Five years ago $10M ARR meant 30+ engineers. In 2026 we've shipped products with teams of 5 hitting that number. The playbook is real. The compression is real. AI is part of it - but only part.
- 1 Product engineer + founder: Owns product direction and ships features. Lives in Cursor + Linear all day.
- 2 Senior full-stack engineers: Ship features end-to-end. Frontend, API, database, deploy. T-shaped.
- 1 Platform engineer: Owns infra, CI/CD, observability, on-call. Keeps the others fast.
- 1 AI/data engineer: Owns AI features, data pipelines, eval harness. Increasingly required by 2026.
What lean teams do differently
Buy primitives, don't build them
Auth (Clerk), billing (Stripe), email (Resend), search (Algolia), feature flags (PostHog), error tracking (Sentry), CRM (Attio). $2-5K/month replaces 3-5 engineers.
Standardize on one stack religiously
Lean teams pick: Next.js + Postgres + TypeScript + Vercel. Every problem gets solved in this stack. No 'try Rust because it's faster' tangents.
Ship behind feature flags from day 1
PostHog feature flags are free up to 1M evaluations. Every feature ships dark, gradually rolls out. Reduces hotfix incidents by 60%+.
Use AI for 30-50% of code
Cursor + Claude Code handle scaffolding, refactors, test generation, doc updates. The senior engineers focus on architecture and review.
Outsource the hard one-time work
Compliance build-outs, mobile apps, design systems. Bring in external teams (like ours) for 8-14 week projects. Avoid the long-term salary commitment.
Async-first, max 3 meetings/week
Every meeting under 30 minutes. Async standups in Slack. Loom for anything you'd present. Engineering time is the constraint.
- Build their own auth system: Wastes 6 months. Use Clerk or Auth.js.
- Pick exotic tech: Rust, Elixir, Haskell - all great. None help you ship a $10M SaaS faster than TypeScript.
- Hire in-house designers full-time: Use Mobbin + shadcn/ui + a design contractor when needed. Save the headcount.
- Spend on dev evangelism early: Direct sales + product-led growth + light SEO. Skip the conference circuit until you're at $5M+ ARR.
The week-by-week shape
- 1
Monday: 30-min team sync + sprint planning
Async written updates over the weekend, 30-min sync to align on the 3-5 things shipping this week. That's the only weekly meeting.
- 2
Tuesday-Thursday: focus + ship
Deep work blocks. PRs go up daily. AI tools handle 30-50% of the typing. Reviews happen in Slack threads, not meetings.
- 3
Friday: demo + retro
30-min demo of what shipped. 30-min retro on what blocked. That's it. Total weekly meeting time: 90 minutes.
"Most teams stuck at $1-3M ARR with 15+ engineers have an org problem, not a product problem."
Lean team vs bloated team at $10M ARR
| Dimension | Lean (5 engineers) | Bloated (15+ engineers) |
|---|---|---|
| ARR per engineer | $2M | $667K |
| Headcount cost (eng) | $900K/yr | $2.7M/yr |
| Tooling spend | $120K/yr | $120K/yr |
| Decision latency | Hours | Days |
| Stack complexity | Boring + standardised | Polyglot + bespoke |
| Hiring rate | 1-2/year | Constant |
| Org politics | None | Real |
What lean teams do that bloated ones don't
Buy primitives religiously
Auth, billing, email, search, feature flags, error tracking - bought, not built. Saves 3-5 engineers.
Standardise on one stack
Every problem gets solved in Next.js + Postgres + TypeScript. No 'try Rust because faster' tangents.
Ship behind feature flags
PostHog free up to 1M evaluations. Reduces hotfix incidents by 60%+.
AI does 30-50% of the typing
Cursor + Claude Code handle scaffolding, refactors, tests, docs. Engineers focus on architecture and review.
Outsource one-time hard work
Compliance, mobile apps, design systems - bring in external teams for 8-14 week projects.
Async-first, max 3 meetings/wk
Engineering time is the constraint. Protect it ferociously.
We're the outside team that ships your hard one-time work.
MVPs, compliance build-outs, mobile apps, design systems. Fixed-price, 8-14 weeks. No long-term salary commitment.
The new normal at $10M ARR
- $2M ARR per engineer is achievable with the right stack and tooling.
- Buy primitives - save 3-5 engineers worth of headcount.
- AI tooling compresses the boring 30-50% of engineering work.
- Async-first culture is the productivity multiplier.
- Outsource hard one-time work; hire for the recurring core.
The compression is real
$10M ARR with 5 engineers isn't a fluke. It's the new normal for product-focused SaaS teams that bought tooling instead of hiring, picked boring stacks, and used AI to compress the boring 30-50% of engineering work.
Most teams stuck at $1-3M ARR with 15+ engineers have an org problem, not a product problem. The playbook to hit $10M is in front of them - they just have to give up the headcount comfort blanket.
The IRPR engineering team ships production software for 50+ countries. Idea → Roadmap → Product → Release. 200+ products live.
About IRPR