Claude Mythos made Liang Wenfeng decide to start financing for DeepSeek.

Jin Se Finance reports that on July 1, according to The Information, the direct reason Liang Wenfeng decided to launch a financing round for DeepSeek was seeing Anthropic's Claude Mythos achieve super capabilities through massive computing power and data training, realizing that if resources were not secured, it would be difficult to sustain. The timeline broadly aligns with Anthropic's release of the Claude Mythos preview in April this year and the first news of DeepSeek's financing. The total financing amount is $7.4 billion, with Liang Wenfeng personally contributing approximately 20 billion RMB (about $3 billion), accounting for two-fifths of the total, making it the single largest investment.

After the financing was completed, DeepSeek announced that the recruitment scale for all departments across the company (currently about 300 people) would "at least double" and established an employee stock ownership plan. The company is also accelerating the adaptation of Huawei chips, but because the training and deployment system was originally built around Nvidia's CUDA, rewriting the underlying software has led to no new model releases in 15 months, missing the programming tool boom driven by Claude Code. Liang Wenfeng stated that programming tools and chatbots are both temporary products in the evolution of AI, and the company's strategy will continue to adhere to open source, low prices, and a focus on AGI, adding that "AI should not be controlled by a minority."