Strategy · Featured Debate

5 guests 12 episodes 2,413 words

30 People, $100M ARR: The New Math of Team Size

Should you stay lean or scale headcount aggressively to capture opportunity?

Gamma has 30 employees serving 50 million users and generating $100M in annual recurring revenue. Boris Cherny at Anthropic ships 10 to 30 pull requests daily without writing a single line of code by hand. Cognition's 15 engineers each work alongside five AI agents simultaneously. Nick Turley runs ChatGPT -- with 700 million weekly active users -- modeled on the WhatsApp playbook of a small team running a global-scale product.

These are not outliers anymore. They are the leading edge of a fundamental reset in what a "team" means.

For two decades, the playbook was clear: find product-market fit, raise money, hire aggressively. Headcount was the proxy for ambition, capability, and seriousness. That era is ending -- not because of some abstract philosophical shift, but because the math has changed. The tools available to small teams now make them disproportionately powerful, and the coordination costs of large teams have become the dominant bottleneck in most organizations.

Should you keep your team small and lean, or scale headcount aggressively when you find traction? In an era of AI multipliers, what is the right model for building a high-output product organization?

The 4 Positions

1 / 4

Stay Radically Small -- 30 People Can Build a Unicorn

Grant Lee · Eric Simons

Grant Lee, CEO and co-founder of Gamma -- which reached $100M ARR in just over two years with a $2B+ valuation while remaining profitable -- did not stay small by accident. It was a deliberate strategic choice, and he attributes much of Gamma's success directly to it.

"Certainly for us, the ability to adapt and move fast in this environment. I attribute a lot of what we've been able to accomplish to a few things. One is we do have a small team and a lean team, one that's able to move really fast."
Grant Lee
▶ 01:16:33

The advantage is not just cost savings. It is velocity. Gamma can go from idea to full-scale experiment in a single afternoon -- ideate in the morning, build a functional prototype, recruit 20 prospective users with zero skin in the game, ship by afternoon, and review results by evening. That speed is a direct product of team size. There are no coordination meetings, no approval chains, no cross-functional alignment sessions about alignment sessions.

But Lee's sharpest insight is about how headcount targets corrupt hiring quality:

"If your goal is to double in headcount and add a hundred people to the company, then that becomes the goal. The goal is no longer to add the best people. It's no longer to maintain a high quality bar."
Grant Lee
▶ 01:36:03

Instead, Gamma follows a mantra of "hire painfully slowly" and obsesses over the first 10 employees getting the DNA right. All 10 of Gamma's original employees are still at the company five years later -- a continuity that compounds in tribal knowledge, shared context, and cultural cohesion.

Eric Simons, co-founder and CEO of StackBlitz/Bolt -- which went from near-death to $40M ARR in five months with 15-20 people -- reinforces this from the survival angle. During 2020-2021, when growing headcount was a KPI that peers celebrated, Simons stayed conservative:

"I think that we ended up doing a good job of being extremely conservative. During a time in which exuberance and growing headcount was like, KPIs of companies. And I'm glad that we didn't heed the advice, because if we had tripled the company and kicked up the burn rate, there would be no Bolt."
Eric Simons
▶ 00:28:02

The discipline of staying small gave StackBlitz the runway for seven years of shots on goal -- until WebContainers proved perfect for AI-era coding and the breakthrough arrived.

When This Applies

when you are pre-PMF, in a fast-moving market (especially AI), or when your competitive advantage depends on iteration speed. Gamma, Bolt, and the B2C companies Lenny profiled all found that small teams plus discipline beat large teams plus resources.

Small Teams Plus AI Multipliers -- The New Paradigm

Scott Wu · Boris Cherny

Scott Wu, co-founder and CEO of Cognition (which makes Devin, the world's first autonomous AI software engineer), and Boris Cherny, Head of Claude Code at Anthropic, are not just arguing for lean teams. They are describing a fundamentally different model where AI agents function as team members.

"Our whole team is only like 15 engineers a year. We use a ton of Devin when we're building Devin. Most folks on the team are definitely working with up to five Devins at once, and so Devin merges like several hundred pull requests into production in the Devin code bases every month."
Scott Wu
▶ 00:00:00

A quarter of Cognition's PRs come from Devin, expected to be over half by year's end. This is not a 15-person team with good discipline. It is a 15-person team with the output of a much larger one -- not through longer hours but through parallel AI agents working alongside each human.

Boris Cherny takes this even further. He has five Claude Code agents running simultaneously and ships 10 to 30 PRs daily:

"100% of my code is written by Claude Code. I have not edited a single line by hand since November. Every day, I ship 10, 20, 30 pull requests."
Boris Cherny

"Productivity per engineer has increased 200%."
Boris Cherny

The structural argument Wu makes is important: unlike previous technology shifts (personal computers, internet, mobile) that grew steadily as hardware distribution expanded, AI has no hardware distribution constraint. The space grows exponentially, which means the companies that adapt fastest -- inherently small teams -- have the greatest advantage. This is not a temporary condition but a permanent feature of AI-era competition.

When This Applies

when your work involves significant code generation, content creation, or data analysis. Cognition, Anthropic, and Every demonstrate that a single person with AI agents can match the output of a traditional team of five to ten. Consider a dedicated Head of AI Operations to make these gains compound.

The AI-Native Operating Model

Dan Shipper

Dan Shipper, co-founder and CEO of Every -- which runs four software products, a daily newsletter, and a consulting arm with just 15 employees and 100% AI-written code -- pushes the insight beyond tools to organizational design.

"No one is manually coding anymore. Organizations like ours, people who are playing at the edge, we're doing things that, in three years, everybody else is going to be doing."
Dan Shipper
▶ 00:00:46

The key enabler is not the AI tools themselves but a structural role: Every has a dedicated Head of AI Operations whose full-time job is to watch how every employee uses AI, identify the most effective patterns, and systematize them across the organization. This creates a compounding advantage -- the team gets better at using AI faster than competitors because someone's job is specifically to accelerate that learning.

When This Applies

when you have exhausted AI leverage and existing team capacity, when the work requires deep human relationships (enterprise sales, customer success at scale), when you are building something that requires hard technical breakthroughs with years of sustained investment, or when you have clear external validation and under-resourcing risks losing the opportunity -- as Robby Stein warns about products that stay too small and die on the vine.

Measure and Reward Impact Per Headcount

Geoff Charles · Matt Mochary · Nick Turley

Geoff Charles, VP of Product at Ramp -- the fastest-growing SaaS startup in history, reaching $100M+ ARR in two years -- operationalized the lean philosophy with a formula: leverage equals impact divided by headcount.

"We try to increase leverage as much as possible -- defined by impact divided by headcount. We reward teams that were able to get more done with very little. For the first two years at Ramp, we had fewer than five PMs and fewer than 50 engineers as we scaled to $100 million in ARR."
Geoff Charles

This is not just a metric. It is a cultural principle that shapes every hiring decision:

"Every headcount that you add inherently slows down the rest of the organization because it increases the cost of communication and alignment within the organization."
Geoff Charles

As Lenny's Newsletter detailed in the "How Ramp builds product" deep dive, Ramp deliberately restricts PM hiring until it is "patently obvious to everyone that product is the bottleneck to velocity." This constraint forces three productive behaviors: everyone does some PM work, each PM has a larger scope eliminating resource politics, and the team genuinely welcomes new PMs when they arrive.

Matt Mochary, the CEO coach who works with founders at Coinbase, Notion, and other top companies, validates this from a meta-perspective across dozens of organizations:

"With fewer people in the organization, things work better. That's the big realization that most people never discover. They hit product market fit, they get tons of money from investors, just hire, hire, hire, but every additional human you have in your organization causes extra overhead and geometrically so."
Matt Mochary
▶ 00:21:17

Mochary names WhatsApp, Instagram, Linear, and Notion as the "real success stories" -- companies that kept teams small through their most important growth phases.

Nick Turley, Head of ChatGPT at OpenAI -- serving 700 million weekly active users -- draws from the same playbook:

"So it is a small team that is running ChatGPT. I take co inspiration from WhatsApp where it was a very small team running a very global-scale product. And then more importantly, you have to treat hiring a little bit more like executive recruiting and less like just pure pipeline recruiting."
Nick Turley
▶ 01:05:51

When This Applies

at any stage. Even if you are scaling, tracking Ramp's leverage formula (impact divided by headcount) prevents the bloat that kills velocity. Restrict hiring until the current team proves they cannot achieve their goals with existing people and better tools.

Evidence from the Archive

Gamma

Gamma reached $100M ARR while profitable with 30 people who fit in a restaurant

Gamma tests across 20 different AI models in production to optimize cost-to-value ratio

Grant Lee · CEO & Co-founder ▶ 00:01:36

Ramp

Built a competitor to Amex in 3 months with ~8 engineers

Built a competitor to Expensify with under 50 total R&D, fewer than 4 engineers and 3 PMs

Geoff Charles · VP of Product ▶ Watch

Every

15 employees running 4 software products, a newsletter, and a consulting arm

Zero handwritten code on the product team

Dan Shipper · CEO & Co-founder ▶ 00:00:46

Mochary Method (CEO Coach)

Mochary let go of an outperformer and the organization ran better within 2-3 months

WhatsApp served hundreds of millions of users with ~55 engineers

Matt Mochary · CEO Coach to Silicon Valley's top founders ▶ 00:21:17

StackBlitz / Bolt

Went from near-death to $40M ARR in 5 months with 15-20 people

Core group of 5-7 people stayed for 5+ years through the lean period

Eric Simons · CEO & Co-founder ▶ 00:28:02

Anthropic

Ships 10-30 PRs daily with 5 agents running simultaneously

Has not edited a single line of code by hand since November

Boris Cherny · Head of Claude Code

The Synthesis

The consensus across eight of nine voices is striking: every additional person you hire makes the rest of the organization slower. The question has flipped from "how many people do we need?" to "what is the minimum number of people who can achieve this outcome?"

Inverted Question

How has the headcount question fundamentally changed?

Human vs. AI Work

What kind of work actually requires humans in the AI era?

Lean as Feature

When is staying small a competitive advantage rather than a constraint?

Conviction-Level Resourcing

When does the lean consensus break down?

The question has flipped from 'how many people do we need?' to 'what is the minimum number of people who can achieve this outcome?' The consensus across eight of nine voices: every additional person you hire makes the rest of the organization slower.

The tasks AI handles well -- code generation, data analysis, pattern matching, content drafting -- are exactly the tasks that drove most headcount scaling. The tasks that still require humans -- judgment calls, relationship building, creative direction, navigating ambiguity -- do not scale linearly with headcount anyway.

Every successful B2C subscription company stayed ruthlessly lean until finding PMF. Calm kept its team under 10. Grammarly was 2-4x leaner than comparable companies. The discipline of staying small was what allowed them to be 'misunderstood for years' while building something that actually worked.

For products built on hard technical breakthroughs that require sustained investment from many people over years -- foundational AI models being the clearest example -- small teams can die on the vine. The key is not to fetishize leanness but to resource appropriately to the conviction level.

Which Approach Fits You?

Answer 3 questions about your situation. We'll match you to the right approach.

Question 1

What is your current team size relative to your ambition?

Question 2

How do you measure team productivity?

Question 3

What type of work does your team do?

Notable Absences

The Bottom Line

Robby Stein's dissent is important precisely because it identifies when the consensus breaks down. For products built on hard technical breakthroughs that require sustained investment from many people over years -- foundational AI models being the clearest example -- small teams can die on the vine. The key is not to fetishize leanness but to resource appropriately to the conviction level.

This insight is reinforced by a pattern Lenny documented in his newsletter on consumer subscription businesses: every successful B2C subscription company -- Calm, Grammarly, Noom, Spotify, Future -- stayed ruthlessly lean until finding strong product-market fit. Calm kept its team under 10 in a one-bedroom apartment. Grammarly was 2-4x leaner than comparable companies. Noom's growth team had fewer than 10 people through $3-4M in revenue. The discipline of staying small was not a constraint -- it was what allowed them to be "misunderstood for a number of years" while they built something that actually worked.

Sources

Grant Lee — "“Dumbest idea I’ve heard” to $100M ARR: Inside the rise of Gamma | Grant Lee (co-founder)" — Lenny's Podcast, November 13, 2025
Scott Wu — "How Devin replaces your junior engineers with infinite AI interns that never sleep | Scott Wu (Cognition CEO)" — Lenny's Podcast, September 8, 2025
Dan Shipper — "The AI-native startup: 5 products, 7-figure revenue, 100% AI-written code | Dan Shipper (co-founder/CEO of Every)" — Lenny's Podcast, July 17, 2025
Geoff Charles — "Velocity over everything: How Ramp became the fastest-growing SaaS startup of all time | Geoff Charles (VP of Product)" — Lenny's Podcast, August 6, 2023