Common SAFE Agreement Mistakes Founders Make

What a SAFE actually does

A SAFE is an instrument that converts into equity at a future priced round — typically the next round in which the company issues preferred shares to investors. Instead of agreeing on a valuation today, the founders and the SAFE investor agree on the terms that will determine the price and amount of shares the investor receives when that priced round happens.

The key economic terms are the valuation cap and the discount. The cap sets a maximum effective valuation for the SAFE investor; the discount gives the investor a percentage reduction off the priced-round share price. The investor converts at whichever is more favorable to them. Other terms — pro rata rights, MFN, and whether the SAFE is on a pre- or post-money basis — further shape the outcome.

Mistake: misunderstanding the valuation cap

The valuation cap is the maximum effective valuation at which the SAFE converts. If the company raises its priced round at a valuation above the cap, the SAFE investors convert at the cap, receiving more shares for their investment. If the round is below the cap, the cap does not apply and the discount usually does.

Founders often set the cap too low — eager to close the round — without modeling how many shares the SAFE investors will receive at the next round's valuation. A low cap is generous to the SAFE investor and dilutive to the founders. The cap should reflect a realistic, slightly conservative view of the company's next-round valuation, not a number chosen to make the round easy to close.

Mistake: ignoring the discount

The discount gives the SAFE investor a percentage off the priced-round share price — commonly 20%. If the next round is priced at $10 per share and the discount is 20%, the SAFE investors convert at $8 per share. The discount rewards the SAFE investor for taking early risk.

The mistake is granting both a low cap and a generous discount, so that the SAFE investor is doubly protected. Each protection reduces the founders' ownership at the next round. Founders should understand which protection actually applies based on the next-round valuation, and avoid granting more downside protection than the round requires.

Mistake: granting pro rata rights casually

A pro rata right lets an investor participate in the next priced round to maintain their ownership percentage. In a SAFE, pro rata rights are optional — not every investor gets them, and not every SAFE includes them.

Granting pro rata rights to early SAFE investors means they can buy more shares at the next round to avoid dilution. This is sometimes appropriate for a lead investor, but granting it to every small SAFE investor can crowd out new investors in the next round and complicate the cap table. Founders should treat pro rata rights as a negotiated term, not a standard clause.

Mistake: signing an MFN you do not understand

A most-favored-nation (MFN) clause entitles a SAFE investor to the benefit of any better terms the company grants to later SAFE investors before the priced round. The intent is to protect early investors from being diluted by later, more favorable SAFEs.

The risk is operational. An MFN means that every time the company signs a new SAFE on better terms, it must go back to the MFN investors and offer them the same terms — creating administrative overhead and a compliance obligation that is easy to miss. Founders who sign an MFN without understanding the notification requirement can inadvertently breach the SAFE and create claims from the investors they were trying to reward.

MFN clauses are appropriate in some early, pre-cap rounds, but founders should understand exactly what they are committing to and whether the administrative burden is justified by the amount being raised.

Mistake: stacking SAFEs with inconsistent terms

Startups often raise multiple SAFEs over time — from angels, from a program, from early institutional investors. Each SAFE can have different caps, discounts, pro rata terms and MFN provisions. When the priced round arrives, all of them convert at once, and the founders discover that the combined dilution is larger than expected.

The mistake is signing SAFEs one at a time without modeling the cumulative effect. Before each new SAFE, founders should model how all outstanding SAFEs convert at a range of plausible next-round valuations, so the dilution is understood before the next round sets the price. A clean, consistent set of SAFEs also makes the priced round simpler for incoming investors and their counsel.

How to negotiate a SAFE you can live with

A SAFE is fast to sign, but the terms are negotiable, and the negotiation happens before signing — not after. Founders who treat the investor's first draft as a take-it-or-leave-it document often accept terms they could have adjusted with a short, informed conversation.

The first step is to model the SAFE's effect before signing. Run the conversion at a range of plausible next-round valuations — low, expected, and high — and see how many shares the SAFE investor receives in each case. This turns abstract terms (cap, discount) into a concrete dilution number the founders can reason about.

The second step is to identify which protections are standard and which are generous. A single valuation cap, or a discount, is standard. Both, stacked, is generous. Pro rata rights for a lead investor may be appropriate; for every small investor, they are not. An MFN clause is appropriate in some early rounds but creates ongoing obligations the founders should understand.

The third step is to keep the SAFEs consistent. If the company raises multiple SAFEs, using the same template with the same core terms — adjusted only for the cap, which may rise between tranches — keeps the next round simple for incoming investors and their counsel. A cap table with six SAFEs, each with different terms, is a red flag even if each individual SAFE is reasonable.

A SAFE reviewed by counsel before signing, with the dilution modeled and the terms understood, is a SAFE the founders can live with. A SAFE signed to close the round quickly, with terms discovered later, is a common and avoidable source of founder regret.