Lovable expands Google Cloud deal to scale AI software creation worldwide

Lovable has deepened its dependence on Google Cloud in a multi-year expansion that gives the AI software maker a fivefold larger footprint on the platform and broader access to Anthropic’s Claude. The agreement, announced at Google Cloud Summit Nordics ’26 in Stockholm, puts Lovable more firmly inside the orbit of the same infrastructure giants that are rapidly shaping the next phase of AI development.

Google Cloud said the expanded collaboration makes it one of Lovable’s primary technology partners and is intended to help the company scale its AI-powered software-creation platform to millions of users globally. The cloud provider also said Lovable users are now processing more than one million new projects every week, a scale that turns the startup’s appetite for compute into a strategic issue, not just a technical one.

The deal builds on a relationship that was already tightly intertwined. Lovable launched Lovable Cloud and Lovable AI on September 29, 2025, and said AI usage in published apps was free for a limited period in late September and early October 2025, powered by Google Gemini models. On February 5, 2026, Lovable added Claude Opus 4.6 as a core model, further tying its product stack to outside model providers as it pushed toward broader adoption.

That growth has been fast enough to make Lovable one of Europe’s most closely watched AI startups. By March 2026, the company said it had about 8 million users and 146 full-time employees, and it raised $330 million in December 2025 at a $6.6 billion valuation. Those numbers help explain why access to large-scale cloud capacity now sits at the center of its strategy: the company is not just buying software support, it is building its business on top of a narrow set of platforms that can keep pace with demand.

The broader picture is even clearer in the backdrop to the Lovable deal. Google and Anthropic have already disclosed a much larger cloud expansion that gives Anthropic access to up to one million Google TPUs and well over a gigawatt of capacity coming online in 2026. Together, those arrangements show how AI infrastructure has become a site of power concentration, where startups, model makers and cloud providers are locked into a small group of dominant suppliers that control the compute needed to grow.