On the 176% annual return of a YC startup index
As my long-time readers know, I’ve written dozens of data-driven blog posts on Y Combinator startups over the years, but I’m yet to answer the single most important question about YC startups from the perspective of early-stage investors — how much money do they make??
YC has published data that hints at this, like the Top YC Companies by Valuation list, but no one has gone so far as to estimate actual early-stage investor returns for YC startups… until now.
At Rebel, we build and maintain the largest database that exists of YC startups perhaps outside of YC itself, which we use to feed our proprietary Rebel Theorem 2.0 machine learning algorithm and ultimately inform our investment decisions targeting the top 5–10% of YC startups each year. This puts us in a great position to estimate the overall investment returns of a hypothetical YC startup “index,” which is the goal of this post.
Before getting into the data, let me caveat that we had to make some estimates of YC startups' initial & current valuations and dilution by funding round (Series A, B, C, D, etc) since this data is non-public. However, between our internal database and third-party data sources like Pitchbook, I think we were able to get pretty close.¹
Now let’s get into the numbers!
To start, let’s talk about how ridiculously well YC startups perform as an “asset class.” According to our math, if you had invested in every YC startup at Demo Day since the accelerator’s founding in 2005, you would have made a whopping 176% average annual return net of dilution² (practically speaking this would be impossible, but more on that later).
This is far greater than other asset classes and even top-decile venture capital funds:
Below are overall seed-stage YC startup investment returns broken down by year. Even ‘modest’ years like 2010–2011 produced 70–80% average annual returns, and the stronger years like 2012–2013 saw 300%+ average annual returns. The big outlier is 2009 at 2500%+, which included both Airbnb and Stripe (note that was a recession year btw…)
I wouldn’t get too worried about the small bars in 2019–2021… those batches are still quite young and the companies still maturing.
Now let’s switch gears and zoom in on the top 100 YC companies by seed-stage investor net return multiples.
As I examined in On the eery predictability of YC startup valuations, YC startup outcomes follow a steep power law curve, with the decacorns really dominating. Even the smallest bars in the chart below represent companies producing >100x net return multiples for their seed investors, yet they’re dwarfed by startups like Airbnb, Stripe, Dropbox, Instacart and DoorDash which produced >1,000x net returns for their early investors.
Startups are really a winner-take-all game. According to our estimates, about 5–6% of YC startups from recent batches will become unicorns, and of those, about 10–12% will become decacorns — which, by definition, are 10x more valuable than the unicorns. And even amongst the decacorns, just a couple of companies dominate the returns chart. The power law curve of YC startup returns remains steep no matter how much you zoom in.
YC Startup Investment Strategy
By now you’re likely asking yourself, “So how do I make a juicy 176% annual return investing in YC startups?”³
The bad news is you probably can’t.
Given the winner-take-all nature of startup outcomes, it’s vital to invest in the top-performing companies, and that’s only guaranteed with an index approach. The problem is that no such YC startup index exists for outside investors and probably never will.
Only YC could offer this, as it would be virtually impossible for a retail investor (or even a fund like us) to invest in every YC startup in every batch. The startups would have to agree to this one by one, and the best startups are disincentivized to do so since their rounds are often oversubscribed, and thus, they get their choice of investors. Why would a founder take money from an index when they can instead take it from a strategic and value-added investor who can help them grow?
YC actually tried a version of this starting in 2011 called YCVC, which eventually involved 4 outside investors (Andreessen-Horowitz, General Catalyst, Maverick Capital, and Khosla Ventures) each investing $20k into every YC startup. However, YC eventually shut it down in 2014 due to signaling risk and information issues. YC’s latest standard deal involves a $375k uncapped SAFE for each startup, which is definitely a sort of YC index at seed-stage valuations, but just for YC :-)
So, given that you can’t invest in every YC startup to make sure you catch the outliers, the next best option is to figure out how many startups you can support in each batch and try hard to pick the best ones. Your target portfolio size should be a function of your screening and portfolio management capabilities, investable capital, and ownership targets.
The tricky part is that even a little bit of adverse selection due to poor screening (i.e., ability to pick future winners) or access (i.e., ability to get on the best companies’ cap tables) is detrimental in a world where ~5% of startups drive the overwhelming majority of investor returns. As an illustration, if you can invest in 20 startups in a YC batch, we’d statistically expect 1 or maybe 2 of them to really drive returns… what happens if you miss those 1 or 2?
At Rebel, we have massive data, selection and access advantages over other investors, including a handful of YC unicorn founders as partners, and it’s still hard for us to pick the best companies in every YC batch (though we do much better than most!)
My advice to investors would be to maximize your shots on goal, have a disciplined screening process, offer a strong value proposition to portfolio companies, only invest in what you understand, and avoid adverse selection at all costs. Either that or invest with a fund manager who can.
¹We estimated average seed-round valuations by YC batch based on industry averages from Pitchbook adjusted by a “YC premium” of approximately 2x that we’ve observed over the past several years of investing. We also pulled industry average dilution by round (Series A, B, C, D, etc) from third-party data sources and applied them on a per-company basis based on announced funding rounds.
²Since we only have reliable valuation data for companies on the YC Top Companies list, we assumed a 0x return for all non-YC Top Companies, which is extremely conservative. Average annual return (AAR) is not the same as internal rate of return (IRR) since the former doesn’t factor in compounding or the time value of money — it’s simply total return divided by number of years.
³Since startup equity is an illiquid asset, you wouldn’t really ‘make’ this return until the companies exit. However, you’d probably end up with an even higher return once they do since most top YC startups are still growing quickly.