Startups Can Be Both Healthy and Bubbly at the Same Time

Startups Can Be Both Healthy and Bubbly at the Same Time

Author:

Peter Walker

Read time:

2 minutes

Published date:

September 6, 2025

Graduation rates improving while certain valuations hit bubble territory is not a contradiction—the divergence between median and top of market captures both...

Are startups healthy OR are we in a dangerous bubble?

Why not both :)

Let’s take one metric: graduation rate.

The percentage of startups that raise Series A who end up raising a Series B (you could use Seed --> A, B --> C, whichever you prefer).

I hear the comments already: "The goal of a startup is not the next fundraise, this is a bad metric".

Eh, I disagree. Sure, fundraising does not equal success - but if you're a VC investor, you want companies you invest in to get marked up by the next round. And if you're raising a Series A as a founder, you very likely want to get to Series B. Not all! But most.

So what does "healthy" look like in this metric? We can plot the past 7 years of data to find out. I took the startups that raised Series A rounds in Q1 of each year from 2018-2024 and then looked to see how many of them made it to Series B in 2 years or less.

Graduation Rate Scores for Q1 cohorts

What’s Different about Today

Based on this metric alone, I'd say there's increasing optimism around VC-backed startups. But what about the AI bubble?

Our current bubble is very different from bubbles past. In the 2021 boom, basically all kinds of startups were getting rocket boosters. Rounds were rising alongside valuations.

In the AI boom, more and more money is going into fewer and fewer companies. So the headlines are full of term sheets with crazy numbers and yet the graduation rate is way below our last bubble period.

What happens when a concentrated bubble pops? Betting we find out in the next 12 months 😬

Author: Peter Walker

Peter Walker runs the Insights team at Carta, focused on discovering key data and narratives across the private capital ecosystem. In a former life, he was a marketing executive for a media analytics startup and led the data visualization team at the Covid Tracking Project.