Guide

SAFE agreements and startup seed financing explained

Harbor Labs closed its seed round without a priced equity round. Four angels and one micro-fund signed post-money SAFEs (Simple Agreements for Future Equity) totaling $1.8 million: three checks at a $12 million valuation cap with a 20% discount, one at $10 million cap only, and a $400k uncapped MFN (most-favored-nation) note from an accelerator. Eighteen months later, the company raised Series A at a $24 million post-money valuation. Finance modeled founder dilution at 18%; after conversion the actual hit was 26%. The gap came from stacking SAFEs issued at different caps, misunderstanding how post-money SAFE ownership is fixed at signing, and forgetting that the 20% discount applied to the lower of cap price or round price — not both simultaneously in the way the CEO assumed.

A SAFE is a standardized contract published by Y Combinator that lets startups raise cash today and issue equity later, typically at the next priced venture capital round. Unlike a loan, it has no maturity date or interest. Unlike priced preferred stock, it does not set a valuation until conversion. This guide covers SAFE variants, valuation caps and discounts, post-money vs pre-money mechanics, conversion math at Series A, stacking multiple SAFEs, side letters (pro-rata, MFN), the Harbor Labs cap-table walkthrough, a financing-instrument decision table, pitfalls, and a founder checklist before you sign.

What a SAFE is (and what it is not)

Y Combinator introduced the SAFE in 2013 to replace slow, lawyer-heavy seed documents. The investor wires cash; the company promises shares when a qualifying equity financing occurs (almost always a priced preferred round above a minimum size). Until then, the SAFE holder has no voting rights, no board seat, and no guaranteed return of capital.

Critically, a SAFE is not debt. If the company shuts down without a priced round, SAFE holders are treated as common-equity claimants in dissolution — usually behind creditors and often receiving nothing. It is also not preferred stock: provisions like liquidation preference and anti-dilution attach at conversion, when SAFEs typically become the same series as new investors (or shadow preferred with equivalent economics).

The standard YC forms include: valuation cap only, discount only, cap + discount, and MFN (no cap, no discount) with a promise to match better terms offered later before the next priced round.

Valuation cap and discount mechanics

The valuation cap sets a maximum effective pre-money valuation for conversion. If Series A prices the company higher than the cap implies, the SAFE investor converts as if the company were worth the cap — receiving more shares per dollar invested. The discount (typically 10–30%) gives a lower price per share than Series A investors pay, rewarding early risk even when the round valuation is modest.

When both cap and discount are present, the investor converts at whichever method yields the lower price per share (better for the investor). Example with Harbor Labs’ $500k SAFE (cap $12M, 20% discount) converting into Series A at $1.20/share (implied $24M post-money on 20M fully diluted shares):

Cap price: $12M cap ÷ pre-money share count before new money ≈ lower share price than round price when the company outgrows the cap.
Discount price: $1.20 × (1 − 0.20) = $0.96/share.
Winner: whichever is lower — here the discount often wins at moderate step-ups; at large step-ups the cap wins.

Founders frequently mis-model by applying cap and discount sequentially. The documents do not work that way.

Post-money vs pre-money SAFEs

YC revised the standard SAFE in 2018 to a post-money form. Under post-money SAFEs, each investor’s ownership percentage is defined at signing:

Investor ownership = Investment ÷ Post-money cap

A $500k investment on a $12M post-money cap equals 4.17% of the company immediately upon conversion, before the Series A primary shares are counted. Multiple SAFEs on the same cap stack additively consume the option pool and founder slice — a $1.8M raise on a $12M cap can represent 15% of the company before Series A investors buy a single share.

Pre-money SAFEs (older or custom forms) dilute differently: SAFE conversion shares are calculated before new round money, and stacking is harder to eyeball. Most US seed rounds today use post-money SAFEs, but European and corporate investors sometimes insist on pre-money or priced seed extensions. Always read the specific form version in the PDF header.

Conversion at a priced equity round

Conversion triggers on an Equity Financing — usually selling preferred stock above a dollar threshold with a lead investor. The sequence at Harbor Labs Series A ($6M new money, $24M post-money):

Calculate fully diluted share count before the round (founders, employees, prior SAFEs if already converted, option pool).
Convert each SAFE into shares at its cap/discount price; post-money SAFE holders take their contractually fixed percentages first.
Issue new Series A preferred to incoming investors at the round price.
Refresh or expand the option pool if the term sheet requires (often 10–15% post-close).

Harbor’s $400k MFN SAFE had no cap; when a later angel received a $10M cap, the MFN holder automatically inherited that cap. Finance had modeled it as uncapped common-equity risk; the MFN match added 4.0% instead of the modeled 2.5%, shifting another 1.5 points of dilution onto founders. MFN clauses are convenient for accelerators but dangerous if you forget to track them.

Stacking SAFEs and side letters

Seed rounds are rarely one check. Each SAFE is a separate contract with its own cap, discount, and side terms. Common add-ons:

Pro-rata rights — the investor may invest their ownership percentage in the next priced round. Useful for angels who want to maintain stake; adds cap-table complexity when many small pro-rata notices arrive at Series A.
MFN — matches better economic terms granted to later SAFE investors before conversion. Reduces negotiation friction early but can ratchet economics against founders silently.
Information rights — quarterly financials; low burden at seed but sets expectations.
Board observer — rare at true seed, more common when a single investor contributes >25% of the SAFE round.

Maintain a SAFE ledger from day one: investor, wire date, form version, cap, discount, MFN status, pro-rata Y/N, and modeled ownership at conversion. Update after every new signature before you quote a Series A lead on available founder ownership.

Harbor Labs cap-table walkthrough

Simplified pre-Series A (post-money SAFEs only, 10M founder shares, 2M option pool authorized unissued):

Investor	Amount	Terms	Implied % at conversion
Angel A	$500k	$12M cap, 20% disc	4.17%
Angel B	$500k	$12M cap, 20% disc	4.17%
Micro-fund	$400k	$12M cap, 20% disc	3.33%
Angel C	$400k	$10M cap only	4.00%
Accelerator	$400k	MFN → matched $10M cap	4.00%

Total SAFE overhang: 19.67% before Series A. After $6M primary at $24M post-money, founders dropped from 100% nominal to roughly 54% fully diluted (not 72% as the CEO projected before the ledger review). The lesson: post-money SAFE percentages are not “maybe” dilution — they are committed unless investors waive conversion terms (extremely rare).

Technique decision table

Instrument	Best for	Trade-offs	When to choose
Post-money SAFE	Fast seed checks; transparent % ownership	Stacking dilution surprises; no investor governance until conversion	US-style pre-seed/seed with many angels; standard YC ecosystem
Convertible note	Investors wanting maturity, interest, or downside protection	Debt on balance sheet; maturity forcing events; legal cost	Bridge between A and B; investors uncomfortable with pure SAFE risk
Priced seed (preferred)	Lead investor sets governance early; larger single check	Slower close; full term sheet (board, prefs, legal fees)	$2M+ seed with institutional lead demanding control
Revenue-based financing	Non-dilutive growth capital with predictable payback	Cash drain; not equity; unsuitable for pre-revenue	SaaS with ARR; founders avoiding any dilution at seed
Convertible bonds	Later-stage corporate debt with equity optionality	Public-market covenant complexity; not a seed tool	Mature companies; not startup seed

Common pitfalls

Ignoring SAFE overhang at Series A — leads price rounds assuming 100% minus option pool; SAFE conversion happens first.
Mixing pre-money and post-money forms — cap-table models become unreliable; standardize one form per round.
Uncapped SAFEs without board approval — uncapped investors convert at the round price with no reward for early risk; fine for friends/family, expensive for strangers.
MFN amnesia — granting a better cap to a later angel silently upgrades every MFN holder.
Discount-cap double counting — modeling both benefits stacked instead of the better-of election.
Pro-rata sprawl — twenty angels with pro-rata can block or complicate Series A allocations.
No qualifying-round threshold — custom SAFEs that convert on any issuance can trigger accidental conversion on advisor shares.
Securities law oversight — SAFEs are securities; 506(b)/506(c) rules, blue-sky filings, and accredited-investor verification still apply.

Founder checklist

Confirm YC form version (post-money vs pre-money) before countersigning.
Model total SAFE ownership as sum of (investment ÷ post-money cap) for each holder.
Track MFN and pro-rata side letters in the same ledger as primary SAFE terms.
Run conversion math at three Series A price scenarios before quoting a lead.
Reserve option pool expansion needed for Series A term sheet in dilution models.
Verify accredited-investor status and file Form D within 15 days of first sale.
Align with counsel on state blue-sky notices and any foreign investor restrictions.
Negotiate a single cap for the round when possible instead of one-off sweeteners.
Plan SAFE conversion logistics in the Series A closing checklist (signatures, wire timing).
Educate co-founders: SAFE dilution is real at signing, not hypothetical.

Key takeaways

SAFEs defer valuation and equity issuance until a priced venture round.
Post-money SAFEs fix investor ownership % at signing: investment ÷ cap.
Cap and discount convert at the better (lower) price per share, not both stacked.
Stacking multiple SAFEs can consume 15–25% of the company before Series A.
MFN and pro-rata side letters change future economics silently — track them.
Harbor Labs modeled 18% Series A dilution; SAFE conversion pushed actual founder dilution to 26%.