Microsoft CEO Satya Nadella warned that companies entirely dependent on specific AI research institutes or single models will ultimately fail to survive. In a CNN interview aired July 27 (local time), Nadella stated companies should be wary of excessively transferring core information, including data and prompts, to AI model providers. He said companies that fail to secure control over data are outsourcing their core thinking, making it difficult to sustain themselves. Nadella emphasized companies must directly possess and manage all metadata generated when operating AI systems without transferring it to external platforms, as securing data sovereignty is essential to build proprietary customized models. He advised against entirely relying on built-in development tools from specific AI companies, such as Anthropic’s Claude Code or OpenAI’s ChatGPT Codex, explaining that separating development tools and context from AI models allows companies to use multiple AIs according to individual strengths. Nadella concluded that by securing data sovereignty, companies can maintain control over their own destiny even if a specific model disappears.