Recently, Li Kaifu, CEO of Zero One and Chairman of Innovation Works, was a guest on Gelonghui's high-end interview program 'Gelong Doctor Room'. During the conversation, Li expressed bullish views on China's excellent companies, believing that their market cap will also have the potential to reach trillions of dollars. After 40 years of reform and opening up, China's GDP has gone from being ranked low to the second largest economy in the world, but unfortunately, China has not produced any huge listed companies like Apple, Microsoft, or Google. Li Kaifu analyzed that the characteristics of these great companies lie in their courage to do things that nobody had done before and their internationalization capability. These two points are not the strong points of Chinese companies, which have strong execution, produce good products, iterate continuously, and have good business sense. However, this type of thinking, where you believe what you see, is not the standard way of thinking in China. Nevertheless, China still has many powerful companies, from Contemporary Amperex Technology to Huawei, Alibaba, Tencent, and so on. They all have the opportunity to become trillion-dollar market cap companies. Li frankly stated that these companies have the opportunity to reach trillions of dollars, but there are also very large challenges. This depends on whether they can catch the next wave, whether they dare to make huge, high-risk investments, and do something unproven. Do they have great CEOs like Jobs and Musk? Li Kaifu believes that if any of these conditions are successful, there is a possibility. The CEOs of the above-mentioned four companies are all very outstanding, so there is potential. In addition, there is also an adverse factor of globalization. The United States can naturally sell everything to the whole world, while China still needs to explore in this area. However, recently, for example, TikTok of ByteDance, Temu of PDD Holdings, and Shein, from these companies, it can be seen that even when Chinese tech companies go global, they are still very competitive. Regarding the resistance that Chinese companies face when going overseas, Li Kaifu stated that unfortunately, geopolitics is a fact that we must face, and this kind of resistance will continue to exist, but the world is not only the American market. Like TikTok, which has many users in various regions around the world, including Southeast Asia and Southeast America, it will miss some business opportunities, and the TikTok case in the United States is still pending. If TikTok is taken down in the US, does ByteDance still have the opportunity to become a great company?

During the conversation with Gelon, Kai-Fu Lee once again talked about Nvidia.

As early as December 2016, Kai-Fu Lee accurately pointed out the most important AI investment target at the Glon Hui Global Investment Carnival event.nvidiaIn the past seven years, Nvidia has grown by 50 times, and by 338 times in ten years.

At this point, if Kai-Fu Lee were to create a portfolio of 3-5 future AI companies to invest in (whether they are listed or not), which ones would he choose?

In response, Kai-Fu Lee said that if it includes companies that are not yet listed, he is very confident in Zero One Huanwu.

Up to now, there has not been a great company in China with a market cap of a trillion U.S. dollars, and AI will be a huge space.Kai-Fu Lee believes that any company in the top tier models is worth investing in right now.For example, the investment target, Anthropic, especially OpenAI.

He thinks that OpenAI will increase by ten times in three years, of course, it may not necessarily be successful, but it is high probability.

So if he had to choose one, he would choose OpenAI.

Because its risk is controllable, and there is still a lot of upside room.

Looking at it from a different perspective, which companies, even without a top tier model, can still achieve tremendous growth through the popularization of large models throughout their businesses.He believes that Microsoft is definitely one of them, as is Tesla.

Tesla's recent autonomous driving technology is very impressive, in a sense, it is similar to a large model, and it is an end-to-end solution, greatly improving the entire autonomous driving experience.Of course, Microsoft and Tesla may not have as much growth potential.

Businesses in China have a good chance to capture a massive value of AI space because there has not been a trillion-dollar market cap company in China yet.

Kai-Fu Lee thinks that Microsoft is definitely one of them, as is Tesla.

Tesla's recent autonomous driving technology is very impressive, in a sense, it is similar to a large model, and it is an end-to-end solution, greatly improving the entire autonomous driving experience.

Which other companies will benefit? Li Kaifu thinks it'll be cloud companies since AI is tied to the cloud. AI will bring about the next stage of advancement in cloud computing, so companies like Nvidia or other leading cloud companies are worth considering. Although Nvidia's market cap is already high, Li Kaifu believes there is still room for growth for the next two years. Looking further into the future, there will be a major turning point. Today, most AI computing power may be used in training because it is expensive and time-consuming. Nvidia has an absolute advantage in training chips, but the bigger opportunity in the future will definitely be for inference because its large models will be running behind every search, e-commerce transaction, and ad viewed. Inference is the use of these models. However, Li Kaifu believes that Nvidia doesn't have a particularly large advantage in inference because it can be achieved with simpler chips, including Chinese and American companies. Moreover, Nvidia's culture is built on dominating everyone with technology, it's what they're most proud of. Li Kaifu remembered his first meeting with Huang Renxun when it wasn't yet the era of AI. He asked Huang Renxun if he was worried about small graphic chip companies surpassing him. Huang Renxun replied that even if they did, he's aiming to solve the most difficult problems. There will always be someone who values the most difficult hardware. Therefore, Nvidia's training is so outstanding that it's difficult to surpass. However, when the inference market becomes larger, everyone will have more investment opportunities. Currently, it's not yet clear which company will have the advantage in inference, but AMD is a possibility. AMD's chips are not as good as Nvidia's for training, although they are cheaper. Based on the speed at which other companies use AMD for inference, Li Kaifum can see that there might be a bullish trend for AMD's inference chips, but it may not occur until a year and a half later. In short, Nvidia is definitely bullish for these two years, but it may not be as bullish later on.

AI will bring about the next stage of advancement in cloud computing, so companies like Nvidia or other leading cloud companies are worth considering.Amazon orAlibabaCompanies such as Nvidia are worth referring to.

Although Nvidia's market cap is already high, Li Kaifu believes there is still room for growth for the next two years.

Looking further into the future, there will be a major turning point. Today, most AI computing power may be used in training because it is expensive and time-consuming.

Nvidia has an absolute advantage in training chips,but the bigger opportunity in the future will definitely be for inference, because its large models will be running behind every search, e-commerce transaction, and ad viewed. Inference is the use of these models.However, when the inference market becomes larger, everyone will have more investment opportunities.

Regarding inference,Li Kaifu believes that Nvidia doesn't have a particularly large advantage,because it can be achieved with simpler chips, including Chinese and American companies. Moreover, Nvidia's culture is built on dominating everyone with technology, it's what they're most proud of.

Li Kaifu remembered his first meeting with Huang Renxun when it wasn't yet the era of AI. He asked Huang Renxun if he was worried about small graphic chip companies surpassing him. Huang Renxun replied that even if they did, he's aiming to solve the most difficult problems. There will always be someone who values the most difficult hardware. Therefore, Nvidia's training is so outstanding that it's difficult to surpass.

However, when the inference market becomes larger, everyone will have more investment opportunities.

Currently, it's not yet clear which company will have the advantage in inference, but AMD is a possibility.

AMD's chips are not as good as Nvidia's for training, although they are cheaper.

Based on the speed at which other companies use AMD for inference, Li Kaifum can see that there might be a bullish trend for AMD's inference chips, but it may not occur until a year and a half later.

So in short, Nvidia is definitely bullish for these two years, but may not be as bullish later.

So in short, Nvidia is definitely bullish for these two years, but may not be as bullish later.

AMD may have a bullish trend after one and a half years.But another company may outperform AMD, after all the threshold for logic chips is not that high.

For more exciting and complete content, follow Dr. Gelon's Studio on Douyin/Short video platform.