Independent · verified · confidence-tagged
Teardowns for the
AI economy.
See exactly how AI businesses are actually built: the wedge, the go-to-market engine, the stack and the moat. Then steal the opportunities they leave on the table. Every figure sourced and tagged by confidence.
Exactly how they were built
Not founder mythology. The real wedge, channels, spend, stack and growth loops, reverse-engineered from public evidence.
Inspiration you can act on
Every teardown ends with adjacent businesses the company leaves on the table — named customer, tech wedge, MVP scope, difficulty.
Verified, not vibes
Every number links to its source and carries a confidence tag: verified, inferred or estimated. No naked claims.
- The wedge
- Full GTM engine
- Tech stack
- AI infrastructure
- Moat assessment
- Adjacent opportunities
- Sources & verification log
Latest teardown
Lovable
The Stockholm vibe-coding studio that reached $400M ARR with 146 people by locking non-technical builders into the only stack they will ever need.
Lovable's core moat is compressing the entire software creation stack (AI model orchestration, code generation, backend provisioning and deployment) into a single natural-language chat interface locked to a fixed, opinionated stack (React plus Supabase), making the product so frictionless for non-technical users that it drives viral, bottom-up enterprise adoption; but this same architectural rigidity (no backend flexibility, delegated RLS security) is both its speed advantage and its structural ceiling for production-grade and regulated workloads.
Read the teardown →More teardowns
WPAI
The lean WordPress AI specialist that built a proprietary training-data moat and exited to Automattic 26 months after shipping its first product.
Read →Moltbook
A weekend-built Reddit for AI agents that Meta acqui-hired 42 days after launch.
Read →Photo AI
A solo founder's AI photography studio that reached $100K/month in revenue by attaching a thin API wrapper to a decade of distribution.
Read →tabiji.ai
A founder-plus-agents travel-safety publisher proving distribution, not generation, is the moat.
Read →Cursor (Anysphere)
The AI-native code editor monetising developer velocity at consumer-app prices.
Perplexity
An answer engine betting that AI search can unseat the ten blue links.
Every teardown, free to read in full.
No paywall and no preview wall. The whole breakdown is open: the wedge, the full GTM engine, the stack, the moat and the adjacent-opportunity map. Tairdown documents the indie AI economy: not just the giants everyone already covers, but the solo founders and micro-businesses you can actually learn from.
Start with Lovable →