- By Prateek Levi
- Tue, 28 Jan 2025 03:17 PM (IST)
- Source:JND
Liang Wenfeng DeepSeek: Liang Wenfeng, a hedge fund billionaire, is the driving force behind DeepSeek, an AI startup that has caught the world’s attention, shaking up the dominance of Silicon Valley and raising concerns across the globe. Founded just a year ago, DeepSeek has rapidly developed models that rival the most advanced AI systems from tech giants like OpenAI, all while spending a fraction of what their U.S. counterparts invest.
What makes Wenfeng’s story even more fascinating is his background. Unlike many tech entrepreneurs who come from engineering or computer science fields, Wenfeng has a financial background. A graduate of Zhejiang University, he co-founded a quantitative hedge fund, High-Flyer, in 2015. His deep interest in AI led him to explore machine learning to predict market trends, which ultimately allowed him to dive into AI research without the pressure of seeking external investments.
His journey into AI took a pivotal turn in 2021 when he started hoarding Nvidia chips. Although this move seemed eccentric at the time, it proved to be foresight. By 2023, Wenfeng launched DeepSeek, initially as a side project, with the goal of challenging China’s biggest tech firms as well as the leading U.S. companies in AI. He built a team by recruiting top talent from China’s leading universities, offering competitive salaries that rivaled those of the biggest names in tech, and focusing on homegrown AI innovation.
ALSO READ: Janus Pro 7B: DeepSeek’s New AI Model Outperforms DALL-E 3, Claims Company
DeepSeek made waves in November 2023 with its first model, DeepSeek Coder, an open-source AI tailored for coding tasks. By 2024, the company had launched the DeepSeek LLM, a 67-billion parameter model that put it on the map as a serious competitor in the AI space. But it wasn’t until 2025 that DeepSeek truly began challenging the global AI scene. The DeepSeek-R1 model, released in January 2025, wowed the tech world with its advanced reasoning capabilities, rivaling OpenAI’s GPT-4, all while operating on a budget that was shockingly low. The $5.6 million Wenfeng spent on computing resources for R1 was a mere fraction of the hundreds of millions that U.S. companies typically pour into similar models.
Despite the skepticism that surrounded his early ventures, Wenfeng’s vision of building a cost-effective, world-class AI product has turned DeepSeek into a formidable player. Unlike the massive investments made by U.S. giants, DeepSeek has relied on innovative techniques like reinforcement learning and model distillation, which make their AI systems not only more efficient but also cheaper to operate. By using less powerful chips, such as Nvidia’s H800s, DeepSeek has been able to deliver impressive results at a fraction of the cost.
DeepSeek’s models, which are open-source, also present a huge advantage for smaller businesses. While companies like OpenAI and Google keep their models closed and charge steep fees for API access, DeepSeek’s low-cost and open approach makes advanced AI technology more accessible to a wider range of companies, particularly startups or businesses that lack the resources to build their own AI solutions. The company’s competitive pricing strategy has already triggered price cuts across China’s AI market, with major companies like ByteDance, Baidu, and Tencent adjusting their prices to stay competitive.
However, there are also geopolitical concerns surrounding DeepSeek’s rise. The company’s timing is notable, coming at a moment when U.S.-China tensions over AI and technology are at an all-time high. While some critics worry about the potential for censorship in DeepSeek’s models, especially given its ties to China’s Communist Party, the company’s open-source approach could help foster trust and accountability in the AI community.
Yet, while DeepSeek is proving itself as a powerful force in the AI space, questions remain about where India stands in this race. Without a major AI lab or startup on par with OpenAI or DeepSeek, India faces a tough challenge in keeping up with this rapidly evolving field.
In short, Liang Wenfeng and DeepSeek are changing the global AI landscape, offering a cost-effective, powerful alternative to the expensive models developed by U.S. companies. As they continue to grow, the implications for the future of AI, both technologically and geopolitically, are only just beginning to unfold.