New ICONIQ data (as of 08/04/2026, covering 2022 through Q2 2026) on median headcount change by revenue growth band. ICONIQ’s own framing was that AI isn’t freezing headcount at hypergrowth companies, and that part is right. The more useful story is in the other three rows of the chart.
N-size: 390 (2022-2023), 200 (2024), 195 (2025), 57 (2026)
#1. The 100%+ Growers Are Hiring Harder Now Than They Did in 2022
119% in 2022-2023, then 65% in 2024, then back to 119% in 2025, now 133% in the first half of 2026.
2024 was the discipline year. Even the fastest-growing companies in the dataset cut their headcount growth roughly in half. It did not stick. The companies growing 100%+ today are adding people faster than they were at the top of the 2021-2022 market.
Who’s in that bucket now is most of the explanation. Growth above 100% in 2026 is overwhelmingly AI-native B2B, usually well-funded, usually in a land grab against three or four direct competitors. When you’re tripling, nobody in the building is running a headcount efficiency exercise. Are the fastest growing AI+B2B startups leaner than a generation ago, overall? Yes, in most cases. But they aren’t staying super lean as they scale. They’re still growing headcount triple digit percentage a year.
#2. The Biggest Change in the Chart Is 46% to 25% in the 50%-100% Band
Companies growing 50%-100% cut their headcount growth from 46% to 25% in a single year. Nothing else in the data moved that much.
The opposite of the fastest growerrs. These are not struggling companies. Growing 60%-80% at scale is a great year. But they aren’t AI rocketships.
The 25%-50% band did the same thing at smaller scale: 16% down to 10%. A company growing 35% is now adding 10% more people. In 2022-2023 that number was 17%, and five years before that it would have been 25%-30%.
If AI leverage is showing up anywhere in this dataset, it’s here. Not in layoffs at companies that missed their number, but in healthy companies quietly planning smaller teams for the same growth.
#3. Sub-25% Growers Stopped Cutting, and … Then Stopped Hiring
(6%) in 2022-2023, (5%) in 2024, +4% in 2025, +3% in H1 2026.
The layoff cycle in this cohort is over. What came after it wasn’t a rebound. It’s flat. A company growing 20% is now adding 3% more people, which at 200 employees is six hires for the year across every function.
For a lot of B2B companies that crossed $50M-$100M in the 2021 cohort and settled into 15%-25% growth, the flat team is now the plan, not a phase.
#4. At 133% Headcount Growth, the Top of the Market Is Getting Less Efficient As It Scales
Rough math, using the bucket floor: in 2024, companies growing at least 100% added 65% more headcount. Revenue per employee went up meaningfully. In H1 2026, the same cohort is adding 133% more headcount. Unless the median company in that band is growing revenue faster than 133%, revenue per employee at the very top of the market is flat to down.
That runs against the story most people are telling about AI right now. The fastest-growing AI companies are not doing more with fewer people. They’re hiring engineers, forward-deployed engineers, and solutions people at a rate that outpaces almost everything from the last cycle.
One honest caveat: 100%+ is a floor, not a midpoint. Some companies in that bucket are growing 200%-300%, which would put headcount growth back below revenue growth. The chart can’t distinguish. But the direction of travel from 65% to 133% in two years is real.
A Few Things to Note
A few things to hold loosely:
- N=57 for 2026 versus 195 for 2025. That’s a small sample, and it’s a specific portfolio (ICONIQ’s private venture and growth companies plus a select set of public software companies), not the whole market.
- It’s a half year. The 2026 number covers Q1 and Q2 only.
- ICONIQ’s post says “average,” the chart says “median.” Those are different numbers, and in headcount data with a few extreme outliers, very different.
- The bands are self-selecting over time. The companies in the 100%+ bucket in 2026 are mostly not the companies that were in it in 2022. The chart is tracking growth bands, not the same companies moving through time.
Three Different Hiring Playbooks For Three Different Growth Rates
The useful output here isn’t a benchmark to hit. It’s the ratio you’re implicitly choosing when you build the plan.
Using bucket midpoints, companies growing 50%-100% added about 0.7 points of headcount for every point of revenue growth in 2025. In H1 2026, roughly 0.4. That’s the number that changed.
So three decisions to make on purpose rather than by default:
- If you’re growing 100%+, the constraint isn’t efficiency. The cohort that’s winning right now added 133%. Under-hiring against an AI efficiency thesis while a competitor doubles the team is how you end up in the 50%-100% band next year.
- If you’re planning 50%-100% growth in 2027, and your headcount plan grows the team by 45%, you’re planning like it’s 2025. Your comparables cut that to 25% in one year. Either you have a reason yours is different, or you have a plan nobody rebuilt.
- If you’re growing under 25%, a flat team is now the market-normal way to run. The 3% number means the growth has to come from price, mix, retention, and better sales productivity. Hiring your way out is no longer what your peer group is doing.
Source: ICONIQ, quarterly financial and operating data from private venture and growth portfolio companies plus a select dataset of public software companies, 2022 through Q2 2026, as of 08/04/2026. For educational purposes, not investment advice.
