The founder of DeepSeek has now become a billionaire.

As a student, he became fascinated with the idea of automating stock trading, recognizing AI’s potential in finance.

‍

In 2013, he founded Hangzhou Jacobi, his first investment firm. Two years later, he co-founded High-Flyer with two university friends—an AI-driven hedge fund designed to predict market movements. The company invested in Nvidia chips to train predictive models, achieving an impressive 71% return in 2020. However, 2021 proved challenging due to AI model errors, but High-Flyer rebounded in 2022 with a 15% gain.

‍

In 2023, Liang took his ambitions further by launching DeepSeek, funding it with his own earnings. He assembled a team of young graduates from China’s top universities to develop cutting-edge AI technologies. On January 20, 2024, DeepSeek released its R1 model, which quickly outperformed ChatGPT and Claude in U.S. app stores. Within two weeks, it had over 3.6 million downloads, turning Liang into a billionaire.

‍

DeepSeek operates on a simple yet effective business model, offering developers access to its AI models at a highly competitive rate—just $2.19 per million tokens compared to OpenAI’s $60. According to Forbes, the company generates approximately $6 million for every million paying users and could already be worth over $1 billion.

‍

Unlike OpenAI and xAI, DeepSeek has no external investors. Liang owns 84% of the company and has chosen to self-fund, ensuring the freedom to operate without outside pressure. Additionally, he has made DeepSeek’s models open source, allowing anyone to access them for free.

‍