From Throwaway Prototypes to Production
Eric Simons, CEO of Bolt, on consumer vs. B2B vibe coding
The real battleground is moving from making demos fast to making software that a company can keep using after the demo. Prototype tools win the first five minutes because anyone can prompt out a mock app, but retention shows up only when that app plugs into the company’s codebase, design system, auth, and data. Once a product is only helping teams sketch disposable ideas, it is easy for Figma Make or the next prompt box to replace it.
-
Bolt’s own shift reflects this. After a broad consumer launch, it found the strongest retention with product and development teams, and expects company based usage to exceed half of revenue this year. Those buyers care less about one off prototypes and more about code that can ship to customers.
-
The money in the stack is already concentrating below the prototype layer. As vibe coding apps grew, backend platforms like Supabase benefited because every generated app still needs database, auth, and storage, and those services keep billing as long as the app stays alive. That layer has stronger retention than app creation itself.
-
Figma is dangerous precisely because prototype generation is now bundled into an existing seat base. Figma reached about $1.05B in 2025 revenue, and Figma Make is already used weekly by roughly 30% of enterprise customers spending over $100K annually. If a startup only offers fast throwaway prototyping, it is competing with a feature inside a standard design license.
Going forward, the winners in vibe coding will look less like idea generators and more like production workflow software. That means deeper enterprise setup, stronger integration with internal systems, and more revenue tied to hosting, auth, data, and ongoing app usage. The products that survive will be the ones whose output becomes part of the company’s actual software stack.