Google, Microsoft, and Meta Restrict Use of Competitors' AI Tools, Citing Security, Capacity, and Distillation Risks
English summary
In early 2026, major AI companies began imposing mutual restrictions. Google limited Meta's Gemini access due to compute capacity shortages, delaying internal Meta projects. Google banned most employees from using Claude Code and Codex, while Microsoft revoked internal Claude Code licenses and later restricted Claude Fable5 over data retention concerns. Meta also restricted the use of Claude and Codex in model building, fearing that model outputs could enter its training pipeline and trigger distillation or legal risks. These moves mark the end of AI tools' free-trial era, as resource scarcity, data security, and asset protection became the three key defensive barriers.
Chinese summary
2026年上半年，AI巨头开始互相设限：谷歌因算力容量不足限制Meta使用Gemini，并禁止大多数员工使用Claude Code和Codex；微软取消了内部Claude Code许可证，并因数据留存要求限制Claude Fable5；Meta则在模型构建中限制使用Claude和Codex，担忧输出数据进入训练流程引发蒸馏或法律风险。这些措施标志着AI工具自由试用期结束，算力紧缺、数据安全和模型输出保护成为三道防线。
Key points
In March 2026, Google capped Meta's Gemini usage due to compute capacity shortages, delaying Meta's internal projects.
2026年3月，谷歌因算力容量不足限制Meta对Gemini的使用，导致Meta部分内部项目延期。
In April 2026, Google barred most employees from using Claude Code and Codex for security reasons; in May, Microsoft canceled internal Claude Code licenses and pushed developers to GitHub Copilot CLI.
2026年4月，谷歌以安全为由禁止多数员工使用Claude Code和Codex；5月，微软取消Claude Code内部许可证并引导转用GitHub Copilot CLI。
In June 2026, Microsoft restricted Claude Fable5 due to Anthropic's data retention policy, and Meta limited Claude and Codex in model building to avoid distillation risks and terms-of-service violations.
2026年6月，微软因Anthropic数据留存要求限制Claude Fable5；Meta在模型构建中限制Claude和Codex，以规避蒸馏风险和违反服务条款。
Three barriers have emerged: resource limits (token/GPU shortages), data barriers (preventing sensitive internal data from entering external models), and asset barriers (protecting model outputs from being used in competitor training).
形成了三道防线：资源防线（算力与token瓶颈）、数据防线（防止内部敏感数据进入外部模型）、资产防线（防止模型输出被用于训练竞品）。
The unrestricted 'trial period' for AI tools has ended; AI is now treated as a core production input requiring quotas, audits, and boundaries.
AI工具的免费试用期已结束，AI被视作核心生产资料，需要配额化、审计化和边界化管理。