WE THINK：The Situation is Developing in Favor of China, with DEEPSEEK Leading the Breakthrough Beyond the

In January, due to the Lunar New Year holiday falling on January 28th this year and the inauguration of the new U.S. president on January 20th, both domestic policies and capital markets were characterized by a cautious wait-and-see attitude. On one hand, there was a focus on summarizing the past year and planning for the new year; on the other hand, there was interest in observing the reactions to the new policies of Donald Trump, who was returning to the presidency. During this month, the Wind All-A Index and the CSI 300 Index fell by -2.32% and -2.99%, respectively, leading to a general decline in investor sentiment. However, overseas sentiment towards China warmed up during the Spring Festival.

In this edition of WE THINK, we will explore several recent hot topics: Trump's new policies, the breakout success of DEEPSEEK, gold prices hitting record highs, and signs of weakness in the U.S. stock market. We will delve into these subjects for further discussion.

After the Trump administration announced a 25% tariff on Canada and Mexico, the capital markets cast a vote of disapproval!

On February 1st, local time, the Trump administration, citing a series of issues and reasons, announced a 25% tariff on imported goods from Canada and Mexico. The Canadian government swiftly responded with strong retaliatory measures. On February 2nd, Canada announced a 25% tariff on $155 billion worth of goods exported from the U.S. to Canada... The U.S. capital markets reacted negatively almost immediately, with stock prices of companies affected by the tariffs plummeting sharply, triggering a broad market decline. When the U.S. stock market opened on February 3rd, major indices experienced significant drops. Additionally, Trump's threat to impose tariffs on the EU prompted a similarly strong response from the European Union. The Wall Street Journal published an article criticizing Trump's tariff war against Canada and Mexico as "the dumbest trade war in history." Later on February 3rd, the U.S. announced a one-month delay in implementing the tariffs on Mexico, pushing the effective date to March 4th. Following this news, U.S. stocks quickly rebounded, narrowing their losses significantly. By the end of the day, the S&P 500 and Nasdaq indices had fallen by only 0.76% and 1.2%, respectively. After the market closed, news emerged that the U.S. and Canada had also agreed to a one-month delay in tariffs, with both sides continuing to communicate and negotiate.

Looking ahead, it remains highly uncertain whether the U.S. will follow through on Trump's initial intentions to impose tariffs on Canada, Mexico, and the EU. If the tariffs are indeed implemented, strong countermeasures from the affected countries will be inevitable, given their national interests and dignity. This would inevitably impact many U.S. companies' markets and performance, further unsettling the capital markets. Since the stock market is an area Trump personally values highly, a bearish market would be something he absolutely cannot tolerate or ignore. As circumstances often outweigh individual will, U.S. policies may not proceed as easily or freely as the rhetoric during the campaign suggested. Faced with choosing the lesser of two evils, careful consideration of the pros and cons will be necessary.

Overall, if the U.S. government makes misguided moves, it could marginally improve the external competitive environment for Chinese companies. At the same time, given the widespread uncertainty surrounding tariffs and domestic policy instability following political changes in various countries, China's political stability has become a rare advantage on the global stage. Manufacturing investments outside China may face renewed hesitation and scrutiny, as tariff risks are no longer unique to China but have become a widespread concern.

With DEEPSEEK successfully breaking through the barriers, the technological blockade holds strategic significance.

During the Lunar New Year holiday, the most talked-about and sensational topic was undoubtedly DEEPSEEK! Not only did it dominate discussions across domestic official media, new media, and self-media platforms, but it also became the hottest topic in the global tech and media circles—a phenomenon that is extremely rare in my over 30 years of experience in the industry!

Since the official release of the DEEPSEEK-R1 model on January 20th, it topped the download charts on both Apple's App Store and Google Play Store globally by January 26th. Within 18 days of its launch, cumulative downloads surpassed 16 million, maintaining a leading position across 140 markets (surpassing ChatGPT's first-month downloads of 9 million but still far behind its current 400 million monthly active users). As of January 31st, DEEPSEEK's global MAU (monthly active users) reached 45 million, with a DAU (daily active users) of 22.15 million. In comparison, ChatGPT's global MAU stands at 350 million, with a DAU of 53 million. Notably, DEEPSEEK's DAU has already exceeded Doubao's 17 million DAU, showing rapid growth. It took ChatGPT 244 days to reach a DAU of 15 million.

In various evaluations comparing DEEPSEEK with other major global AI models,opinions vary. However, the general consensus is that DEEPSEEK has surpassed OpenAI's o1 level and is slightly below OpenAI's o3 level, but outperforms in Chinese language environments. This places DEEPSEEK among the global top-tier leaders in AI. Tech giants such as NVIDIA, Microsoft, Meta, and OpenAI have all highly recognized DEEPSEEK's innovations in algorithm technology.

On January 27th, after the U.S. market opened, tech stocks, led by NVIDIA, experienced a sharp decline. NVIDIA's stock price dropped by 16.97% that day, wiping out $570 billion in market value in a single day. Tech stocks, often seen as the most resilient fortress in global capital markets, were unexpectedly shaken by the DEEPSEEK whirlwind, causing some panic. Although Microsoft, a shadow stock of OpenAI, saw a narrower decline that day, its stock price gradually stepped down in the following trading sessions, resulting in significant market value losses.

Why would a relatively unknown Chinese startup cause such a stir among U.S. tech giants and prompt global investors to sell off stocks of dominant players like NVIDIA and Microsoft?

For AI large models to achieve success, the industry generally agrees on three critical factors: data, computing power, and algorithms. Previously, the U.S. aimed to maintain its dominance by restricting China's access to advanced computing power, believing that its overwhelming lead in computing scale and capabilities would leave China behind in the AI-driven era. However, the DEEPSEEK team has now introduced groundbreaking innovations in algorithms, achieving remarkable results with significantly less computing power (using just over 3,000 GPUs compared to clusters of 10,000) and lower training costs (55 million versus an average of 100 million, roughly one-twentieth of the cost). These advancements have enabled DEEPSEEK to match or even surpass the AI capabilities of U.S. tech giants, particularly in areas requiring deep reasoning. This breakthrough has shattered the illusions of those who believed in monopolizing AI technology through sheer computing power.

Since DEEPSEEK is open-source and has published related research papers, the U.S. tech community has been able to study and verify the effectiveness—and even the leading edge of China's breakthrough.

This moment in AI history echoes similar patterns seen in other technological fields. For instance, the original inventions of photovoltaic cells and lithium batteries did not come from Chinese scientists. However, it was the collective efforts of Chinese scientists and engineers that enabled large-scale industrial production, high quality, low costs, and economic viability. Historically, the U.S. scientific community has often played the role of the original inventor, while China's scientific and industrial sectors have excelled in transitioning innovations from the lab to mass production. This division of labor has proven effective, provided that the distribution of benefits between original intellectual property and industrial application is handled fairly. If managed well, this dynamic could be a boon for both China, the U.S., and humanity as a whole.

The same logic applies to the productization and application of AI. If astronomical levels of computing power and capital expenditure are required, the high costs will ultimately be passed on to consumers, hindering widespread adoption. China's advancements in AI algorithms have made a significant contribution to the industry's development and progress, paving the way for more accessible and scalable AI solutions.

At the beginning of new year, our team reached a consensus that 2025 will be the year of AI productization. At that time, there were some lingering concerns, especially regarding whether the "brain" of our AI terminal products would be competitive. We were confident in the hardware technology and cost, but we worried that if our AI products' intelligence lagged too far behind that of GPT, consumer experience would be significantly lacking. The emergence of DEEPSEEK has greatly alleviated these concerns! Confidence has increased that Chinese companies will not miss this wave of AI technological advancement. As one of our tech team members put it, "China's AI has suddenly brightened!"

The price of gold has reached a historical high!

Since the beginning of 2025, international gold prices have continued to rise gradually. Following a 27.23% increase in 2024, they have risen by another 7.46% this year. On the morning of February 4, the London spot gold price reached a historical high of $2,820 per ounce. Amid severe dollar overissuance and the lack of constraints on U.S. debt, along with an uncertain international environment, the status of gold as a global currency reserve is increasingly being emphasized and recognized, making precious metal assets worthy of attention.

Recent Strategy

Recently, the U.S. stock market has shown clear signs of high-level fluctuations. I have always believed that U.S. technology and Chinese manufacturing are the two most competitive fields globally today, and they are also the most noteworthy areas for investment! In recent years, due to various factors, global investors' enthusiasm for U.S. technology and the continuous profit-making effect have led many to develop an "investment faith" in leading U.S. tech companies, resulting in substantial benefits from that belief. Similarly, this has created a certain "disdain chain" towards Chinese assets, with an overweight in U.S. stocks and an underweight in Chinese assets becoming a widely accepted fact and belief among many.

Years of experience in the capital markets tell me that when most people adhere to a particular investment strategy as a "faith," it may indicate that the strategy has reached an extreme! Recently, in conversations with some friends, I have a feeling that by the end of 2024, U.S. tech stocks might resemble the core assets in China at the beginning of 2021. The companies may still be good, but their returns could significantly lag in the following years, possibly due to valuation issues.

We do not provide any advice on shorting U.S. stocks. However, we believe that going long on quality Chinese assets in 2025 could be a good option!

WU Weizhi

4 Feb 2025

本期《偉志思考》簡體中文版鏈接：

伟志思考：情况朝着有利于中国的方向发展，DeepSeek率先突破小院高墙