Lovable Doubles to a $13.3B Valuation in 6 Months

On August 12, Stockholm-based “vibe coding” startup Lovable announced a fresh $400 million funding round that pushed its valuation to $13.3 billion. That’s roughly double the $6.6 billion valuation it disclosed just eight months earlier, in December. Lovable’s pitch is simple: describe the app you want in plain language, and the software writes it for you — no code required.

Who led this round

The round was co-led by Menlo Ventures and EQT’s Scaleup Europe Fund, with more than a dozen other investors participating, including Tencent and Balderton Capital, spanning Europe, Latin America, and Asia. Notably, this is Lovable’s second major raise in under a year — its December Series B brought in $330 million. The new round’s dollar amount is only modestly higher, but the valuation doubled outright, a sign that appetite for “describe it and it builds itself” tools is running well ahead of what most people would have guessed.

The numbers behind the valuation: 60 million projects, a million new ones a week

What’s propping up that valuation is a growth curve that looks almost out of control. Lovable’s annual recurring revenue hit roughly $500 million in June, up from just over $200 million at the start of the year — nearly tripling in half a year — and the company is tracking toward close to $600 million by the end of August. The platform now hosts more than 60 million projects, adds roughly a million new ones every week, and pulls in about 900 million visits a month. Those numbers would be striking for any software company, let alone one that only started out two years ago on the premise of building apps purely through chat.

A former CERN physicist’s company

Lovable was co-founded in 2023 by Anton Osika and Fabian Hedin, based in Stockholm. Osika’s background is unusual for a startup founder — he was previously a particle physicist at CERN before moving into software entrepreneurship. In just two or three years, the company has landed on TIME’s 2026 list of the most influential companies and become one of the fastest-growing names in the vibe-coding boom.

The shine has a couple of scuffs on it too

Lovable’s road hasn’t been entirely smooth, and it’s stumbled more than once. The first time was in May 2025, when several industry figures — including Replit CEO Amjad Masad — publicly warned that tools like this make it too easy for non-technical users to accidentally expose private data. Two Replit employees, Matt Palmer and Kody Low, tested it themselves and found at least 170 Lovable-hosted sites with similar vulnerabilities, while a separate software engineer, Danial Asaria, posted on X that he’d pulled usernames, emails, financial information, API keys, and even personal debt figures and home addresses out of Lovable-built sites within 47 minutes. The second and more serious incident came in April 2026: a security researcher found that any free account could read another user’s source code, database credentials, and even their full AI chat history. This time Osika issued a direct public apology, writing: “Lovable had an incident this week… things shouldn’t have gone this way in the first place, and I take accountability. I’m sorry.” He acknowledged the team learned about the incident on X in the first place, meaning the company’s own vulnerability-disclosure process had broken down. Lovable subsequently patched the flaw, made every historically public project private again, restructured its HackerOne triage process, and brought in security firm Aikido Security for automated penetration testing. Both episodes point to a challenge the whole vibe-coding category has to reckon with: when the barrier to writing software drops this low, users often lose the security instincts a trained engineer would normally bring along with it.

Not just renting AI from someone else — Lovable is training its own models too

Alongside the funding news, Lovable has quietly been repositioning itself technically: the company has started training its own AI models rather than relying entirely on outside large-language-model providers, while also signing a multi-year deal with Google Cloud for roughly 5x its current usage. That two-track strategy — buying more cloud capacity while also building in-house models — reflects how much Lovable wants to secure its footing in AI coding: not wanting to hand its whole future to an outside vendor, while still needing enough raw compute to keep up with a million new projects a week.

Quietly expanding into hardware, too

This round also came with an investment in Atech, a Danish startup extending the “vibe coding” concept into hardware design — meaning users might eventually be able to talk their way not just into an app, but into a physical product’s design process. If that actually works, it stretches Lovable’s story well beyond pure software.

The vibe-coding arms race is just getting started

Lovable’s surge isn’t happening in isolation. Over the past six months, the entire “describe it in plain language and get working code” category — from OpenAI’s Codex to a growing list of competitors — has been expanding fast, with capital and users pouring in simultaneously. Lovable doubling its valuation and its revenue in the span of six months is, in a sense, a microcosm of that broader arms race: once “you don’t need to know how to code to build software” becomes the mainstream pitch, the real question stops being whether these tools work at all, and becomes whether these companies can hold quality, stability, and a real business model together while growing this fast. At $13.3 billion, investors aren’t really buying today’s Lovable — they’re betting it can keep up with the pace it’s already set for itself.