Vibe Coders Hit the Stripe Wall. A Lovable Investor Wants Revenue Shares Instead of Subscriptions.

Nine months after Jason Liu published his case for outcome-based pricing at AI coding platforms, the argument still hasn't landed — at least not with the platforms themselves. Lovable, the viral app-builder that Liu holds a small stake in, still runs on seat subscriptions and credit packs. So do most of its competitors. That might be the point.

Liu, an AI consultant who also runs a course business generating around $800,000 a year, made the argument in a June 2025 essay that has circulated in AI-adjacent circles since. His pitch is direct: platforms like Lovable charge for access, but access isn't what users want. They want revenue. If a platform only gets paid when a user gets paid, it has every reason to help them get there.

The proposal he floats — a tiered 'Lovable Partners Program' — would have the platform take between 5% and 30% of user revenue in exchange for services that no current subscription tier covers: payment infrastructure setup, customer support, CDN optimization, infrastructure migration. At 15%, Liu argues, the arrangement would still undercut the cost of stitching together Stripe, Podia, Maven, and Kit.com independently.

What sharpens the framing is where Liu locates the problem. He coined 'vibe coders' to describe builders who can ship a working app through AI tools in an afternoon but grind to a halt at what he calls the 'Stripe Wall' — the tangle of webhooks, subscription logic, and edge cases that payment infrastructure requires. His own job board, pulling in $2,000–$3,000 a month, doesn't have a proper Stripe integration; it runs on Zapier workarounds because the engineering lift isn't worth it at current revenue. Outcome-aligned pricing, his argument goes, would give platforms a reason to solve that problem rather than leave users to figure it out alone.

The counterargument isn't hard to find. Subscription and credit-based pricing is predictable in ways that revenue shares aren't — for the platform, anyway. A company like Lovable can model its growth, staff to a forecast, and invest in infrastructure without tying its unit economics to whether a cohort of vibe coders breaks through. Revenue-share arrangements also raise awkward questions about data access: to take a cut of what users earn, you need visibility into what they earn. That is a closer relationship than most SaaS companies want with their customers, and most customers want with their SaaS providers.

There is also the question of selection effects. Platforms that price on outcomes would likely find themselves disproportionately serving early-stage projects where the expected value of a revenue share is low and the operational cost of white-glove support is high. The economics only work if users succeed — and most don't.

Liu's implicit answer to that is a data flywheel argument: every manual service the platform delivers teaches it something it can eventually automate, turning forward-deployed human effort into scalable product capability. It is a reasonable bet on a long enough timeline, but it requires patience and capital that most startups do not have in abundance. It also assumes the platform can execute on services — payment operations, customer support, infrastructure — that are well outside the core competency of a product-led coding tool.

Whether Lovable or any competitor moves in this direction is an open question. Liu published his case nine months ago and nothing has visibly shifted. What his argument gets right is the underlying tension: the platforms generating the most genuine user value — the ones whose tools let someone build a real business — are not capturing much of that value. Whether that is an opportunity or simply a reflection of how hard outcome-based models are to operate is where the interesting debate starts.