According to CoinMeta, as the competition between OpenAI and Anthropic over data privacy issues intensifies, more and more companies are beginning to pay attention to open models and "sovereign AI." When choosing AI suppliers, data privacy has become a key consideration, with many executives worried that using these companies' AI models could lead to the leakage of their "secret formulas" to competitors. In June, Anthropic began to enforce a 30-day chat record retention policy, which sparked strong opposition from the industry. Microsoft CEO CEO Nadella and Palantir CEO Satya Nadella also warned companies not to rely on the models of these cutting-edge AI companies, stating that they might steal intellectual property. Although OpenAI and Anthropic claim not to use corporate data for model training, many data security experts believe that these companies may obtain operational insights into their customers in more subtle ways. To address these concerns, companies are seeking to diversify their AI usage and reduce their dependence on leading AI companies. The concept of sovereign AI is gradually gaining traction in the corporate world, referring to the ability of companies to control their own AI stacks, including owning the chips that run AI workloads and establishing proprietary cloud infrastructure.