Y Combinator's SAFE in Brazil: why the convertible loan agreement still wins

Y Combinator's official SAFE documentation recently got an overhaul: besides the three versions of the post-money SAFE (with valuation cap, with discount, and the "uncapped MFN"), YC now offers the Send a SAFE tool, which generates, signs, and sends the contract in about two minutes, plus a programmatic CLI built for AI agents. The prompt suggested by YC installs the yc CLI, logs in via device, and exposes commands like yc safes create, yc safes update, and yc safes send, with an important warning: sending triggers an email to a real investor and cannot be undone.

The instrument is impressive for its standardization. Created by Carolynn Levy in 2013, the SAFE has already been used to raise more than $15 billion across YC's portfolio. The point for anyone building a tech business in Brazil isn't the tooling, it's what the contract assumes underneath, and why that premise doesn't survive the legal crossing.

What the SAFE assumes that Brazil's Ltda. doesn't deliver

The SAFE is neither debt nor equity: it's the right to receive stock in the future. Conversion is automatic when the startup does a priced round (equity financing) and the investor receives preferred stock. All the mechanics described by YC (conversion into non-participating preferred, senior to common and junior to debt in a liquidity event, 1x back on dissolution) presuppose a capital structure with multiple share classes, an option pool, and the issuance of preferred shares.

This is exactly where the translation breaks down. The sociedade limitada, or Ltda. (Brazil's standard limited liability company structure and the default vehicle for the vast majority of Brazilian pre-seed startups), doesn't have shares, it has quotas, and it doesn't support preferred classes in the sense the SAFE requires. There's no "preferred stock" for the SAFE to convert into. YC acknowledges this limit by publishing versions only for companies in Canada, Cayman, and Singapore, and by restricting the tool to companies incorporated in the US. There's no Brazilian version, and that's not an oversight: the Delaware C-Corp is part of the design.

Clause by clause: where each term loses its footing

| SAFE term (YC) | What it does in the US | What happens in Brazil | |---|---|---| | Valuation cap (post-money) | Caps valuation at conversion; ownership sold = investment ÷ cap | Survives as a calculation, but depends on a share-conversion event that the Ltda. doesn't have | | Discount | Converts at a discount (typically 10 to 20%) on the round price | Only makes sense if there's a priced round with a defined price per share | | MFN | Automatically adopts the best terms from future SAFEs | Depends on a series of standardized instruments that the Brazilian market doesn't widely use | | Pro rata | Right to maintain percentage, in a separate side letter | Transferable, but needs to be rewritten in Portuguese with real corporate effect | | Automatic conversion | Becomes preferred stock at equity financing, no threshold | There's no preferred stock; conversion becomes a case-by-case negotiation |

The post-money cap detail illustrates the elegance that gets lost. YC teaches the direct calculation: a $1M target selling 15% implies a post-money cap of ~$6.7M ($1M ÷ 0.15). $500k at the same cap is ~7.5%. This ownership transparency is the main virtue of the post-money SAFE, YC's standard since 2018. But it's only "immediately calculable" because there's a defined share base to divide against. Without that, the number becomes an estimate.

Why the convertible loan agreement still wins in Brazilian pre-seed

The instrument that Brazilian accelerators and funds adopt in practice is the convertible loan agreement (contrato de mútuo conversível): a loan that converts into equity in a future event. It solves the structural problem because it doesn't require the immediate issuance of anything. The investor lends, and conversion becomes an amendment to the articles of association (quota entry or capitalization) at the agreed trigger.

The irony is that the convertible loan agreement is, technically, what YC classifies as the more complex model. In the source's own comparison, the difference between the SAFE and the convertible note is clear: the note is debt, accrues interest, and has a maturity date; the SAFE has no interest or maturity, which makes it simpler, faster, and cheaper to sign. The Brazilian mútuo inherits precisely the debt nature, with interest and a term, that the SAFE was created to eliminate.

In other words: the Brazilian founder trades the SAFE's simplicity for the mútuo's legal robustness not by preference, but for lack of an alternative. The SAFE is more elegant; the convertible loan agreement is what the legislation and the Ltda. structure support.

The counterpoint: when the original SAFE makes sense in Brazil

There's one path where YC's SAFE works without adaptation: when the Brazilian startup sets up a Delaware topco early on, with the BR operation as a subsidiary. In this setup, common in theses aiming at American capital or an international exit, the original SAFE applies directly, YC's tool works, and the investor receives exactly what the contract promises.

The trade-off is one of cost and timing. Structuring a Delaware C-Corp with a flip of the Brazilian operation involves lawyers in both countries, international accounting, and tax implications that are rarely worth it before there's a clear thesis for raising money abroad. For the founder who's going to raise R$500k from a local fund or a Brazilian accelerator, setting up a topco just to use the SAFE is premature optimization.

What this means for decision-makers

The practical takeaway is straightforward. If the raise is local, with a Brazilian investor and no immediate plan to flip, the convertible loan agreement remains the right instrument, and YC's SAFE serves as a conceptual reference, not a contract to sign. It's worth using YC's SAFE calculator to model cap and ownership, but the drafting has to go through a Brazilian corporate lawyer, translating each term into its equivalent effect in quotas.

If the thesis is already built around international capital from the start, then the cost of the Delaware topco pays off, and it's worth adopting the original SAFE right away, including the agent-friendly tooling. The costly mistake is the middle ground: taking the SAFE PDF, having it translated, and signing it over an Ltda., creating a contract that promises conversion into preferred shares that the company isn't legally able to issue. In that case, you gain neither the SAFE's simplicity nor the mútuo's legal security, you're left with a document that neither country validates.