This editorial appeared in the June 26th, 2025, issue of the Topline newsletter.
Want all of the latest go-to-market insights without the wait? Subscribe to Topline and get the full newsletter, including bonus commentary and takeaways, delivered to your inbox every Thursday.
Exactly 12 months ago, the market cap of all private AI companies was $283B. Today, it's $658B - a 130% increase in just one year. During the same period, public AI companies added $2T in total market cap. T for trillion. You can't see this data and not take AI seriously.
Yet naysayers are everywhere, claiming we're in an AI bubble. When you show them the revenue growth, they dismiss it as "experimental revenue" that won't stick. "It's not real ARR," they say.
The naysayers resist because these AI companies look nothing like what they're used to. They're built differently - just like SaaS companies looked alien to those selling perpetual licenses. The patterns don't match, so they assume it's broken.
Dismissing what you don't understand can be an expensive mistake.
Instead, we need to try to understand how these AI companies are different, and that’s what we are going to do today.
By The Numbers
We've all heard about the revenue explosions at companies like Cursor, Bolt and Codium. You probably know someone at a company growing insanely fast - like our friends at Fyxer who went from $0 to $10M in five months.
But what about AI companies as a whole?
The Emergence Capital team dug into this and found that AI companies are growing 4X faster than SaaS companies. Not 10% or 20% faster - four times faster. On average.
The skeptics love to cry "it's not real ARR" and "who knows if it'll renew." Yet the average net revenue retention for AI companies is 132% compared to 108% for SaaS. That means an AI company with $10M in revenue grows to $13.2M from existing customers, while a SaaS company hits $10.8M.
They do this with 7-8X fewer employees per dollar of revenue than the median SaaS peer. The gold standard for public SaaS companies is $300K per employee. Leading AI startups are already well past that.
So AI startups grow revenue faster, expand it more, and do it with a fraction of the headcount. The quantitative differences are clear. But the cultural differences run even deeper.
The Cultural Shift
The first thing you notice about these companies is the work ethic. 12+ hour days, six to seven days a week, many in office full time. Harry Stebbings and Jason Lemkin have both commented on this shift, and I've seen it firsthand. They believe there's a massive prize to be won, and they refuse to let work ethic be what costs them the win.
This intensity extends to how they think about team building. They get zero status from large headcount. What they want is the leanest possible team for their stage, and they're not shy about discussing revenue targets or profitability goals. It's common to find founders aiming for ARR milestones within one to two years that would have taken five to ten in the SaaS era.
With such ambitious targets and lean teams, they can't rely on traditional competitive defenses. They know any feature can be copied in days or minutes, so they've abandoned the old playbook of building moats. Instead, they bet everything on execution speed and quality. Move fast enough, and competitors can't keep up.
That velocity extends to distribution too. These digitally native founders understand social media instinctively. When they build a feature, they're simultaneously planning how to announce it. Their announcements are cinematic, authentic and pointed. This enables them to punch above their weight and accelerate revenue faster than anyone thought possible.
What you're witnessing is a fundamental shift in startup culture. These companies operate with completely different assumptions about speed, scale, and what's possible. The approach is working, so don't expect a return to SaaS-era norms.
Hundreds of thousands of professionals will need to change their beliefs about building and scaling businesses. Those who don't will become the naysayers still talking about "experimental revenue" while trillions of dollars of value get created without them.
