Y Combinator's SAFE doesn't speak Portuguese: where the contract breaks in Brazilian angel funding

Y Combinator's SAFE (Simple Agreement for Future Equity) is today the default pre-seed funding instrument worldwide. Created in 2013 by Carolynn Levy and standardized in the post-money version in 2018, it has already been used to raise more than US$15 billion in the YC portfolio alone, according to YC's official documents page. The promise is appealing: a short, free contract in which the only variable usually up for negotiation is the valuation cap. The investor puts in money now and receives shares later, through the automatic conversion that happens when the startup does a priced round.

The problem for anyone building a tech business in Brazil isn't in what the SAFE promises. It's in what it silently assumes: that the company is a Delaware corporation (or, in the international versions, one from Canada, Cayman, or Singapore), that the investor is accredited under SEC rules, and that conversion will result in preferred stock, a structure that Brazil's Corporations Law (Lei das S.A.) simply doesn't reproduce in the same way. YC itself makes this explicit: the U.S. form doesn't work for other jurisdictions, and the Send a SAFE tool only supports companies incorporated in the U.S.

How the SAFE converts, and why that's the sensitive point

In the post-money SAFE, dilution math is transparent: the stake sold is the amount invested divided by the cap. US$500,000 at a US$6.7 million cap gives about 7.5%; US$1 million at the same cap gives about 15%. That clarity is the biggest gain of the post-money version over the old pre-money one, and YC even offers a calculator to run the numbers.

But notice the trigger. The SAFE is not debt (no interest, no maturity date) and is not equity until it converts. Conversion is automatic and only happens when the startup sells preferred stock in a priced round. In other words, the instrument assumes that, down the road, there will be a standardized preferred share class with liquidation rights (the investor gets the greater of 1x the investment back or conversion into shares). That's where a literal translation breaks down.

| Original SAFE assumption | Corporate/foreign exchange reality in Brazil | |---|---| | Conversion into Delaware preferred stock | A Ltda. (Brazilian limited liability company) doesn't issue shares; conversion requires an S.A. (corporation) or an adapted convertible loan | | Accredited investor (SEC rule) | No direct equivalent; CVM (Brazil's securities regulator) has its own regime for angel investors | | Investment via USD wire, cap in dollars | Foreign capital inflows require registration with the Banco Central (Brazil's central bank) (RDE-IED) | | SAFE is not debt, no maturity | Brazil's Angel Investor Law (Lei do Investidor-Anjo) treats the investment as not part of share capital |

Where most founders go wrong

The market standard in Brazil isn't the pure SAFE, it's a variation of the convertible loan (mútuo conversível) or the angel investment contract set out in Brazilian law. Many founders download the YC PDF, copy the cap and discount logic, and tie it to a Brazilian Ltda. It works on paper until conversion time, when three gaps show up.

The first is entity structure. The SAFE only makes full sense in an S.A. (corporation), because only that structure issues the share classes the instrument assumes. A Ltda. has no preferred stock; forcing conversion turns into a corporate renegotiation from scratch at the Series A, exactly the friction the SAFE promised to avoid.

The second is foreign exchange. A foreign investor who signs a SAFE in dollars and wires the money needs to register that inflow as foreign investment with the Banco Central so that, on exit (sale or dividends), the money can go back out without punitive withholding. YC's FAQ even mentions the investor sending "a little less or more" because of the exchange rate, treating it as a trifle. In Brazil, that exchange-rate difference isn't an accounting footnote: it affects the registered amount and can create a mismatch between what was contracted and what was actually brought into the country.

The third is the MFN clause and signing timing. The SAFE has an elegant rule: the MFN adopts the best terms of any SAFE issued later, and the date that counts is the later of signing or receipt of the wire. YC's guidance is to close the lower-cap SAFEs (with the money already sent) before issuing an uncapped MFN. That date choreography assumes instant settlement by wire. With foreign exchange, deal closing, and BC (Banco Central) registration in the mix, "you sign last and control the date" stops being trivial.

What's still worth importing

The honest counterpoint: none of this means the SAFE is useless here. Its conceptual engineering is excellent and worth copying. Three things in particular:

• Separate rights into a side letter. YC recommends keeping the SAFE untouched and putting pro rata, MFN, and information rights into a separate document. It's a discipline that avoids a polluted cap table and applies to any Brazilian angel contract.
• Think about dilution as the pair invested amount/cap, not as a percentage negotiated by feel. That keeps the math transparent for both sides, as in the post-money model.
• Calibrate rights to the size of the check. As the FAQ itself says, a US$500,000 investor has a case for pro rata; a US$10,000 one doesn't. That applies just as much to the angel putting in R$20,000 as to the fund.

There's also a 2025 detail that changes the operational flow: YC made the Send a SAFE tool agent-friendly, with a CLI (yc safes create, yc safes send) that an AI agent can operate. It's a sign of where the U.S. market is heading: issuing a funding instrument as a terminal command. Except, again, the CLI only covers U.S. companies, and the prompt itself warns that yc safes send triggers a real, irreversible email to the investor.

The takeaway for decision-makers

The question a Brazilian founder needs to answer before copying the SAFE isn't "which cap to negotiate," it's "where will my company be incorporated when this contract converts." If the plan is to raise dollars from a foreign fund and the likely exit is a sale or a round abroad, it makes sense to set up a holding company in Delaware or Cayman and use the real SAFE, in the jurisdiction it was written for, with local counsel (YC itself requires this in the non-U.S. versions).

If the operation is 100% Brazilian, with a Brazilian angel and a domestic S.A. or Ltda., the safer path is to use the convertible loan or the angel investment contract and import the SAFE's logic, not its text: the post-money cap, the side letter, automatic conversion tied to the next priced round. Translating the YC PDF line by line for a Ltda. is the shortcut that saves days in angel funding and costs weeks of corporate renegotiation at the Series A, exactly when the big money is on the table and the cost of getting it wrong is highest.

Source 1: Y Combinator, SAFE: Simple Agreement for Future Equity (official documents) (https://www.ycombinator.com/documents)