As part of our 2025 Cloud Challengers Report Notion's Talent Director, Michelle Cheng, discusses how Cloud Challenger companies are sharpening their talent strategies and prioritising lean, AI-fluent teams over sheer headcount.

Key insights:

1. The lean minimum viable organisation: Cloud Challenger companies are trending ten fewer people than in previous years
2. AI is permeating not just what products are being made, but also how work gets done
   

The old benchmark of doubling headcount each year is already outdated—tomorrow’s $100M companies won’t need 1,000 employees to get there. A mass decoupling of growth from hiring is underway, paving the way for the rise of the lean, outcome-driven organisation.

Team Size

This year saw a striking shift in team formation. The overall median team size dropped to 14 - over 40% fewer people than last year’s median of 25. In the world of startups, this is no small adjustment; it’s a fundamental redefinition of the building blocks needed for growth. To borrow from Bezos’ “two-pizza team” ideal team size rule - these startups are running on a lot fewer pizzas than before.

The story gets even more interesting when you look at the spread. In the chart below, you can see much higher variability in the blue dots (last year’s cohort) compared to a closer concentration of black dots (2025). In the course of just a year, the concept of what constitutes an ideal ‘Minimum Viable Team’ for early stage startups has gotten tighter. This shift suggests a growing consensus around the essential composition of early stage startup teams:new challenger companies are arising with consistent clarity on a productivity-first mindset compared to previous years.

In the chart , you can see much higher variability in the grey dots (last year’s cohort) compared to a closer concentration of green dots (2025).

Teams & Sector

Within the 2025 cohort, the vertical software companies had the highest median headcount, which may allude to the heftier professional services requirements associated with selling and implementing vertical SaaS. Interestingly,Dev Tooling companies were the most common type of company in the Cloud Challengers cohort, and also had some of the smallest teams. Interestingly Fintechs were able to operate with the leanest teams.

Are AI companies killing all our jobs? Not quite. This year’s cloud challenger’s cohort was almost evenly split betweenAI Native and non-AI native companies, and their respective median head counts were not dramatically different.

My interpretation of this is that it still takes roughly a similar number of people to build an excellent early stage company, whether it’s an AI product or not.

Median headcount of AI Native versus Non Native companies

However - what’s different is the rate of adoption.The Non-Native AI companies employed almost 200more people in total than the native AI startups.There was much broader standard deviation across the non-AI companies, which to me means that there’s still lots of iteration happening on how best to operate efficiently, whereas the AI native companies are evolving via a more uniform, ‘AI from day one’ philosophy.

Founding Teams

This year, we saw 26 solo founders- more than in previous CloudChallengers cohorts. Founding teams were generally smaller, reflecting a broader trend of leaner teams and greater risk tolerance. As AI adoption streamlines workflows, founders can experiment more with limited runway, making entrepreneurship more accessible than ever.

This is likewise true for founders from underrepresented backgrounds - we had three companies with sole female founders, the most since the conception of this report.

‍

Gender

This year 15 out of 100 companies had mixed-gender founder teams and we had three companies with sole female founders.

For more insights, check out the full 2025 Cloud Challengers Report.