Prompt to app

Placement: a judgment drawn from the evidence, open to correction (dispute it).

The model

Definition: Natural-language prompts are converted directly into deployed, working software (websites, apps, internal tools) without the user writing code, priced by usage (tokens, compute, or agent-hours) rather than by seat.
In one line: People describe software in plain language and get a working, deployed application, paid by usage.
What compounds: Each app built teaches the platform; more builders bring more templates and integrations.
Who accelerates: Anyone can build software; professional developers move up to harder problems.
What it displaces: Agencies and freelancers building simple apps and sites; no-code tools.
Stops working when: If most generated apps are abandoned rather than run in production, so revenue churns.

Companies 2
Evidence items 2
Grades 0A 2B
Latest dated 2026-04-13
Counter-evidence 0

1. Flow through · loop gain about zero. The curve helps it, but the gain passes straight through to customers and rivals.

Selling abundance · money lands in volume of cheap intelligence.

Cost-curve loop · cheaper models, lower unit cost, more volume, more revenue.

Company	Claim	Date	Gr	Source
Lovable Sweden · scaleup · best B	Lovable's ARR trajectory: $100M (Jul 2025), $200M (Nov 2025), $300M (Jan 2026), $400M (Feb 2026), with 146 employees	2026-03-11	B	TechCrunch
Vercel (v0) US · scaleup · best B	Vercel's ARR run rate reached $340 million by end of February 2026, up from $100 million at the start of 2024 (240% growth over roughly two years), with CEO Guillermo Rauch stating about 30% of apps running on Vercel's platform were generated by AI agents rather than hand-written; the article did not break out v0's standalone revenue	2026-04-13	B	TechCrunch

None recorded in the launch batch.

TechCrunch techcrunch.com · 2026-03-11
TechCrunch techcrunch.com · 2026-04-13