[jasonkwon]

It's clear you put time and thought into this post, so it deserves an equivalent amount in response. I think you’ve understood some things correctly, but not others, but that’s why we're on HN - to help clarify.

(1) The modeling you’ve done for the premoney safes is correct, but it’s incorrect for the postmoney scenario. That’s because Angelcalc hasn’t been updated yet for postmoney safes that track the one we released. Angelcalc includes the Series A option pool increase in both flavors of safes, because what people were doing when flipping standard premoney cap safes to postmoney cap safes is they were just changing the pre to post, and nothing else. We deliberately took out the Series A pool increase for reasons that are all detailed in our post. That means both we and the safe holders share the Series A pool increase with the founders, which is not how it’s working on Angelcalc (but we will update it soon).

Also, in your postmoney scenario, the valuation cap for the $2M safe needs to be adjusted to be a $12M postmoney cap safe.

So if you update the postmoney scenario using all of your variables based on the postmoney safe we released, the results are different. I did it by hand on excel - here’s a screenshot:

https://imgur.com/m4V51SH

Happy to send you a copy of the excel file. Also, to be perfectly transparent, these examples are somewhat artificial because they assume a 0% option pool issuance in both cases, which is unlikely to be the case. Safe investors will do better than in the screenshot I sent the more options that are issued before the Series A round. They also have the option now to ask for a template side letter to participate pro rata in the Series A round itself.

(2) The YC deal should be viewed together with the money founders will raise at demo day, i.e. as one continuous round, and thus the combined % of the company you end up selling. That combined % for YC and demo day safes was often too high in the old deal because founders had a hard time understanding how dilution was unfolding. Safe rounds may not have been priced correctly because of that lack of clarity. With these new changes, the days of raising on safes and not knowing how much you owned are over. The days of planning a Series A fundraise not knowing how much you’ve already been diluted are over. We strongly believe that founders will end up less diluted by the combined % of YC and demo day safes. It’s interesting that you would characterize an uptick in YC ownership as “downside” for the founders. I don’t think founders look at our ownership - they look at theirs.

(3) An underlying assumption of your post is that the safes and YC deal are changing, but everything else — valuations, option pools, amounts people raise and dilution transparency (or lack thereof) — will remain the same. The point of us doing this though is that we expect it to change all of those other things. Everything is tied together. As Michael already pointed out, once you can see what’s happening, both investors and founders can take better actions on both fronts. High-res fundraising should also become easier, as Carolynn points out on http://ycombinator.com/documents.

[sethbannon]

Thanks Jason for the thoughtful reply, and thanks also for your work on simplifying and improving the YC SAFE. I think these improvements will benefit the entire ecosystem (founders & investors & employees) by making it easier for everyone to understand SAFE dilution.

Still not clear on how (in most cases and assuming there is not a 0% option pool pre-equity round) this will not lead to increased expected dilution for founders from the YC deal as compared to the old deal, so would love to play around with your Excel sheet. My email is my HN username at gmail.

[sroussey]

I think they expect to get a better deal but that it will be offset by demo day investors.

[jasonkwon]

Sure - just sent to you.

[sethbannon]

Played with your excel and while the difference is not the same as I calculated with AngelCalc, it still seems the dilution from this new YC deal will be greater than the old YC deal post-equity round in basically every circumstance.

Essentially, with this new deal, after equity financing YC will own 7% minus the dilution from the equity round minus dilution from any options pool increase [1]. Previously, after equity financing, YC would own 7% minus the dilution from the equity round minus the dilution from the SAFE round.

While it's true founders are getting a little bump on YC absorbing the dilution from a Series A option pool re-up, in my experience these are typically 5% to maximum 15% increases. Whereas the dilution from post-YC SAFE rounds are typically 15% to maximum 30%. So YC is assuming a potential 5-15% dilution in their ownership while avoiding a 15-30% dilution in their ownership. Translation: YC will own more post-equity financing than they would in the old deal.

This puts a burden on the founders to make up for that increased dilution by raising the post-YC SAFEs at a higher valuation than they otherwise would, which will likely make those raises harder. Alternatively, they can raise their Series A at a higher valuation than they otherwise would to make up for YC's extra ownership, but that will make those raises harder than they otherwise would be. So there is a real dilution downside for founders here.

1: in reality YC will continue to own 7% after the equity round because they'll exercise their pro-rata right during the equity round but that doesn't change the underlying point being made here so will ignore it for simplicity.

[jasonkwon]

I answered your first model comparison with point #1. Can't use it here since I don't know what other scenarios you're modeling out.

I think points #2 and #3 continue to apply. You can't say that the new framework will result in more dilution in "basically every circumstance," because that assumes all of the surrounding aspects of the circumstances will continue to remain the same. I can model out a scenario for you as well that shows a thoughtful founder now able to better plan out a round using a series of escalating valuation caps, rewarding the earliest investors, to raise the same amount of money but for less overall dilution, in precisely the manner described by PG in his original post. Or another scenario where investors are ok with slightly higher valuations because they have certainty of ownership. Or another scenario where a founder who previously would've raised an unnecessary extra 5% now doesn't, because the dilution math is clearer.

Your point about safe dilution being 15-30% is exactly right, and exactly one of the reasons we're doing this. The high part of that range is too high. People are raising too much, on too dilutive terms, because of the lack of transparency. Founder dilution is driven by what everyone else on the cap table is getting, not just us. If that sum total of "everyone else" is still less, founders are still net better off. If you want to frame the dilution aspect all indexed to YC ownership, that's ok, but the outcome was already pre-figured by your choice of framing in the first place. We can agree to disagree on that.

That said, I think one thing we can definitely agree on is that the success or failure of this new framework will be judged by whether founders are more or less diluted in the end, and have a harder or easier time raising money. We've already made our bet, so there's not much to do here but let it play out.

[abalone]

For us mere mortals following along... I don't necessarily understand what you guys just debated about the impact of options pools but I did grok this from the spreadsheets: In the new model,

Founder stake goes 48.8% -> 49.6%, a modest 1.63% increase.

YC stake goes 3.68% -> 4.55%, a whopping 24% increase.

Other SAFE investors stake declines 12.5% -> 10.8%, a 13% decrease.

So while the immediate impact on founders appears negligible (assuming all variables stay the same), it would seem that this new model represents a fairly large transfer of ownership from other SAFE investors to YC.

Wouldn't this be a cause of concern for founders? If SAFE investors have a particular percentage ownership in mind, wouldn't this push founders to raise more from them (or give a discount, etc.) and thus increase dilution?

[sethbannon]

You've nailed it. Becaise of YC's increased ownership, founders will either have to raise less from SAFE investors or take more dilution, because this change won't make those SAFE investors be willing to pay higher valuations. That's the downside.

[jasonkwon]

The impact on safe investors will be less than in those 2 models because those models assume no pool, ie no hires between Safes and Series A. That’s why I said it was artificial, and that safe investors will do better than depicted. In reality there will be some amount of pool in most cases, so the safe investors won’t be bearing the full impact of the dilution from the 10% pool, ie they will own more than in the screenshot posted.

You’re right that many safe investors have a particular % in mind when the invest. The problem with the old framework is that they had no certainty on that. They could invest $1m at $9m premoney cap and think they got 10% but if the company raised more money on a bunch of other Safes or caps, they ended up with a lot less. The dilution confusion impacted both founders and investors alike. In the model you saw, the safe investors got exactly what they bargained for, 16.7% (2m invested at 12m post = 2/12 = 16.7%), and then ended up with ~11% because they were diluted by the Series A round. That’s what happens - each round dilutes the next. It’s also what would happen if those investors just did a priced round instead of using a safe.

The safe investors also now have the option to ask for a side letter that gives them the right to invest Pro Rata at the Series A. They didn’t have this in the prior framework. And if they did do Pro Rata here, they would probably end up with more ownership than they would in the old framework. This is actually more consistent with how most investors invest. Spread bets early and then double down on the ones that are working, based on an ownership level that they know they have today.

[xmly]

YC = angel round + Demo day seed round. This is my understanding. Just currently the angel was "pissed off" by the dilution in the demo day round....