Growth, AI, or Both: What’s Driving Premium Valuations
I have read a few reports recently on software valuations that have concluded that growth explains a company’s multiple far better than profitability does.
A recent report from ICONIQ weights growth 2.7x more heavily than margin, while other data from Meritech shows the same pattern over time. Take each group’s multiple and divide it by its own growth rate, and the top 10 traded at 2.3x the median company’s level before the ZIRP era, 2.0x at the height of the 2021 boom, and 2.9x today. The widest gap between the most expensive companies and everyone else wasn’t the 2021 bubble, it’s now.
But is it all about growth?
Tomasz Tunguz recently made the point that within a category, growth doesn’t explain who’s on top. He highlighted that Shopify trades at 8x its category median, while Salesforce, growing the slowest in its category, still trades at 4.0x forward revenue against a 2.3x category median. So, on the surface, that’s a direct contradiction of everything in the first section.
But it isn’t, necessarily. ICONIQ and Meritech’s numbers cover the whole market, where growth does explain a lot of the picture. Meritech’s own data even shows the gap between the top 10 and everyone else widening over time. That made me curious whether growth still explains the gap once you’re only looking at the companies already at the top, so I pulled Bessemer’s own cloud index to look specifically at the top 10 myself.
Testing it
A simple first check is whether the fastest-growing companies in the index are actually the ones commanding the top multiples. I ranked all 62 companies by growth rate and looked at where the top 10 fastest growers ended up.
Source: Bessemer Cloud Index
Six of them made the top 10 by multiple, so growth clearly is a large factor here. But Figma is the second-fastest grower in the index, at 48%, and trades at 5.9x, while Zeta grows 44% and trades at 3.7x. So, growing fast doesn’t automatically get you into the top 10.
Then I isolated the ten most expensive companies by EV/forward revenue multiple (Palantir, CrowdStrike, Cloudflare, Palo Alto Networks, Datadog, Snowflake, JFrog, Rubrik, DigitalOcean, Samsara) and ran the same growth regression on just that group.
Source: Bessemer Cloud Index
Across the full index, growth explains 43% of the variation in multiple. Within the top 10, it only explains 24%, but if you remove Palantir (company with 93% growth & 43% FCF margin) it drops to just 2%.
Snowflake and CrowdStrike are a good comparison to illustrate this point. Both companies sit on almost identical Rule of 40 scores (~54%), but CrowdStrike trades at more than double Snowflake’s multiple, despite growing more slowly.
Growth explains the market, but it doesn’t reliably explain who gets into the expensive tier, or the ranking once you’re there. It also explains less of it than it used to, in a gap that’s now wider than pre-ZIRP and wider than the 2021 bubble.
What explains it, then
I checked FCF margin, gross margin, R&D spend, S&M efficiency, and SG&A and all were basically negligible. FCF margin explains 29% of the variation before removing Palantir, but just 9% after. Everything else is close to zero once Palantir’s out.
So, is it an AI story as Tunguz suggests? Well, ICONIQ’s report points the same way, even though their headline number is more growth weighted. They point to factors such as data moats (e.g. Crowdstrike’s Threat Graph processing trillions of daily security events or Samsara’s 25 trillion data points in physical operations), AI-driven workloads (e.g. the 3x growth in Snowflake accounts using AI) and AI cost-efficiencies (e.g. average revenue per employee up 30% at Samsara) as drivers of premium valuations.
Meritech also splits the software index into companies with a clear AI tailwinds and everyone else. The AI cohort is up 153% since January 2024, against an 18% decline for the rest. That gap has widened sharply through 2026, so something about AI exposure is clearly moving prices.
Source: Meritech Software Pulse – 11-September-2026
Look at the fastest growers across the whole index who didn’t make the top 10 by multiple. Figma is the second-fastest grower in the index, at 48%, and trades at just 5.9x, clearly illustrating that growth on its own isn’t enough. Its a similar story with Zeta. Both companies have an AI narrative, but sit in categories exposed to AI. Neither has the kind of structural moat a company such as Samsara has either. So, growth and AI labels aren’t enough if the market doesn’t trust what’s underneath (similar to the Thoma Bravo argument about software more broadly I have covered before).
On the other side of the coin, CrowdStrike commands one of the highest multiples in the market on the slowest growth rate in its own top 10 because its AI story is the one the market believes most. Fast growth without a convincing AI story gets discounted, and a strong AI story commands a premium even without fast growth.
Growth certainly explains the market, but belief in the AI story explains who’s winning it.
Sources: