By ByteBot
Stockholm-based Lovable raised $330 million at a $6.6 billion valuation on December 18, backed by Nvidia, Google’s CapitalG, and Salesforce. The vibe-coding startup—which turns natural language into complete applications—tripled its valuation in five months, jumping from $1.8 billion in July. The company hit $200 million ARR within 12 months and generates 100,000 new projects daily. Provocative timing: this massive raise lands days after reports showed developers are slower and more skeptical of AI coding tools the more they use them.
What is Vibe-Coding and Why It’s Different
Vibe-coding—Andrej Karpathy’s term, Collins Dictionary’s 2025 Word of the Year—isn’t GitHub Copilot’s autocomplete. You describe an application in plain English. AI generates the React frontend, Supabase backend, authentication, and deployment. No code required.
A journalist told Lovable to “build a tool that tracks global south tech news.” Minutes later: a Bloomberg terminal-style interface, deployed. At Zendesk, teams compressed six-week development cycles into three hours. McKinsey engineers delivered solutions in hours that previously required six months.
Lovable’s CEO Anton Osika wants the company to be “the last piece of software companies ever buy.” The vision: instead of purchasing separate CRM and project management tools, organizations build exactly what they need on demand. Ambitious. Possibly delusional. $6.6 billion expensive.
Here’s what separates Lovable from the pack: 63% of users are non-developers. GitHub Copilot and Cursor target engineers. Lovable targets everyone else.
The Metrics Behind the $6.6B Valuation
Lovable hit $100 million ARR in eight months, doubled to $200 million four months later. The platform generated 25 million projects in year one. Enterprise customers include Uber, Klarna, and Zendesk.
The valuation math is aggressive. $6.6 billion on $200 million ARR is a 33x multiple. Cursor raised at $9.9 billion on $500 million ARR—a 20x multiple. Lovable’s multiple is higher despite smaller revenue, pricing in that non-developer market premium.
The Counternarrative: Developer Skepticism
Yesterday, ByteIota published research showing developers using AI coding tools are 19% slower in controlled studies. Stack Overflow’s 2025 survey: only 29% trust AI tool accuracy, and trust declines with experience.
Vibe-coding has documented problems. May 2025: security researchers found 170 of 1,645 Lovable apps had vulnerabilities allowing unauthorized data access. The New Stack reported senior engineers facing “development hell” with AI-generated codebases. Sixty-three percent of developers spend more time debugging AI code than writing equivalent code manually.
AI startup bubble warnings are loud. Sam Altman called it “insane” that startups with “three people and an idea” get astronomical valuations. Google DeepMind’s CEO: “There’s obviously a bubble in the private market.” MIT reported 95% of enterprises get zero return on $30-40 billion in generative AI investment.
Why did Nvidia, Google, and Salesforce just hand Lovable $330 million?
Why VCs Bet Billions Despite the Data
Lovable isn’t competing with Copilot. Professional developers are skeptical because they maintain complex systems where AI-generated code creates technical debt. Lovable’s users are 63% non-developers building prototypes and simple internal tools—use cases where speed trumps maintainability.
Zendesk’s six-week-to-three-hour story isn’t about production code. It’s about validating ideas fast. McKinsey’s timeline isn’t replacing engineers. It’s bypassing them for throwaway internal dashboards. That’s a legitimate use case.
Nvidia’s investment signals validation—they don’t back everything. Enterprise customers (Uber, Klarna, Zendesk) show commercial traction. $200 million ARR in 12 months proves people are paying, not just trying.
But the valuation assumes Lovable becomes “the last piece of software companies ever buy.” That’s where skepticism is warranted. Custom software has always been possible, just expensive. Lovable makes it cheaper and faster, but “last software ever” sounds like hyperbole that precedes a correction.
The Verdict: Bubble or Breakthrough?
The truth is in the middle. Vibe-coding works for prototypes, simple internal tools, and applications where non-developers need something functional fast. It doesn’t work—yet—for complex, maintainable systems where security and technical debt matter.
Lovable’s $6.6 billion valuation assumes the first category is enormous and the second doesn’t matter. With Nvidia, Google, and Salesforce’s money behind it, that’s a bet worth watching.
Developer skepticism isn’t wrong. Productivity studies, security vulnerabilities, and technical debt warnings are real. But VCs see a market of non-developers who want software without hiring engineers, and Lovable proved that market will pay.
2026 will answer the question. Either Lovable’s revenue scales to justify the 33x multiple, or it joins AI startups that raised at bubble valuations and crashed when reality arrived. Watch the ARR growth and security improvements. If those stall, the correction will be brutal.
