Eighteen months. That's all it took.
Lovable, a Stockholm-based platform that lets people build software by chatting with AI, announced on June 9, 2026, that it had reached a $500 million annual revenue run-rate—a company claim not independently verified. The milestone came with the sort of metrics that sound almost absurd: 1 million new projects spinning up every week, 50 million projects built in total, more than half the Fortune 500 using the tool in some capacity. The company doing all this? A team of 146 people, according to its own reporting.
The figures—self-reported and unaudited, as startups' often are—represent a growth curve that defies most assumptions about how software companies scale. TechCrunch, covering the announcement, noted that Lovable had added $100 million in ARR in a single month, jumping from $400 million to $500 million faster than most companies scrape together their first million in revenue.
Whether the trajectory holds, or whether this is the sort of hypergrowth that eventually meets hard limits, remains an open question. For now, though, Lovable stands as perhaps the most striking case study in AI-era compression: the idea that entire business cycles—from concept to product-market fit to enterprise adoption—can collapse into timelines that would have seemed impossible just a few years ago.
The Unlikely Genesis
Anton Osika wasn't trying to build one of Europe's fastest-growing startups. Before co-founding Lovable in late 2023 with Fabian Hedin, his resume read like a tour through ambitious technical projects: working on ATLAS at CERN, serving as the first engineer at Sana Labs, co-founding Depict.ai—a YC-backed company that pulled in $20 million in funding. Hedin, a fellow student from KTH, had designed interfaces for Stephen Hawking's computer system.
But the real spark came from something Osika built almost on a whim. In June 2023, he released GPT-Engineer, an open-source experiment in "text-to-app" coding. The project racked up more than 54,000 stars on GitHub and became an early proof point that AI could do more than autocomplete functions—it could architect entire applications.
What Osika and Hedin saw, though, went beyond developer tooling. They glimpsed something broader: a platform where non-technical people could build production-grade software just by describing what they wanted. No syntax. No frameworks. Just conversation.
They called it "vibe-coding," which is either brilliant branding or the sort of term that will age poorly. Time will tell.
Shipping at Breakneck Speed
Lovable went public in November 2024 with a deceptively simple pitch—chat with an AI, get a working web app. Under the hood, the product was more sophisticated than the interface suggested. Users could export code, integrate with GitHub, connect Supabase backends, and eventually lean on an "Agent Mode" that planned and executed changes with minimal human handholding.
The pace of feature releases bordered on reckless. Agent Mode arrived in June 2025, enabling autonomous plan-and-act loops. Visual editing tools, a versioning overhaul, and "Lovable Cloud" (a managed backend service) followed throughout the year. By May 2026, the company introduced "Skills"—reusable templates that let teams standardize their building patterns. A month later came automatic security scanning with optional auto-fixes at publish time.
Reading through the changelog felt less like watching careful iteration and more like watching a team race against some invisible deadline. Perhaps they sensed that the window for this kind of product—somewhere between developer tool and no-code platform—wouldn't stay open forever.
The Funding Frenzy
Lovable's cap table tells the story of a company in a hurry. An €6.8 million pre-seed landed in October 2024 from Hummingbird Ventures and byFounders. Four months later, in February 2025, Creandum led a $15 million pre-Series A when the company had 500,000 users and $17 million in ARR. Respectable numbers. Not jaw-dropping.
Then things got strange.
By July 2025—just eight months after launching publicly—Accel led a $200 million Series A at a $1.8 billion valuation. Lovable had hit unicorn status before many of its enterprise customers had even finished their procurement processes. Five months later, in December 2025, CapitalG and Menlo Ventures co-led a $330 million Series B at a $6.6 billion post-money valuation.
The investor syndicate read like a Silicon Valley census: Khosla Ventures, DST Global, NVIDIA's NVentures, Salesforce Ventures, Databricks Ventures, plus strategic arms from Deutsche Telekom, Atlassian, and HubSpot. Total announced funding through December 2025 reached roughly $552.5 million.
The revenue metrics, at least as the company reported them, suggested the valuations weren't completely unmoored from reality. According to TNW's June coverage, 80% of Lovable's builders are non-technical, and projects on the platform collectively pull in 720 million monthly visits. Whether those visits translate to sustainable enterprise revenue is a different question.
The Google Stamp of Approval
Nothing signals mainstream credibility quite like a Google Cloud partnership—or at least, that's the bet Lovable made. On June 3, 2026, Google announced an expanded multiyear collaboration that would increase Lovable's cloud footprint fivefold. The deal included deeper access to Gemini models, a listing for Lovable Agent in Google's Enterprise Agent Gallery, Wiz integration for securing AI-generated code, and procurement through Google Cloud Marketplace.
The announcement name-dropped enterprise users: HCA Healthcare, HubSpot, Microsoft, Uber, Zendesk. Lovable claimed more than half the Fortune 500 had adopted its tools "in some fashion"—a phrase carefully worded enough to encompass everything from skunkworks experiments to approved enterprise rollouts.
The company's own case studies, published in November 2025, painted a consistent picture. Delivery Hero used Lovable for rapid prototyping. Sentry built internal tools to sidestep Salesforce limitations. The common thread: teams moving faster than their traditional development pipelines allowed. Whether "faster" ultimately means "better" remains an open debate in enterprise software circles.
Stumbles Along the Way
Hypergrowth rarely unfolds without casualties, and Lovable's ascent came with its share of bruises.
In April 2026, Swedish outlet EFN reported that an access control flaw had exposed code, chat logs, and project data in public repositories for roughly two months—from early February through late April. Employees from NVIDIA, Microsoft, Spotify, and Uber were among those potentially affected. The founders issued an apology that acknowledged mistakes in both product security and internal communication.
A month earlier, Swedish tax authorities had flagged the company for unpaid VAT, an issue Skatteverket characterized as "serious." Lovable said it was registering for VAT and planned to settle outstanding amounts. Not exactly the kind of operational maturity you'd expect from a $6.6 billion company, but perhaps understandable for one growing this fast.
Then there was the phishing problem—maybe the most intractable issue of all. In August 2025, security firm Proofpoint observed tens of thousands of Lovable URLs being weaponized in malicious campaigns dating back to February. The platform's ease of use had made it attractive not just to legitimate builders, but to bad actors deploying phishing sites and scams.
Lovable responded with AI-driven protections and takedown procedures, but the incident laid bare an uncomfortable truth: tools that democratize software creation also democratize software-enabled crime. It's the sort of challenge that doesn't have an obvious solution.
The Efficiency Paradox
One hundred forty-six employees supporting $500 million in annual recurring revenue works out to roughly $3.4 million per person—exceptional by any measure, especially for a company still in hypergrowth mode.
Part of that efficiency comes from AI leverage, both in what the product does (helps people build software) and how the product itself gets built (AI helping build the AI coding tool). But the math also raises uncomfortable questions. Can a team this lean actually support enterprise customers expecting 99.9% uptime, dedicated account managers, and white-glove support?
The company's November 2025 outages—traced to upstream vendor issues at PostHog and Netlify—suggested operational maturity might be lagging commercial success. Running lean is admirable until systems start breaking.
Swedish sources reported in June 2026 that Lovable's valuation "could double" in upcoming funding discussions, though these things have a way of not materializing. Whether that happens likely depends on the company's ability to convert viral adoption into durable enterprise contracts, and to transform what started as a clever developer tool into mission-critical infrastructure.
What Comes Next
For now, Lovable exists as a peculiar artifact of this moment in technology—eighteen months from launch to a $500 million revenue run-rate, built by two Swedish founders and a team small enough to fit on a single office floor. The company has proven that AI can compress timelines in ways that would've seemed fantastical even five years ago.
What it hasn't proven yet is whether this kind of growth is sustainable. Whether the platform can evolve from a tool that helps people build things quickly to one that helps them build things well. Whether enterprise customers will trust AI-generated code for mission-critical applications, or whether Lovable will remain a prototyping platform—powerful but peripheral.
The answers to those questions will likely determine whether Lovable becomes one of Europe's defining tech success stories or a cautionary tale about moving too fast. Eighteen months in, the company has bought itself the chance to find out. That might be the real achievement.
