Comment by mlinsey

A 33% per-round dilution would be much more sensible (and more common). In the right fundraising environment this might even be 20% or 25%, but 33% is the right default assumption if you have no information.

Another hidden variable is the presence of liquidation preferences for VCs. Some funding agreements allow VC's to "double-dip", ie take their liquidation preference and then also take their full percentage of the leftover amount. However, with YC advising them and with the Heroku founders having a reputation inside YC as being good fundraisers, I think there's no way Heroku's VC's got this.

If you assume no double-dipping, a 33% dilution per round, a YC-average 6% stake in Heroku, and an average amount per YC startup of $20K, YC's share of Heroku is enough to pay for their investment in every YC company ever up until this point, and also all the companies in the next two batches assuming they are the same size as the most recent (and largest ever) batch. Anything they eventually get from Dropbox, AirBnB, etc would be pure profit.