Any individual or organization that is not good at using AI tools to transform itself will be eliminated.
“We’ve reflected on it, and there’s one thing we underestimated – our understanding of the power grid wasn’t fast enough.”
When saying this, Hui Hengyu’s tone was very calm, without the deliberate solemnity of reviewing “lessons”. As a managing partner of Chaoxi Capital, he has been investing in the manufacturing industry for more than a decade. He has witnessed the “531” new photovoltaic policy in 2018, the complete cycle of lithium prices dropping from 600,000 yuan per ton to tens of thousands of yuan, as well as the booming new energy trend in 2023 and the global rise of the lithium – battery industry in the past two years. He knows that a downward trend is the norm, and an upward trend will eventually come.
In the past, when the industry predicted the installed capacity, it was always a 50% annual growth. However, hardly anyone seriously asked: Can the power grid handle the electricity generated? “New energy is not the final product; it’s just a power – generating product.” When facing this huge market worth trillions of yuan, he didn’t come up with any astonishing conclusions but corrected a industry – wide cognitive bias. This “bottleneck” was ignored by everyone when the industry was advancing rapidly. It was only when the supply – demand mismatch really broke out that people realized they might have been wrong from the start.
This pragmatism runs through Hui Hengyu’s judgment of the entire cycle. He doesn’t avoid problems or exaggerate anxiety. When talking about the invested enterprises, he said that none of them became “zombie enterprises” because of the cycle. When talking about energy storage, he said, “The market won’t converge to just three to five companies like the photovoltaic market.” When talking about internationalization, he said, “This is already a must – answer question, not an optional one.” Every sentence is supported by data and cases, without redundant rhetoric.
But he’s not the kind of person who only focuses on the present. The so – called “new energy” such as wind, solar, and energy storage in people’s traditional perception is just a small part of what he pays attention to. He closely follows AI, believing that it has opened up a new world for the development of energy and electricity. After the demand increases, the supply side will surely see new players. He also spent a long time explaining the underlying logic of “space photovoltaic + space computing power” and deduced the competition logic of ground – based new energy. “It will take at least five years to verify this space – related direction,” he said. “But we have to track it from the angel and seed rounds.”
This rhythm of “looking far ahead but not being hasty in action” might be the underlying reason why Chaoxi Capital hasn’t been caught up in the bubble or frozen by the cold winter in several cycles. Hui Hengyu put it more straightforwardly: “In the trough, you have to hold on even if you don’t have the courage – no one will take over, and the actual controller is also unable to help you exit. All you can do is stand with the invested enterprises and move towards each other.”
He doesn’t be emotional or beat around the bush. He never takes a step back when he should shoulder the responsibility. This style is somewhat similar to the manufacturing enterprises he invests in – not flashy but solid.
The following is an edited transcript of Hui Hengyu’s interview with “China Entrepreneur” (with some deletions):
In the past two or three years, the most drastic changes have undoubtedly occurred in the lithium – battery and photovoltaic industries. The lithium – battery industry started to decline at the end of 2022 and will emerge from the trough in 2025. The photovoltaic industry entered a downward trend at the end of 2023. The general consensus in the industry is that an inflection point will appear in about a year.
To be honest, since we come from an investment background in the manufacturing industry, we’re mentally prepared for the cycle. In investment, 90% of the work is done before investment – when making an investment, you have to predict whether the enterprise has the ability to resist adversity when the downward cycle comes, whether the founder has a sense of crisis, and how to pre – research and allocate the capital strategy, talent strategy, and market strategy in advance.
In our investment decisions at that time, we didn’t invest much in the photovoltaic main – material industry that relies on the expansion of production scale. In fact, since Chaoxi started to layout minority equity investments, investments related to photovoltaics have accounted for less than one – fifth.

Source: Respondent
Currently, none of the enterprises we’ve invested in have seen their value go to zero due to the cycle. However, we also have some reflections. There’s one aspect where our analysis was insufficient – the impact of new energy on the power grid came faster and more violently than we predicted.
New energy is not the final product; it’s just a power – generating product. To transform from power – generating equipment into actual electricity consumption, it has to cross the hurdle of the power grid – grid connection, transmission, and consumption. There are bottlenecks in each link. The current problem facing the global new – energy development is that intermittent energy sources such as wind and solar power have impacted the power grids of various countries, while the construction cycle of the power grid is relatively lagging.
In the past, people were very optimistic about the prediction of installed capacity, with an annual growth of 30% – 50%. However, they underestimated the carrying capacity of the installed capacity. Now, on the surface, it seems to be a demand problem. In fact, it’s because the demand was estimated too optimistically, leading to a bold expansion on the supply side and a huge supply – demand mismatch. From leading enterprises to small companies, all are suffering from the pain of declining profits, operating losses, and tight cash flow.
As institutional shareholders, what we need to do is not to “step on” the enterprises during the downward period but to help them soberly navigate through the cycle: reduce unprofitable bids, not take market share as the only indicator, ensure cash flow, restrain capacity expansion, and continue to invest in new technologies.
The current relationship between new energy and the power grid, as well as the high demand of AI for stable computing power, have made energy storage even more indispensable. Last year, the industry’s installed capacity was nearly 600 gigawatt – hours, a 50 – to 60 – fold increase in five years. Even so, compared with the existing installed capacity of wind and solar power, it’s still far from enough. The existing global installed capacity of wind and solar power is nearly 4 terawatts. If 50% of it is equipped with energy storage for 2 to 4 hours, the market scale will exceed 4 terawatt – hours.
However, energy storage is fundamentally different from photovoltaics: it doesn’t rely on manufacturing attributes. Photovoltaics are basically commodities, winning by scale and cost. Energy storage is different. It’s not a heavy – asset industry but relies more on solutions – understanding the users and the power grid. The power – grid structures in different countries and regions are completely different. Is it grid – forming or off – grid support? Is it single – phase or three – phase? How to install during delivery? How to conduct subsequent maintenance? There are no standard answers to these questions.

Source: AI – generated
So, the energy – storage market won’t be like the photovoltaic market, which will eventually converge to only three to five companies. It will be a fragmented market. In the large – scale energy – storage field, the patterns of CATL, Hichen Energy Storage, and Eve Energy have been initially established. In the household – energy – storage field, MaiTian Energy, Deye Technology, and Siger New Energy each dominate in regional markets. The industrial and commercial energy – storage field has another set of strategies.
The capital market has a very high enthusiasm for energy storage and can offer a price – earnings ratio of more than 100 times. For enterprises, going public is not the end. There are two real decisive factors in energy storage: one is the in – depth understanding of the power grid, which determines how much profit you can make; the other is the internationalization ability, which determines how much market share you can capture and how far you can go.
The equipment industry is an area where we’ve invested more. Companies like Laplace, Shanghai Lianfeng, and Hefei Xinyihua… still hold leading positions in their respective fields. The characteristic of the equipment industry is that when the downstream expands, orders are full, but when it contracts, the order – confirmation cycle becomes longer, and accounts receivable are the biggest risk.
However, it’s a light – asset industry. Apart from human resources, it basically has nothing else, mainly focusing on R & D, assembly, and on – site debugging. So, from the beginning, we told equipment enterprises: They should make a platform – based layout, taking orders from a variety of sources and from different industries.
The founder of one of our invested enterprises has a very rational saying – if an equipment industry has a continuous stream of orders every day, it’s abnormal. No industry is always expanding production; it’s just a stage – by – stage phenomenon. Therefore, being prepared for danger in times of safety is a compulsory course for entrepreneurs in the equipment industry.
Microtech started to get involved in semiconductor equipment in 2020; Robotek did it even earlier, acquiring an optical – module equipment company around 2019; Shanghai Lianfeng is now engaged in helium recovery – the helium used in semiconductor memory – chip production and rocket launches; Jiangsong Technology is trying out the intelligent parking – lot business…
The advantage of the equipment industry is that the technologies are interoperable, and there’s a large space for platform – based expansion. The speed and determination of transformation depend on the entrepreneurs themselves. It’s really difficult to “go against the current” to explore new battlefields, but if they don’t expand, they’ll be very passive when the cycle comes.
An important change in the past year is that internationalization has changed from a “vision” to a “must – option”. LONGi Green Energy built its first photovoltaic – module factory in the United States. It took about a year from construction to production, and it’s profitable. Canadian Solar’s North American factory was also completed in about a year, which was unimaginable before.
Building a factory in the United States is the most complex. EPC (Design – Procurement – Construction) can’t be outsourced to Chinese people. Land acquisition, land leasing, approval, hearings, environmental protection, resident noise, and employment all have to go through a long process.
The greatest contribution of the photovoltaic industry to China is not just products and technologies but the cultivation of a large number of international talents. In the past 20 years, although some photovoltaic companies have fallen, the talents have remained around the world. They understand the social rules in the United States and Europe and can communicate seamlessly with local governments and residents.
This is the foundation of internationalization ability: It’s not about whether you can speak English but whether you have a group of talents who can be trusted by the decision – making level of Chinese companies and accepted by local society.
Energy – storage enterprises are also following this path. In the future, they need to produce, assemble, and create employment locally, becoming “citizen enterprises” in the local area to dispel customers’ concerns about supply – chain security and subsequent maintenance.
One of our invested enterprises, Hichen Energy Storage, has a saying: “Without a stable domestic market, it’s hard to stand; without an international market, it’s hard to be strong.” The domestic market still needs to be grasped, but many rules are still in the exploration stage, and there may be a temporary imbalance between cost and benefit. Although the overseas market has thresholds, the market doesn’t recognize “low price means good”.
The most certain impact of AI on energy is currently on the demand side. Global computing – power construction is driven by electricity, and these demands are currently mainly concentrated in the United States. However, the energy structure in the United States is mainly based on natural gas, so the demand for gas turbines is triggered first, followed by other energy sources such as photovoltaics. However, all power generation needs to be equipped with energy storage, so the market elasticity of energy storage is actually greater.
This is just the story on the ground. Elon Musk is promoting “space photovoltaic + space computing power”. The sunlight duration in space is three to four times that on the ground. Once the heat – dissipation problem is solved, the unit computing – power cost in space may be lower than that on the ground. However, it will take at least five years to verify this direction, requiring the mass launch of starships, the deployment of space photovoltaics, and a service life of more than 10 years.

Source: AI – generated
So, AI has opened up a new world for energy, not only expanding the demand but also presenting new technological paths. Currently, we’ll track early – stage startups that can clearly explain the first – principle and connect with the ecosystem. It’s not yet the time for large – scale investment.
AI also has a revolutionary impact on cost – reduction and efficiency – improvement on the energy supply side. For example, for the battery electrolyte formula, it used to take two to three months to conduct experiments manually, but with AI tools, it may only take a week. For equipment – debugging parameters, in the past, masters had to try one by one, but now AI fixes the best parameters, and the production – line adjustment will be very fast… Any individual or organization that is not good at using AI tools to transform itself will be eliminated.
Actually, in our internal strategy meetings in the past two years, “AI for Energy, Energy for AI” has long become a core concept that we can’t ignore. In our energy investments, the integration ratio of technology and energy is constantly increasing, which is also the unique solution of our industrial investment institution when facing the new world.
This article is from the WeChat official account “China Entrepreneur Magazine” (ID: iceo – com – cn), author: Miao Shiyu. Republished by 36Kr with permission.
该文观点仅代表作者本人，36氪平台仅提供信息存储空间服务。
36kr Europe (eu.36kr.com) delivers global business and markets news, data, analysis, and video to the world, dedicated to building value and providing business service for companies’ global expansion.
© 2024 36kr.com. All rights reserved.

source