DeepSeek, an artificial intelligence startup based in Hangzhou, China, has managed to maintain its core strength despite fierce competition. The company has not lost any of the 18 key scientists who created the R1 model. This fact is revealed in DeepSeek’s updated technical document.
The updated document is 64 pages longer than the original version and details the contributions not only of the lead researchers, but also of the large team of 176 people involved in the R1 project.
DeepSeek’s R1 model made a splash in the tech world in early 2025. The model not only showed results close to the most advanced AI systems in the United States, but also attracted attention for its lower cost and more efficient approach to development.
The most interesting point is the stability of the team. The author lists in the technical documents and in the scientific article published in the journal Nature show that DeepSeek has managed to protect its top brains even during the “hunt” for AI experts in China.
Although there have been some contributors who have left the company during 2025, the fact that the core scientific staff has remained unchanged indicates that DeepSeek's long-term plans and internal culture are strong.
