In recent years, the most significant trend in the tech sector has undoubtedly been artificial intelligence (AI).

Major tech behemoths are pouring substantial investments into AI. In the United States, we have OpenAI, Microsoft, Meta, Google, Apple, and xAI, while in China, Alibaba, Tencent, Baidu, ByteDance, and Xiaomi are all vying to stay ahead in the race, fearing obsolescence.

Of course, from the vantage point of AI-centric companies, few, if any, have yet to turn a profit from AI. Be it OpenAI, Google, or Meta, they are all in the phase of heavy investment without immediate returns.

Nevertheless, the AI industry as a whole has still managed to contribute to profits.

Projections indicate that by 2026, the entire AI industry chain will generate a staggering $637 billion in profits. These profits, however, are not emanating from the AI companies themselves but rather from the upstream sectors of the AI industry.

This raises a pertinent question: Who is reaping these profits, and who is reaping more?

First, let's examine a chart. This chart provides a clear overview of the situation.

From a country and regional standpoint, the United States emerges as the top earner, capturing 49% of the total profits, which equates to $314 billion. South Korea follows closely behind, securing 35% of the profits, or $223 billion.

It can be argued that the United States and South Korea are the true champions, collectively amassing 84% of the total AI profits.

Next in line is Taiwan, China, which earns $47 billion, accounting for 7.4%. Mainland China, on the other hand, earns 4.1%, or $26 billion. Japan and Europe trail behind, with even smaller earnings: Japan secures $14.2 billion, representing only 2.2%, and Europe earns $13.7 billion, also around 2.2%.

What accounts for the United States' substantial earnings, and why does South Korea rank second?

Let's delve into the earnings of specific companies to gain a better understanding.

NVIDIA stands out as the top earner, raking in $207 billion alone. Given that all AI systems necessitate AI chips, NVIDIA, which supplies AI chips to AI companies, commands a 90% market share and, consequently, reaps substantial profits.

Then we have SK Hynix, Samsung, and Micron. All AI systems require memory, and these three companies collectively monopolize around 85% of the global memory chip market, thus also enjoying significant profits. Together, these three companies earn even more than NVIDIA.

Next is TSMC, which manufactures chips for NVIDIA and others, earning $33 billion on its own. There are also other supply chain companies that provide ancillary products for NVIDIA, Samsung, and the like, such as Broadcom, Kioxia, Foxconn Industrial Internet, InnoLight, and Yitai Technology.

Overall, in the entire AI industry chain, AI chip and memory manufacturers emerge as the big winners. These two types of companies are the ones raking in the most money, followed by foundries that secure a smaller share, and finally, the supporting companies that scrape by with the leftovers but still benefit to some extent.

China (primarily referring to Mainland China, excluding Taiwan Province) actually holds a relatively modest share in this industry chain. This is attributed to our limited participation in the main AI chips and memory chips, with only some involvement in the supporting areas. However, the value of these supporting areas is not the highest, so we still need to redouble our efforts and strive for higher-value segments to earn more money.