微软首席执行官萨蒂亚·纳德拉在CNN节目中发出警告，指出完全依赖单一AI提供商的企业可能无法长期存活[1]。纳德拉表示，"任何没有这种控制权的公司，我敢说将不复为一个公司，因为你本质上已经把自己的思维外包出去了"[1]。他强调企业需要保有数据的自主权和多模型选择能力，以维护独立的竞争地位。
为了避免过度依赖AI实验室，纳德拉建议企业采取关键举措[1]。首先，企业应保留围绕数据的所有元数据，用于训练自有权重或开源模型[1]。其次，企业应避免过度依赖AI实验室内置的编码工具，如Claude Code和ChatGPT Codex等[1]。第三，企业可以通过将接口层与AI模型分离、将上下文和记忆机制从模型中独立出来的方式，灵活使用多个AI模型[1]。纳德拉指出，"通过让接口层独立于模型、让上下文和记忆独立于模型，你完全可以使用多个模型"[1]。
值得注意的是，微软本身是Anthropic和OpenAI的投资者[1]。纳德拉还强调，消费者与企业的情况有所不同，对消费者而言共享数据是使用服务所需付出的代价[1]。
Microsoft Chief Executive Satya Nadella has cautioned businesses against placing complete reliance on a single artificial intelligence model provider, arguing that such dependency could threaten their survival [1]. Speaking on CNN, Nadella stated that "any firm that doesn't have this control, I will claim will not remain a firm because you've essentially outsourced your thinking" [1].
To maintain independence and competitive advantage, Nadella recommended that enterprises retain ownership of their usage data and metadata to train proprietary models or leverage open-source alternatives [1]. He emphasized the importance of separating AI harnesses—built-in coding tools offered by providers such as Claude Code and ChatGPT Codex—from the underlying model itself [1]. "By keeping the harness separate from the model and the context and memory separate from the model, you absolutely can use multiple models," Nadella explained [1]. This architectural approach would enable organizations to adopt multiple AI providers simultaneously without vendor lock-in [1]. Nadella distinguished between enterprise and consumer usage, noting that while consumers may share data as the price of service access, businesses should maintain strategic control over their AI infrastructure [1].