Mark Cuban: Lovable and Replit have the potential to outlast AI labs.

      At the RAISE Summit in Paris on Wednesday, Mark Cuban, investor from “Shark Tank,” advocated for AI coding tools such as Lovable and Replit, as reported by Business Insider. He claimed that these tools can compete effectively with major AI labs because they now offer bundled services that a standard language model lacks. Since Cuban is an investor in Lovable, he has a vested interest in this perspective. However, his argument directly addresses the main concern surrounding “vibe-coding” startups: if companies like Anthropic or OpenAI can easily introduce new features, what protection do these startups have? Cuban expressed this concern during a discussion with Lovable CEO Anton Osika, stating, “When we evaluate many front-end tools like Lovable and Replit, we wonder if foundational models from the big players will simply replace them. But what you're indicating is that you have a localized base of data.”

      Transitioning from a coding tool to an “AI cofounder,” Osika suggests that Lovable has evolved beyond being merely an AI software engineer. Now, founders can set up their companies and manage payments within the platform, rather than just generate apps. “People previously viewed us as a product that acts like an AI software engineer,” Osika noted. “Now, we’re increasingly recognized as a partner—essentially their AI cofounder.”

      This distinction is significant from a commercial standpoint. A code generator competes solely on the quality of the model, while a comprehensive business platform is not as easily replaceable. Lovable has capitalized on this broad scope, reportedly achieving $500 million in annual revenue with a small team.

      The recent months have intensified concerns. Following the release of Anthropic’s Opus 4.6 model earlier this year, many founders and developers announced on X that they had canceled their subscriptions to Cursor and Lovable in favor of Claude Code. This trend raises alarms for investors, who continue to compare vibe-coding applications against base models. They fear that a single update to Claude Code could potentially eliminate a startup's competitive advantage overnight. In response, some tools are developing their own models to reduce reliance on the major labs.

      The implications are clear. Lovable explicitly acknowledges this threat. In a podcast from March, the company's head of growth, Elena Verna, stated that competing startups do not concern her the most. “I always worry about the big players in the industry,” she remarked, mentioning OpenAI, Anthropic, Google, and Apple. Her reasoning centers on distribution; these giants can reach billions of users and easily integrate coding assistants into everyday applications. Cuban’s rebuttal is straightforward: by owning the entire workflow surrounding code, including incorporation, payments, and data, Lovable creates a barrier that the labs cannot easily breach.

      The critical question remains whether this barrier will hold as models advance. A significant portion of European startup value currently hinges on this answer.