Google and Microsoft restrict competitor AI coding tools

Google and Microsoft have begun barring employees from using rival AI coding assistants. Google stopped most staff from accessing Anthropic’s Claude Code and OpenAI’s Codex, citing security concerns that internal source code, meeting notes or other confidential material could be inadvertently fed into external models. Microsoft similarly revoked Claude Code licences for many workers, steering development toward its own GitHub Copilot platform. Both moves aim to safeguard proprietary data, control model output usage, and protect against “model distillation” by competitors, while also reflecting growing competition for limited compute resources.

In parallel, analysts note that AI‑powered programming tools are reshaping the economics of software rewrites. By automating code generation, refactoring suggestions, documentation, and test assistance, AI lowers some barriers to updating legacy systems. However, the underlying costs of understanding existing code, ensuring business‑logic compatibility and managing deployment risk remain, so AI is more likely to shift the decision threshold for rewrites than to eliminate the associated risks entirely.