Google Now Lets Its Own Engineers Use a Rival AI to Write Code

Google has quietly opened the door to a competitor's AI. The company now lets many of its own software engineers use Claude, the coding-focused AI made by Anthropic, to help write and fix computer code. That is a notable reversal: Google used to restrict staff from using outside coding assistants like Claude Code or OpenAI's Codex, keeping those tools to a few elite teams. Gemini, Google's homegrown AI, remains the official default.

So what? AI coding assistants (AI that writes computer code for you) are becoming the assembly line of the software industry. When two fierce rivals push each other on the same task, the tools get better and cheaper fast. That trickles down to you: the apps and websites you use get built quicker, cost less to maintain, and ship fixes sooner. It also signals something honest and slightly uncomfortable — even the company that built Gemini reportedly heard complaints that it stumbled on hard programming problems.

The catch is control. Claude is not handed out freely. It sits inside Google's own developer platform, called Antigravity, and each engineer gets a limited monthly allowance, or quota. Once they hit it, they are back to Gemini. So Google gets the benefit of a stronger tool for tricky jobs while keeping its own AI as the default — the supplemental tool never becomes the main one.

The bigger twist: Google is both a competitor to Anthropic and a major investor in it, with plans to put in as much as $40 billion. Amazon does the same thing, letting staff use outside AI while building its own. For you, the practical takeaway is that AI tools are blending together. Expect fewer lock-ins, more choice, and more of your work quietly done by whichever AI happens to be best that day.