How DeepSeek Started
DeepSeek is a Chinese artificial intelligence (AI) company that has quickly become a major player in the AI world. It was founded in 2023 by Liang Wenfeng, a former stock trader and AI expert. Before starting DeepSeek, Liang co-founded a hedge fund called High-Flyer in 2016, which used AI for stock trading.
By 2019, High-Flyer relied entirely on AI-powered trading algorithms. Realizing the potential of AI beyond trading, Liang began collecting Nvidia GPUs (graphics processing units) to develop more powerful AI models. Before the U.S. restricted the sale of advanced chips to China, he had already stockpiled about 10,000 Nvidia A100 GPUs. This early preparation helped DeepSeek build powerful AI models even under international trade restrictions.
Key Milestones in DeepSeek’s Development
- July 2023: High-Flyer officially launched DeepSeek as a separate AI company, with a focus on developing large language models (LLMs) to compete with Western tech giants.
- May 2024: DeepSeek released DeepSeek-V2, an AI model that performed well and was much cheaper than models from companies like OpenAI and Google. This led to a price war among Chinese AI companies, forcing big firms like ByteDance, Tencent, Baidu, and Alibaba to lower their AI service prices.
- January 2025: DeepSeek launched its AI chatbot, which quickly became the most downloaded free app in the U.S. Apple App Store, even surpassing ChatGPT.
DeepSeek’s Technology and Cost Advantages
One of DeepSeek’s biggest achievements is its ability to develop powerful AI models at a lower cost. For example, its DeepSeek-R1 model performed almost as well as OpenAI’s GPT-4, but with a much smaller budget:
- DeepSeek-R1 training cost: $6 million
- GPT-4 training cost: Over $100 million
- DeepSeek GPUs used: About 2,000
- GPT-4 GPUs used: About 16,000
This efficiency has made DeepSeek a serious competitor in the global AI market.
Impact on the Tech Industry
DeepSeek’s rapid rise has shaken the global AI industry. After its AI chatbot became popular, major technology stocks took a hit.
- Nvidia’s stock dropped by nearly 18%, and other tech giants lost billions in value.
- By January 28, 2025, U.S. tech stocks lost nearly $1 trillion in market value due to the AI competition.
Controversies and Challenges
Despite its success, DeepSeek has faced several challenges:
- Accusations from OpenAI: OpenAI suggested that DeepSeek may have copied ChatGPT’s model using a technique called distillation (extracting knowledge from another AI model). However, no solid proof has been provided.
- Censorship Issues: DeepSeek’s chatbot follows Chinese government censorship rules and avoids certain political topics. This has raised concerns about free speech and data transparency.
How DeepSeek is Changing AI
DeepSeek’s success shows that AI development is not just about having the most money or computing power. With smart strategies and careful planning, even companies outside the U.S. can develop world-class AI. This challenges the dominance of OpenAI, Google, and Microsoft, proving that China is a serious player in AI innovation.
DeepSeek’s journey from a hedge fund experiment to a global AI leader is a fascinating example of how smart planning, resource management, and innovation can lead to rapid success. As AI continues to evolve, DeepSeek’s role in shaping the future of technology will be one to watch.
-1738753949.png)
