the founder understands what was sold. The post-money SAFE locks in a specific ownership percentage. That percentage will dilute again at the priced round. A founder who treats the SAFE as a casual transaction misses the point that they sold a fixed slice of the company. A founder who treats it as a structured equity sale, modeled and tracked, walks into Series A with the right framing.
The Proof Layer
Every fundraise compresses two things into a single conversation: the cap table math and the operating proof. The math is what a SAFE locks in. The proof is what an investor decides to trust. First-time founders almost always over-index on the math (the document, the cap, the discount) and under-index on the proof (what the company actually is and what evidence supports the claims in the deck).
SeedForge was built around this mismatch. A 30-minute AI session with a founder produces a structured Living Profile that an investor can review before the first call. The investor sees the founder's real answers to the questions they would have asked anyway: how the first ten customers were acquired, what the revenue trajectory looks like, what breaks if a competitor copies the product. The SAFE conversation moves from "trust me, sign this" to "here is what the company is, here is what the cap table will look like at Series A, here is the math." The proof comes first. The paper follows.
This is the inversion of the usual pre-seed dynamic. Most first-time founders raise on a SAFE because they have nothing else to offer. SeedForge replaces "nothing else" with structured, sourced proof, accessible at one link. The SAFE becomes a financing instrument, not a substitute for substantive evidence.
Practical Checklist Before You Sign a SAFE
A first-time founder should never sign a SAFE without doing the following six things.
Build a fully diluted cap table model in a spreadsheet or in Carta, Pulley, or another cap table tool. Include every SAFE you have signed, every SAFE you plan to sign in this round, and a placeholder Series A. Run the conversion math.
Decide your target Series A founder ownership. Carta's Founder Ownership Report finds that median founders hold about 56% of their company at the seed stage and 36% by the time the Series A closes. The 36% median is pulled down by founders who raised too many SAFEs at the wrong caps. Target the top of the range, not the median.
Set a maximum total dilution budget for the entire SAFE round (across all checks). A common target is 15 to 20%. If the math says you would need to give away 35% to raise the round at the caps you are being offered, the round is mispriced for you.
Pick one SAFE structure (cap only is the cleanest) and stick to it for the entire round. Mixing cap-only, cap-and-discount, and MFN-only SAFEs in the same round creates conversion math that is painful to track at Series A.
Read the side letter, if any. Some investors negotiate side letters with pro-rata rights, information rights, or other terms that do not appear in the standard SAFE. Those rights survive into Series A and can affect future negotiations.
Engage a startup lawyer for the first SAFE round, not the third. The cost is modest. The value is that the lawyer will catch the structural mistakes before they compound.
Frequently Asked Questions
What is a SAFE note in plain English? A SAFE is a one-page agreement where an investor gives a startup cash today in exchange for shares at the next priced round. No interest, no maturity date, no debt. The investor's ownership is fixed at the post-money valuation cap, divided by the investment amount. A $500,000 SAFE on a $5 million cap converts to 10% of the company.
What is the difference between a SAFE and a convertible note? A SAFE is an equity claim with no interest, no maturity, and no repayment. A convertible note is a debt instrument with interest (usually 5 to 8%) and a maturity date (usually 18 to 36 months). At pre-seed, SAFEs reached 92% of rounds in Q3 2025 per Carta, while convertible notes dropped to 9%. SAFEs win at pre-seed; convertible notes still appear in seed-stage bridges and later-stage financings.
What is the difference between a pre-money and a post-money SAFE? A pre-money SAFE measures the SAFE holder's ownership before all the SAFE money is counted, so multiple SAFEs in a round dilute each other. A post-money SAFE locks the holder's ownership in after all SAFE money but before the priced round. Post-money is now the standard: over 85% of all SAFEs signed in 2024 were post-money, with that share continuing to rise through 2025. The trade-off is that founders bear all dilution from subsequent SAFEs and the priced round, not the SAFE holders.
How much dilution should I expect from a SAFE round? A clean single SAFE round of $1 to 2 million on a healthy cap typically dilutes founders by 15 to 20%. Stacked rounds at different caps can easily double that. The Crunchbase News analysis by Kong and Michelman (news.crunchbase.com/venture/startup-funding-safes-kong-michelman) cites Carta data showing median dilution from large SAFE rounds hovers around 20%. The compounding effect at Series A (SAFE conversion plus new money plus 10% option pool refresh) is what takes founders from a typical 56% post-seed ownership to a 36% post-Series A median, per Carta's Founder Ownership Report. Stacked SAFEs at low caps push that final number well below 36%.
What is a valuation cap on a SAFE? A valuation cap is the maximum company valuation at which the SAFE will convert into shares. If the priced round happens at a higher valuation than the cap, the SAFE converts as if the company were worth the cap. In 2025, Carta data showed median caps of $7.5 million for sub-$250,000 rounds (Q2 2025, up from $6.5 million in Q1), $10 million for $250,000 to $1 million rounds, and $15 million for $1 to $2.5 million rounds. The cap protects the investor on the upside; it does not set a floor for the founder.
Are SAFEs safe for first-time founders? SAFEs are the right instrument for most pre-seed rounds, but only if the founder models the dilution math before signing. The mistakes that hurt founders are not in the SAFE itself; they are in stacking multiple SAFEs at different caps without tracking the cumulative dilution. The post-money structure protects the SAFE investor by design, which means the founder bears the full burden of subsequent rounds. Model first, sign second.