On Y Combinator batch quality at scale

Since Rebel Fund invests exclusively in Y Combinator startups (with 80 YC startup investments to date) I’m often asked by other investors what I think about YC’s quickly growing batch sizes. Since I’ve been at every YC Demo Day since my own batch’s in Winter 2013, I’ve watched batch sizes balloon from 46 startups in my batch to 400+ in the most recent Winter 2022 batch.

The main concern investors have is whether YC can maintain the quality of startups it admits as the quantity increases roughly 10x over the past 10 years.

Since Rebel maintains the most comprehensive database that exists of YC startups outside of YC itself, we’re in a great position to answer this question. Our fund collects over 100k data points on each new YC batch, in large part to feed our Rebel Theorem machine learning algorithm which helps us predict and invest in tomorrow’s top-performing YC companies.

This data also helps us measure YC batch quality in several ways, from simple valuation growth to earlier predictors of success such as the percentage of startups backed by top-tier funds, to those having accomplished repeat founders, to those with founders graduating from top universities or working at top tech companies associated with startup founder success, and more.

However for the purposes of this post, we’ll examine the most simple measures of YC batch quality over time:

• Percentage of startups becoming a Top Company
• Percentage of startups becoming a unicorn ($1B+ valuation)

Before getting into those, here’s an illustration of how quickly YC batch sizes have grown since the accelerator’s founding in 2005:

The growth is clearly exponential, and can be divided into 3 distinct phases:

2005–2012 — During this period batch sizes steadily increased, but YC got a bit ahead of its skis in 2012 when things started to ‘break’

2013–2019 — After a batch size reset in W13 (blue circle), YC was able to continue steady growth after making some internal structural improvements

2020-present — YC had its first virtual batch in S20 (red circle) thanks to covid, which allowed it to unlock a new stage of growth reaching ~400 companies per batch

Moving on to batch quality, let’s start by looking at the percentage of YC startups by batch becoming a Top Company, which means a recent valuation of $150M or more:

This is our first clue that YC startup quality has not only maintained as batch sizes have grown, but actually improved.

The yellow circle highlights the Top Company ratio of batches from 2011–2014. I chose these years because YC was old enough to be mature as a program, but still had batch sizes below 100 startups. The average percentage of YC startups becoming Top Companies from these batches is 10%.

The blue circle highlights the Top Companies ratio of batches from 2015–2018, which had batch sizes >100 startups but are still old enough vintages to have reached their ‘terminal’ Top Companies percentage (i.e., if a startup from 2018 isn’t a top company yet, then it probably won’t be in the future). The average Top Companies ratio for these batches is 14%, which is a 40% improvement over the prior period. Go YC!

Next let’s look at the unicorn ratio for YC startups over these same two periods:

Here we can see that from the 2010–2014 batches, there was a 4% average unicorn rate (which is 4x the industry average of 1% for venture-backed startups). From 2015–2018, this grew to 5%, or a 25% improvement over the prior period, and this likely understates the terminal unicorn rate for this period since it often takes 5+ years for a startup to grow into unicorn status.

As an aside, something special happened in the W16 batch, which had a whopping 23% Top Companies ratio and 11% unicorn ratio. I’m not sure what!

While it’s too early to say what these statistics will look like for the 2019 vintages onward since they haven’t yet had time to mature, the data from previous years certainly suggests that YC is able to scale and improve batch quality at the same time. So, I wouldn’t want to take the opposite side of that bet ;)

Bonus:

In case you’re curious what percentage of YC Top Companies eventually become unicorns, if we look at the 2012–2016 batches (chosen to strike the right balance between a mature YC program and mature portfolio companies) the answer is 43%:

This means that of the 273 companies currently on the Top Companies list, we’d expect well over 100 to eventually become unicorns. Hold on kids!