Lovable says it has reached $500 million in annualized run-rate revenue as of June 2026, according to TechCrunch, up from $400 million in February, a $100 million jump in roughly four months. The usage numbers behind it are the part worth staring at: 1 million new projects are now created on the platform every week, and more than 50 million projects have been built since launch. The company was founded in late 2023, which means it crossed half a billion in run-rate before turning three. Back in August 2024 it projected hitting $1 billion within 12 months, a claim that sounded like founder bravado at the time and now reads like a trajectory.
The product itself is the simplest version of the vibe-coding pitch: primarily non-technical users, founders, designers, salespeople, describe the software they want and the platform builds it. What they build, per TechCrunch, is not toy demos: e-commerce storefronts, websites, and a long tail of internal tools, CRMs, inventory systems, HR platforms, the unglamorous business software that companies used to buy, not make. That last category is why the article frames Lovable inside the so-called SaaSpocalypse: every internal tool a sales lead describes into existence on a Tuesday afternoon is a seat-license some SaaS vendor does not renew.
The honest caveats are structural. No churn figures, no profitability, no valuation update in the report, and 1 million projects a week is a creation metric, not a retention metric, most of those projects are experiments, throwaways, or duplicates, and revenue per project is what separates an economy from a fidget toy. Annualized run-rate is also the most flattering possible lens on revenue. None of that erases the direction: the February-to-June slope is the kind that does not happen on churn-heavy products, and the platform is monetizing people who were never going to hire a developer in the first place, which means the spend is mostly additive to the software market, not cannibalized from it.
For the bigger picture we keep tracking, this is the demand side of the agent story. The Harvard and Perplexity study we covered yesterday measured agents changing which tasks people attempt, more higher-order work, more work outside their own occupation. Lovable at 1 million projects a week is that finding wearing a revenue line: a salesperson building a CRM is cross-occupational task expansion, monetized. The runtime layer we keep writing about, who hosts the agent, who executes the code, who holds the state, is being pulled forward by exactly this: millions of people who do not call it an agent runtime, they call it the thing that builds their store.