Last month, venture capitalist Fred Wilson of Union Square Ventures posted a blog entry titled: The Rich Get Richer. In it, he notes that alarmists of a venture capital fuelled startup bubble are missing the point—it’s not an entire sector run amok, but rather, a small number companies that are driving headlines, consuming capital at a high clip, and reaching ever-higher valuations along the way:
"What is going on is that the late stage market is going crazy... Round sizes have gone up and burn rates [the rate at which companies spend funds] have gone up, but so much of this is limited to a hundred or a couple hundred companies. The rest of the market is more or less where it has been for years. The rich are getting richer. The middle class is stagnant. And the people who can’t raise a round still can’t. Only the top end of the market has really changed over the past five years. Kind of like the entire economy, isn’t it?”
As this brief analysis shows, Fred is right: a small number of later stage companies are skewing the overall numbers. However, there is one other point that Fred did not mention: this trend appears to be geographically concentrated in the Bay Area.
To test his statement, I calculated the statistical dispersion (standard deviation) of venture capital fundraising round sizes annually between 1998 and 2015. I collapse the data into two groups: the early and growth stages (identified as the first three rounds of venture capital for a company), and the later stages (companies raising four or more rounds of venture capital). It is within this group that has drawn Fred’s attention.
As the data show, the dispersion of late-round sizes has increased tremendously in recent years. In fact, the standard deviation today is nearly three times what it was at the height of the dot-com bubble in 2000 when funding round sizes were more similar across companies. Differences among sizes of earlier-stage deals, on the other hand, have held about steady.
However, the figures tell a much different story when excluding the San Francisco Bay Area from the analysis. The chart below plots the exact same data as before, but adds two additional series—the same two dispersion measures, but excluding data from the San Francisco Bay Area (dotted lines).
Here, the story for the first three rounds is basically the same as before, but the data for the later rounds tell a completely different story—one of steady yet slightly growing dispersion of deal sizes over time, rather than one of a substantial dispersion in the last two years as seen before. But, that still leaves the question: which part of the distribution of San Francisco area deals is driving the dispersion?
The chart below shows that while late-stage deal sizes at the bottom half of the distribution have been steady (25th percentile, median), deals at the top (75th percentile) have moved steadily upward since 2010 and sharply in 2015.
What’s also interesting is that the average has moved further away from the median—surpassing the 75th percentile so far this year. This indicates that deals at the very top are getting ever larger, particularly when compared with the rest of the distribution, pulling the average further and further up.
Fred Wilson is right—the fundraising rounds have been relatively steady, save for a few firms at the very top of the late-stage distribution. This trend nationally appears to be driven by activity in the Bay Area.
What this all means is anyone’s guess, so I'll stay away from offering normative judgments here. There are a number of perfectly good reasons why a small number of predominantly Bay Area firms are raising ever larger rounds of late-stage venture capital—including, the growth in winner-take-all markets, the shift towards delaying or forgoing IPOs, and potentially an increasing importance to scale late-stage companies in the world’s leading innovation hub (this last point would be a deviation from early-stage companies, which have seen a geographic dispersion in recent years, as I have written about before).
Or, perhaps, the alarmists have a point—maybe things have gotten crazy. Time will tell.