Summary of Key Points
This industry observation from Gengxin Capital, a primary market investment firm, uses the well-documented cycle of the new energy vehicle industry from 2015 to 2020 as a reference to analyze the current fluctuations in the AI and embodied intelligence sectors. Many people panic when they see the sharp drops in robot-related stocks and the cooling of market financing, but there's no need to worry. The underlying foundation of this AI revolution is electricity, which coincides with the benefits of the energy revolution we have already implemented. The current correction is merely a "deep squatting period" before the explosion of the industry, a rare long-term investment opportunity that comes once in a generation, not a signal to exit the market.
---
Simplified Explanation of Key Points
1. The most valuable asset in the AI industry today is no longer chips, but "waitings in the power grid"
When AI first became popular a few years ago, people were scrambling for GPU chips. Now, the scarce resource is actually the right to connect to the power grid. To build a computing center capable of running large models, you need to wait in line for the power grid to allocate electricity. Global tech giants have already secured the power grid expansion plans for the next 2-3 years in many regions.
Why is this? Consider the reality: every piece of text or image generated by AI requires electricity—electricity is generated in power plants, transported to data centers, used by servers to produce results, and then sent to your phone via fiber optics. This entire process consumes energy. While AI computing power is updated monthly (new, more powerful chips may emerge in 3 months), power grid expansion takes years (new power plants and lines take 3-5 years to build). This time difference between the rapid development of computing power and the slow expansion of the power grid represents the biggest business opportunity. In the future, computing centers will be like new factories, and electricity quotas will be like new industrial land. Whoever secures enough electricity will control the core resources of the next industry wave.
2. The current humanoid robot sector is following the same pattern as the new energy vehicle industry in 2017
Many people are worried about the volatile stock prices of companies like Yushu Technology, thinking the robot industry is on the decline. Look back at 2017: the new energy vehicle industry was hyped for two or three years, with stock prices soaring, only to cool down suddenly. Everyone claimed it was a scam, and Tesla's market value dropped to $40 billion in 2019, with rumors of bankruptcy. However, when the Model 3 went into mass production, Tesla's market value soared to over $1 trillion, a 25-fold increase.
The humanoid intelligence sector is now in that "scam-like" phase—too much hype preceded it, and now people are demanding real results. Companies without mass production or actual orders will surely suffer. But if a company can lower the cost of humanoid robots and achieve mass production, like Tesla with the Model 3, there will be huge gains. The current fluctuations are not a sign of the industry's lack of prospects; rather, it's a transition from the market paying for future potential to paying for actual performance, and the volatility is completely normal.
3. Our generation is particularly lucky: we've coincided with the intersection of two major industrial revolutions
Previously, new energy and AI were seen as separate waves of opportunity, but they are merging into one. In the past five years, our country has laid a solid foundation for new energy: two out of every three new cars sold are electric, and we control a significant portion of the global photovoltaic and battery production capacity. Even the hydropower projects in the lower reaches of the Yarlung Zangbo River have started, with a capacity equivalent to three times that of the Three Gorges Dam. This means we have established a reliable and affordable power supply. As AI becomes more practical, its impact will extend beyond mobile devices; humanoid robots, autonomous vehicles, and industrial intelligent devices all require electricity. Our advantages in both areas are unmatched: the U.S. excels in chip technology but lags behind China in power grid construction; we can reduce power costs while our engineers can develop powerful models with less energy. This puts us at the forefront of the next forty years of industrial development.
4. The "cooling" of financing in the AI sector doesn't mean capital is leaving; people are just more cautious
Some say the AI industry is cooling down, with only over 70 billion yuan in funding in the first half of the year. However, this isn't a withdrawal of capital; it's just the end of the reckless spending phase. Now, investors are more selective, investing in either early-stage startups with promising teams or mature companies with solid orders. Companies that have hyped up their concepts without producing anything or with inflated valuations are ignored.
The situation is similar to the peak of the new energy industry in 2018: all investors are waiting and no one dares to act first. The real breakthrough will come from the first products to hit the market. Startups should focus on saving cash and wait for industry consensus to return, as that will give them a significant advantage.
5. This "low point" is a rare opportunity that comes only once in a generation
The article reminisces about 2019 and Tesla, not the company itself, but the moment when everyone thought it was doomed despite being a clear opportunity. Such moments occur only during major industrial revolutions. The internet and new energy industries experienced this, and now it's AI and robotics' turn. The current volatility is temporary; in the long-term, the industry is moving from paying for potential to paying for real results. Don't let short-term fluctuations deter you from seizing this once-in-a-lifetime opportunity.
*Note: The above is a purely analytical interpretation of the industry and does not constitute investment advice.*