A 15-person team generating $100m in ARR within two months of launch?
It’s the stuff that VC dreams are made of: tiny, super productive teams making investment cash go even further.
It’s also, quite frankly, a fantasy for plenty of startup operators fed up with bloated teams thanks to a post-Covid surplus of venture dollars.
But is smaller really better? And how long does it make sense to stay tiny for?
The tiny team formula
A newer generation of entrepreneurs who founded companies in the last year or two are eager not to make the mistake of a previous generation of founders by raising loads of capital and overhiring. They’re instead building tighter, compact teams that leverage AI to do tasks that once required dozens of people — and avoiding layers of bureaucracy as a result.
“Fewer people need less coordination effort. In principle, you’re able to ship faster and serve your customers better because there is less overhead in terms of keeping people aligned. So we like to be as small as possible,” says Lennard Schmidt, founder of Langdock, an AI adoption platform for mid-market and enterprise customers.
Langdock started out in 2023 as a team of five; now it’s reached $5m in ARR and is serving 1k customers with 20 people. It has five new employees starting in September and is currently hiring for three roles — two product engineers and an AI associate — all focused on the needs of customers.
“We usually hire people across our product engineering and customer success teams — so either making the product better or helping the customer understand and extract more out of the product,” explains Schmidt.
In a tiny team, anything that can get automated, does. At Langdock, that includes preparing invoices and documents for tax consultants, and writing marketing copy. The company also outsources services such as tax, legal counsel and accounting, and focuses on using humans to do very human things. Its marketing person, for example, spends a lot of time running offline events and organising dinners with customers.
The metric VCs love
Dinika Mahtani, partner at Cherry Ventures, says that as team sizes shrink, people will be expected to produce a lot more.
VCs, particularly from later stage funds, are increasingly paying attention to a metric called revenue per employee which they see as a key sign of efficiency, says Mahtani. For example, AI coding platform Cursor, which has 60 employees and is valued at over $1bn, is generating an “amazing” $3.1m in revenue per employee.
“The biggest shift you’ll see on the employee side is in companies that are consumer apps or product-led growth businesses — like Lovable, where basically something hits a viral loop and grows fairly organically or with very targeted marketing spend. These companies need to ship a product, maintain a product but they don’t have to build a whole sales organisation,” says Mahtani.
Cursor, Lovable and Windsurf (which was scooped up this week by AI startup Cognition AI) are what Mahtani calls the “poster children” for “insane revenue per employee metrics.”
But, not all companies will be able to maintain high levels of growth with few hands on deck. “It depends on the type of company, the type of business model and how you need to take it to market,” she adds.
For example, regulated companies in healthcare and finance can’t skimp on having teams to take care of compliance.
It also depends on what kind of customers you serve.
Schmidt says consumer-focused self-serve software like Cursor or Lovable requires minimal staffing as users can resolve issues and complete tasks via FAQs, chatbots and customer portals without the need to talk to a human. Companies selling to enterprise customers — as Lovable is gearing up to — require more people for implementation, compliance and support.
“What we are seeing is that the lower your ACV — the lower the amount that a single user pays — the less people you will likely need at scale, just because the user that buys software doesn't expect to talk to someone. The larger the customer you're serving, the more expectation they have to talk to someone,” says Schmidt. “This is one thing that heavily impacts the way you can build an organisation.”
He adds: “There’s workflows we can augment to serve our customers, but in the end, we also sometimes need to spend time on them — and we want to, because that helps us craft a product roadmap on what they need.”
Lean thinking
For Benjamin Tennmann, the cofounder and chief technology officer of Science Machine, a London-based startup building an AI data scientist that just raised $3.5m and is currently two people strong, keeping a team lean comes down to focusing on a few important things at a time.
“Instead of hiring people to do more, you actually want to prioritise things better,” says Tenmann.
“For example, if we want to do a certain research project for PR purposes, instead of us going out and spending three months trying to hire someone who might not work out, we think about whether we even need to do this thing in the first place.”
Science Machine plans to hire just three people in the next year: two founding engineers and a computational scientist, a domain expert to help with client onboarding and inspecting the quality of its AI data scientist.
The company is mostly looking for generalists; people who are self-starters that can “get stuff done,” says Tenmann.
AI tools also mean that individuals trained in one area can pitch in on others: for example, a machine learning engineer can now do UI work, or a product manager can help out with coding.
“The differentiator now will be less about specialised knowledge and much more about bias to action, at least in startups. The tools will be able to help you a lot with the specialised knowledge, and you'll be able to pick things up much more quickly,” says Tenmann.
Tenmann is eager for his startup to stay five people for as long as possible. Having worked in startups and scaleups previously, he’s seen how big fundraises followed by periods of “overhiring” have significantly harmed companies in the past.
He acknowledges he’ll likely have to hire more people as the company grows — but not too many. “We’d rather pay fewer people more,” he says.
What happens now?
While not every startup can thrive with a tiny team, the model is reshaping founder expectations — and investor criteria.
VCs now quiz founders on AI tooling, coding speed and how they’re getting to their first 100 customers — and AI native companies with big headcounts are seen as “an orange flag” to investors, says Mahtani.
But as AI startups proliferate, Tenmann thinks that increased competition may lead some to grow larger teams to build products, sell and differentiate themselves faster.
With Sam Altman predicting the first billion-dollar startup will be built with one person, the very concept of a “company” may radically shift.
