With the advent of DeepSeek, the AI market in the U.S. has been turned upside down.
At the same time, he began to talk about whether the chip restrictions on China were not working properly.

And everyone is surprised that a similar product comes from China, which is about 5% of the cost used for AI development in the United States.

So let’s summarize what we’re talking about in a news video interviewing Alexander Wang, founder and CEO of Scale AI

an interview summary

U.S. vs. China AI Competition
Alexander Wang says that the competition between the U.S. and China for AI is one of the most important issues today. Recently, the U.S. has maintained a lead in AI, but China is fast catching up. In particular, the recent model introduced by China’s DeepSeek Institute performed on par with the best models in the U.S.

The AI Advantage of the U.S. and the Challenge of China
The U.S. has been ahead of AI thanks to its accessibility to high-performance Nvidia GPUs (such as the A100). However, China is developing at an alarming rate, developing efficient algorithms with less resources. DeepSeek appears to have around 50,000 A100 chips, which is likely a result of bypassing U.S. export regulations.

AI Model Performance Comparison
Various models show strength in certain areas. For example, OpenAI is good at reasoning, and Anthropic is good at coding. As such, different models have differentiated strengths, and the AI market is becoming more and more competitive.

Open Source AI and Future Prospects
Companies are experimenting with open-source models, such as LLaMA, to reduce costs, while utilizing several commercial models, such as OpenAI, Claude, and Gemini. It is highly likely that the future AI ecosystem will be a form where commercial AI and open-source AI coexist.