YC's post-money SAFE: the cap stacking that dilutes the founder without them noticing

The SAFE (Simple Agreement for Future Equity) was created at Y Combinator in 2013 by attorney Carolynn Levy, and since 2018 YC has standardized the post-money version as the official model distributed on the accelerator's documents page. In practice, it's common for this instrument to be copied from a template in bridge rounds without the founder redoing the math on how much they're actually selling.

The promise of the post-money SAFE is simple: the percentage of equity sold is always the amount invested divided by the valuation cap. Raise $1M with a post-money cap of $6.7M (which equals $5.7M pre-money) and you know, at the moment of signing, that you've sold about 15% of the company. This transparency is why YC abandoned the pre-money model: before 2018, the calculation of how much each investor took depended on how many other SAFEs and convertible notes were already outstanding, and nobody knew the real number until the priced round converted everything at once.

The difference YC's documentation makes explicit

YC's own documents page sums up the difference: the cap on a pre-money SAFE is based on the pre-investment valuation, while the cap on a post-money SAFE already includes that SAFE's own investment. But the most important line in the material lies in the detail of what post-money includes and what it leaves out.

The post-money valuation cap is "post" all of the money you raise on SAFEs. It is NOT also "post" the equity financing money.

Y Combinator, Startup Documents (SAFE)

This sentence is the key to the entire stacking problem. The post-money cap of a specific SAFE already accounts for that SAFE's money and for the previous SAFEs issued by the company, but it does not account for the priced round that comes afterward or for the option pool that's usually created or expanded in that round.

The math of stacking, according to YC's own example

The documentation gives a direct example of how founders combine multiple caps in the same bridge round: instead of selling the entire $1M at a single $6.7M cap, it's possible to split the raise into two separate SAFEs, each with its own valuation cap.

• $500k at a post-money cap of $5.5M sells approximately 9% of the company.
• $500k at a post-money cap of $8.3M sells approximately 6% of the company.
• The sum of the two slices reaches the same ~15% as the original example, just split across two instruments with different prices.

The problem isn't the math of each individual SAFE, which is correct and exactly what YC's official calculator returns. The problem is operational: founders raise bridges at different moments, with different investors negotiating different caps, and rarely add up the accumulated dilution before closing the next check. Each SAFE, on its own, looks small. The whole stack doesn't.

Where the dilution really hides: the option pool

The point that most catches founders off guard isn't the second or third SAFE itself, it's what happens when all of them convert in the priced round. The post-money cap is already "post" the option pool that existed before the SAFE round, but YC itself makes clear that it is not "post" a new or expanded pool created in the priced round.

In practice, this means that when the priced round creates or expands the option pool, that entire pool comes out of the founder's slice, not out of the slice already converted by the SAFEs. The typical outcome: the founder adds up 15% sold in SAFEs, thinks they still have 85% plus the old pool, and finds out at the Series A negotiating table that a new pool was stamped on top of that, pre-money, consuming their stake directly rather than the bridge investors'.

The counterpoint: post-money is still safer than pre-money

It's worth acknowledging what the post-money model actually solved. Before 2018, with stacked pre-money SAFEs, the percentage sold to each investor depended on how many other convertible instruments were already outstanding at the moment of conversion, and that number only became clear at the priced round. It was possible to unintentionally sell more equity than the founder thought, because the denominator of the calculation changed with each new SAFE.

Post-money eliminates that ambiguity within each instrument: investor and founder both know, at signing, exactly what slice that specific SAFE represents. The risk that remains isn't miscalculation, it's failure to add up. A founder who stacks SAFEs at speed, especially with tools that make issuance nearly instantaneous, like YC's own Send a SAFE, which the documentation describes as generatable, signable, and sendable in about two minutes and already designed to be operated by AI agents, has less friction to generate the next SAFE and less natural discipline to recalculate the accumulated total before signing.

What changes for those closing the next bridge

Before signing yet another SAFE on top of the ones that already exist, the founder should manually add up, SAFE by SAFE, the total percentage already committed, rather than trusting that each isolated cap remains small. It's also worth negotiating, already during the bridge, a target option pool for the next priced round, so that pool dilution isn't a surprise stamped solely on the founder when the Series A closes.

The difference between closing the round and later discovering you've lost control isn't in the SAFE itself, which is a short, standardized document. It's in treating each instrument as an isolated piece, when the correct cap table is the sum of everything already issued, plus whatever is still going to be stamped in at conversion.