In the era of AI, the concept stocks of humanoid robots continued to rise strongly in 2025. Tesla, Nvidia, Ubtech, Huawei and other giants have been laying out humanoid robots, with Citigroup boldly predicting that by 2050, the market size of humanoid robots is expected to reach 70 trillion US dollars. In 2025, how long can humanoid robots remain strong? Fuguo's Xu Zhixiang recently shared his views. In the list of the top 50 public fund returns in the media statistics as of February 10, Xu Zhixiang's Fuguo New Materials and New Energy products ranked ninth. Xu Zhixiang, senior fund manager of Fuguo Fund Equity Investment Department, has been in the industry for 10 years, formerly a industry researcher at Zhejiang Commercial Securities Co., Ltd. and an investment manager at Ping An Asset Management Co., Ltd. Xu Zhixiang manages a fund size of 4.583 billion yuan, and the funds he manages are as follows: In November 2021, Xu Zhixiang was appointed as the fund manager of Fuguo New Materials and New Energy mixed securities investment fund. Taking Fuguo New Materials and New Energy Hybrid C as an example, the growth rate in the past 3 years was 20.45%, the average return of the same category was 17.66%, and the growth rate of the SSE 300 was -15.13%. The growth rate in the past year was 56.98%, the average growth rate of the same category was 18.16%, and the growth rate of the SSE 300 was 16.05%. The growth rate since the beginning of this year is 26%. In the fourth quarter report of 24, Xu Zhixiang pointed out that he continues to focus on technological innovation in the direction of AI, focusing on stock selection in computing power and applications, paying more attention to domestic chains in computing power, as well as innovative directions such as CPO; AI application hardware mainly focuses on smartphones, glasses, cars, and robots. Currently, the core positions are in the robot, automotive, and computing power sectors. The investment assignment representative has reviewed his past quarterly reports and found that his investments have always been in the field of technological growth stocks. In the first quarter report of 24, he once proposed that the future space of AI applications and robots would reach trillions of levels. Currently, his leading performance continues to support his views. Now the difficulty of investing in the growth sector is still relatively high. The best direction is in stages 1 to 10, with future space reaching trillions of industries. The main areas we have identified include intelligent driving, energy storage, and cloud computing. In addition, we need to rotate in the industries between 0->1 and 10->n. But we firmly believe that there will definitely be industries in stage 01 that can emerge and enter stage 110. This turning point is the singularity of nonlinear growth, and if we can seize this singularity, we will reap generous investment returns. In stages 01 and near 1, the future space will reach trillions of industries. We have identified AI applications, robots, commercial aerospace, wearable devices, 3D printing, etc. In addition, the low-altitude economy, brain-machine interface, high-temperature superconductivity are also developing rapidly. Recently, Fuguo Fund disclosed Xu Zhixiang's views, and excerpts from the investment assignment representative are as follows: In AI applications, he has the most confidence in robots. Q: What made you start laying out humanoid robots in 2022? Xu Zhixiang: The main reason is that I am optimistic about the field of AI applications, and my positions in AI applications are also concentrated in the field of robots. Robots may be the area with the greatest future space in AI applications, and their industrial development is similar to the early investment in new energy. Huang Renxun recently publicly revealed Perception AI, Generative AI, Agentic AI, and Physical AI as the four different stages of AI development, with robots belonging to the fourth stage, with the largest market space. There are two main directions for AI development: General Artificial Intelligence (AGI) and AI democratization. Q: Does the birth of low-cost large models mean that the demand for AI computing power has been disrupted? Xu Zhixiang: Currently, there are two main directions in AI development: one is General Artificial Intelligence (AGI), pursuing extreme performance, aiming to touch the technical ceiling; the other is AI democratization, penetrating downward, making AI technology popular in the mass market. AGI requires very high performance, and the demand for training computing power is very large. Although AI democratization reduces the demand for training computing power, the demand for inference computing power will significantly increase due to the large user base. Therefore, the demand for computing power needs to be considered separately. If AGI is developed, the demand for training computing power will continue to rise, and the demand for inference computing power will also increase; if AI democratization is promoted, the demand for training computing power will decrease, but the demand for inference computing power will significantly increase. This year may be the tipping point for the mass production of humanoid robots. Q: Has the tipping point for humanoid robots come? Xu Zhixiang: This year may be the tipping point for the mass production of robots. By comparing the production pace of robots with the production pace of electric vehicles in the past, it can be divided into three stages, with different production volumes and prices at each stage. From the first stage to the second stage, after the cost is reduced, exports can be made, mainly for B-end customers. After entering the third stage, with a significant cost reduction where it corresponds to the C-end market. Therefore, this year starting the production of training robots, next year may be aimed at the B-end, and the following year may begin to target the C-end. As China's supply chain was introduced early this time, the pace of mass production climbing will be very fast. In the mid-term, there are better investment opportunities, and the focus is on TMT and related AI directions in the future. Q: Is it still suitable to enter now? Which direction of the market is more sustainable? Xu Zhixiang: In the mid-term, there are still good investment opportunities, but the future performance needs to be considered based on the policies that will be successively introduced later. From the perspective of technological development stages, large models correspond to the Generative AI stage, while robots correspond to the later Physical AI stage. In terms of landing speed, large models are expected to achieve commercial applications before robots. In terms of market space, the potential of the robot direction may be greater. Overall, as AI develops further, the focus of investment in AI will lean more towards the application end, and in the future, TMT and related AI directions will be more focused on. In the robot sector, a focus on the automotive and machinery sectors may be beneficial. This article is reprinted from the public account "Investing Work Pro", GMTEight editor: Li Fo.