Anton Osika's Lovable deepens Google Cloud bet as AI coding rivals pick sides

TechCrunch reports the multi-year deal expands Lovable's Google Cloud footprint fivefold and gives it broader access to Claude and Gemini.

By ยท

Why it matters

AI coding companies are turning into cloud-alliance businesses. Lovable's Google Cloud expansion shows that model access, compute capacity, and hyperscaler distribution may matter as much as product growth in deciding who wins.

Anton Osika's Lovable deepening its Google Cloud and AI integration strategy (Museum miniature diorama with handcrafted figurines made from polymer clay and paper, set against a hand-painted backdrop)

Anton Osika's Lovable and Google announced an expanded multi-year Google Cloud collaboration on Wednesday, TechCrunch reported, with a person familiar with the deal telling the outlet that Lovable will increase its Google Cloud footprint fivefold.

Osika, identified by TechCrunch as Lovable's co-founder and CEO, is building in one of the most expensive corners of software: AI coding tools that convert prompts into working apps. Lovable, a Stockholm company TechCrunch described as fast-growing and a long-time Google Cloud customer, is now committing more deeply to the infrastructure layer that determines whether that growth can be served profitably.

What Google and Lovable actually disclosed

Lovable and Google publicly described the agreement as an expanded collaboration to scale AI-powered software creation in a Google Cloud announcement. The public announcement did not disclose a dollar value.

The more material number came from TechCrunch's source: a 5x expansion of Lovable's Google Cloud footprint, including AI usage. The same person told TechCrunch that Lovable will get expanded access to Anthropic's Claude, widely used for coding tasks, as well as Google's Gemini models.

That distinction matters. A 5x footprint increase is not the same thing as a disclosed revenue commitment. Without a minimum spend, term economics, discount schedule, or cloud-credit structure, the figure is best read as a capacity and strategic-alignment signal, not as a clean revenue number for Google Cloud.

Why Claude access is the hinge

The Claude component is the most strategically important part of the announcement. Coding-agent companies live and die by latency, reliability, context-window economics, and model quality. If Lovable can secure better or more predictable access to Claude through Google Cloud while also using Gemini, Lovable gets optionality across two model families rather than being pinned to one supplier.

Google's incentive is also bigger than ordinary cloud usage. TechCrunch reported that Google invested $10 billion in Anthropic in April in cash and compute credits, with another $30 billion possible if Anthropic hits certain performance targets, at a $350 billion valuation. That means Claude demand flowing through Google Cloud can support several Google priorities at once: cloud consumption, Anthropic distribution, and Google Cloud's pitch that developers can run serious AI workloads on its infrastructure.

For Lovable, the upside is more direct. AI coding products can look capital-efficient at the application layer while quietly depending on large and volatile inference bills underneath. A deeper hyperscaler relationship can buy Lovable more supply certainty, better model access, and potentially more favorable unit economics, though none of those terms were disclosed.

The Replit-Microsoft backdrop

The timing is notable because AI coding is becoming a hyperscaler-aligned market. Replit, another prominent AI app-building and coding environment, has been moving closer to Microsoft. That gives Microsoft a path into developer workflows that increasingly start with natural-language prompts rather than IDE setup or cloud-console configuration.

Lovable's expanded Google Cloud deal is the parallel move for Google: secure a high-usage AI coding customer before the category's winners are obvious. For Google, Lovable is not only a cloud account. Lovable is a channel into a new class of software creators whose first interaction with infrastructure may happen through an AI app builder rather than through a cloud sales motion.

For Osika, the bet is that Lovable can keep scaling usage without losing control over the model and compute stack underneath it. The unanswered commercial question is how much of that 5x expansion converts into durable gross margin for Lovable rather than simply higher pass-through infrastructure spend.

Reader comments

Conversation for this story loads after sign-in.