The rise of DeepSeek: new opportunities in the AI ​​agency track

Author: Daoshuo Blockchain

The emergence of DeepSeek is definitely a "black swan" in 2025.

This black swan directly "blacked" the US stock market and broke the mindset of the subsequent development of artificial intelligence.

Previously, Google's former CEO once said that the competition in artificial intelligence is a competition with a threshold of 100 billion US dollars. If you can't cross this threshold in terms of funds, you can't even get a ticket to the competition.

Zhu Xiaohu of Jinshajiang Venture Capital once said that he would not invest in domestic large-model companies at all, because he believed that domestic companies would never catch up with the United States in terms of large models.

The reason why these two people made such an assertion is that there has always been a basic assumption in the AI ​​competition: to win in the competition, especially in the competition of underlying models, computing power and funds are a threshold that cannot be crossed.

Our country's enterprise cards are stuck in computing power - they can't get the best Nvidia graphics cards.

The most direct shock brought to the world by the advent of DeepSeek is that it has achieved extremely excellent performance at a very low cost.

This achievement directly cast the biggest "shock bomb" on the current global competition in artificial intelligence:

If performance improvements and cost reductions can be achieved through algorithm improvement and optimization, then does it still need such huge investments in computing power and energy to continue to promote and develop artificial intelligence applications?

So the founder of "Black Wukong" praised DeepSeek as a national achievement.

However, no matter how excellent DeepSeek is, it is still the result of standing on the shoulders of giants. As Yang Lekun said:

This is the victory of open source big models over closed source big models.

I think DeepSeek has three very far-reaching effects on the subsequent development of artificial intelligence:

First: It greatly reduces the cost of big models, which means that countless small and medium-sized companies may join the competition for AI big models and develop dedicated big models in various segments and industries.

Second: China's artificial intelligence entrepreneurial wave will explode---far beyond what we are seeing now.

Third: China's entrepreneurial teams in the artificial intelligence track will be looked at with new eyes and are likely to re-attract the attention of the global venture capital community.

The success of DeepSeek means that the threshold of computing power is not as high as imagined. Just after DeepSeek went online, many laboratories, including universities in Hong Kong, successfully reproduced DeepSeek's results.

We can fully imagine that a large number of companies will further optimize the model based on DeepSeek's open source, so that we can train good large models with less top-tier chips.

The first of these impacts is now interfering with US stocks.

Because of DeepSeek's amazing results in cost and performance, the core companies in the artificial intelligence track will now face revaluation.

Nvidia fell more than 10% in Monday's opening. Other related companies are facing similar situations.

I immediately thought of Tesla.

I have always believed that Tesla's visual solution is likely to be the final solution for autonomous driving. Tesla's advantages in this solution are large models and data.

Now with the advent of DeepSeek, Tesla's advantage in large models may not be so obvious, and the data advantage is still far ahead.

How long this leading advantage can be maintained depends on the subsequent development of other electric vehicle companies.

Back to the crypto ecosystem.

On the surface, the advent of DeepSeek has not only pulled down the US stock market, but also the crypto ecosystem, especially the AI ​​proxy track. Many people think that the AI ​​proxy track has collapsed. But I think this just shows from another perspective that the competition in the AI ​​agent track is probably far from over.

Because if the big model can iterate at such a fast speed, it means that the leading advantages achieved by the existing AI agents are only temporary. If they cannot keep up with the trend and the iteration of technology in the future, they may be overturned by the up-and-coming companies with the latest technology and the latest models at any time.

At the same time, this also means that the research and development threshold of AI agents is further lowered, and more participants can participate in the development of AI agents. It also means that Chinese local teams have every opportunity to show their talents in this arena.

Today is our last article in the Year of the Dragon. We will meet you again after the Spring Festival of the Year of the Snake. I hope that in the new year we can continue to maintain our enthusiasm, continue to find what we love in this industry, and reap what we gain.