Y-Combinator SAFE conversions decoded
OK, the Y-Combinator SAFE invented in 2013 (we used it back in 2014) and then updated in 2018 to the "post-money" SAFE is a great example of an simple idea that is actually kind of hard to calculate unless you spend some time with algebra. Most of the sites like Carta or Andrew Jenkins, are so automatic (and mainly filled with assumptions) that for something as important as the equity in your company, you should really take apart the math and make sure you get it.
Then you can land in technical discussions about what is pre- and post- on Medium and get even more confused. The basics are pretty easy to understand but hard to calculate:
- A SAFE (Simple agreement for future equity) isn't debt like the typical convertible round, but instead a promise that when there is a real equity round, early investors will get either a discount from the round price of 20% or so or they will get a cap so that if the round is say a $10M pre and they have a $5M cap they get their shares at the $5M valuation.
- The nice thing is that there is no priced round so an investor doesn't have to think too hard about value but knows they are getting a good deal in the future either from the discount if the valuation is low or the cap is the valuation is high.
But, it's incredibly hard to think about how to do the math for this stuff because how do you calculate a price when the number of shares you get is bouncing all over the place. Here is where DLA Piper has a great #ELI5 (explain it like I'm five) explanation.
The first thing to do is to understand the cap. It is actually easier to understand than the discount. So what this means is that for high valuation cases, typically, your ownership falls as the valuation rises for a given investment. Say you put $10 into Tesla in 2013 but the company doesn't get any equity investment until 2021 when it is worth $1T, then you will get $10/$1,000,000,000,000 worth of shares which doesn't feel super great. That's why a cap was created.
So for instance, if you invest $10 in Tesla with a $100 cap, this your ownership decreases as the valuation rises. If the company is worth valued at $20 pre-money (that is pre-investment), then your $10 buys $10/$20 or 50% of Tesla which isn't bad. The main point is that at the cap of $100, your ownership for all higher valuations "sticks" at "your investment/the cap". So in this example, $10/$100 is 10% of Tesla and on the financing round that happens, you will get at least 10% (as an aside making $100B on a $1T pre-money sounds pretty darn good). Now of course most companies will have a round way before $1T and having a cap is an incentive to get a round done quickly as close to the cap as possible.
However, there is another case called the discount in the current round. What you do is see whether the cap or the discount method gives you the most shares and pick that one. I had somehow thought that you got a discount *and* a cap at first, but it is either/or.
But the confusing thing is that a discount means that in the current round, you get a 20% discount compared with the new investor, so it is completely different calculation from a cap. As an example if the new investor buys shares at $1, then you can buy shares at a discount rate of 80%, so you can buy your shares at $0.80.
The confusing part about this math is that most of the time investors don't tell you the price per share, but they say I'm going to pay $10M pre-money that values all the prior shares in your company. But, there is a dynamic process where each additional share I give the SAFE person means the total share count goes up, so the price paid goes down. You can really tear your hair out trying to figure this out.
So you can either do iteration or you can solve the problem analytically and it turns out there are two conceptual things to do. The first is that the discount is like you getting a lower pre-money than the current investor. So if they are offering $10M pre that's like the SAFE investor is getting a $10M * 80% = $8M pre-money.
Ok, so we are getting close. This is the trick we need because we also know the SAFE investment amount which is say, $1M, which means that in the end, the number of shares should work so that the SAFE investors get $1M/$8M worth of the company or 16%. So now you are nearly done and here is where the algebra comes in because while we don't know how shares the SAFE investor is going to get, we do know the current shares outstanding in the company, say that is 50M shares.
So if you start it, you suddenly realize this is a ratio problem. You know a few things, you know the ratio of SAFE to non-safe and you know their percentage ratio that is SAFE/50M = 16%/(1-16%) because the if the SAFE is 16% then the non-SAFE must be 100%-16% = 84%. So suddenly is solved the SAFE shares = non-safe Shares * (%SAFE/%non-SAFE).
To me that's the crazy trick that I've learned a million times in algebra and it's why you see formulas like
x/(1-x) all over the place in many equations.
The home stretch
OK, now the final distinction for the discount is what is the capitalization of the company. That is what are the non-SAFE shares. For the so-called "pre-money SAFE", the capitalization includes the option pool so all options are in there but it doesn't include other SAFEs. Some folks take a series of SAFES which makes life really complicated, you do a $100K at $5M cap then $200K at $10M cap and so on. So then you have this SAFE layering problem which is really confusing and makes my head hurt. So a post-money SAFE means that you don't include the option pool but you do include all the SAFE investments. I mainly hope I never end up in a situation where there are an extended series of SAFEs. 🙂