While the broader market experienced high-level volatility and repeated tug-of-war, the power sector collectively surged. As of May 19th, Jingneng Power secured five consecutive limit-ups, while Shanghai Electric Power and Huaneng Mengdian all hit the limit-up.
Prior to this, power leader Datang Power (601991) had six consecutive limit-ups and recently surged again, with a monthly gain of 77%. Meanwhile, the ChinaAMC Green Power ETF saw net capital inflows for eight consecutive trading days, collectively "absorbing" 264 million yuan, with its scale reaching a record high since its inception.
Beyond the secondary market, energy storage and green power enterprises also ushered in a wave of IPO listings. According to incomplete statistics, since the beginning of this year, several new energy companies including SiGe New Energy, Taijin New Energy, Goode Electric Materials, and others have entered the capital market. More importantly, these companies saw their stock prices soar immediately upon listing.
Data shows that investors behind these IPOs have reaped returns of tens or even hundreds of times; some early-stage investors achieved returns exceeding 600 times, becoming a landmark scene in the primary market's wealth feast. This also marks the long-awaited turnaround moment for "long-silent" new energy investors.
With AI, the Sector Logic Has Completely Changed
Regarding this surge in the new energy power sector, some investors believe it's not a simple traditional catch-up rally, but rather the prelude to a revaluation of new energy assets in the AI era.
Rewind two years, how difficult was life for new energy investors? It can be described as "valuation halved, financing cooled."
At that time, after valuations hit the ceiling, the entire sector fell into the awkward dilemma of valuation shrinkage. Since the second half of 2022, in the secondary market, valuations for sectors like lithium battery materials and power battery cells were directly slashed to one-third to one-fifth of their peak levels; photovoltaic modules and silicon material sectors fell more than 80% from their highs. In the primary market, a large number of projects experienced valuation inversion, where subsequent financing rounds were priced lower than the previous round's valuation. Institutions became very cautious, and the financing pace abruptly cooled.
One investor recalled that the hardest part was not the moment the market took a sharp turn for the worse, but the prolonged industry bottoming period. "In many instances, you might know deep down that the long-term logic of this sector is sound, but you just couldn't see a short-term turnaround."
During this process, influenced by market cycles, policy changes, and investment return expectations, some investors chose to exit: they directly pivoted to sectors like semiconductors and AI, which were seen as more certain at the time; others adjusted their strategies, investing in infrastructure assets with more stable cash flows and stronger anti-cyclical capabilities, such as photovoltaic power stations and charging stations.
However, everything changed in 2026.
As computing power infrastructure and new power grid projects in various regions entered the centralized filing and bidding stages, green power and energy storage became keywords. I noticed that, unlike before, new regulations require that newly built large-scale computing power centers must be equipped with green power supply and energy storage facilities. Projects failing to meet standards will not be filed or connected to the grid.
This means that with the explosion in computing power demand and the approaching summer peak electricity consumption period, the certainty of green power entering an upward cycle has increased.
In fact, previous policies had already given clear signals. In April of this year, the National Development and Reform Commission, the National Energy Administration, the Ministry of Industry and Information Technology, and the National Data Administration jointly issued the "Action Plan for Promoting the Bidirectional Empowerment of Artificial Intelligence and Energy," directly addressing the core pain points in current energy and AI development. Among them, "Computing-Power Synergy" was formally elevated to a national strategy, aiming to preliminarily build a safe, green, and economical energy guarantee system supporting AI innovation and development by 2027.
A series of policies and new regulations transformed the once-burdened sectors of energy storage, green power, and smart scheduling, turning them into some of "the most certain" and "most worth investing in" assets within AI infrastructure.
"Now, power is no longer a traditional manufacturing industry; it is the core underlying layer of AI infrastructure," an investor bluntly stated. When power becomes the underlying energy foundation supporting AI operations, what power determines is no longer just efficiency or cost, but industrial structure and long-term competitiveness.
Thus, a new narrative of "computing power is power, green power is an asset" is gradually unfolding within the new energy power sector.
Valuations of Some Top Projects Have Already Recovered by 60%
The change in sector logic is most directly reflected in investment and financing data.
Since the beginning of this year, investment in the lithium battery industry chain has noticeably accelerated. Data from CVCVSource shows that in the first quarter alone, the 85 expansion projects disclosed had a total investment exceeding 220 billion yuan, close to half of the total scale for the entire year of 2025. Among them, companies like Fuling Precision and Ronbay Technology have successively announced expansion plans worth tens of billions.
On the financing side, accompanied by the capital myth of energy storage dark horse SiGe New Energy, whose stock soared upon listing and saw its market value exceed 160 billion HKD on the first day, financing events in the new energy sector have gradually warmed up. Taking China Resources New Energy, which passed its A-share IPO hearing in April this year, as an example, its fundraising scale of 24.5 billion yuan not only became the largest IPO in the history of the Shenzhen Stock Exchange but also set a new fundraising record in the new energy industry.
However, in the second half of 2024, China Resources New Energy's valuation was only 12 billion yuan. At that time, this green power operation platform was also experiencing the industry's "low point" characterized by subsidy phase-outs, declining electricity prices, and rising curtailment rates for wind and solar power. However, after completing a new round of strategic financing in March 2026, the company's post-investment valuation had risen to 19.5 billion yuan, a recovery of about 62%, with all funds coming from industrial capital and long-term infrastructure funds.
As the logic of "green power = core asset of the computing power base" became a market consensus, money in the primary market flowed more clearly and definitively towards assets related to the "computing power base."
In terms of financing scale, I noticed that the single-round financing amounts in the new energy power sector this year also showed an upward trend. Data shows that Gongji Technology, led by a Tsinghua-affiliated team focusing on grid-style computing power scheduling, completed a nearly 100-million-yuan Pre-A round of financing in May 2026. Additionally, energy storage system solution provider Huazhi Energy and battery material manufacturer Huayou New Materials both secured financing at the hundred-million-yuan level.
Some investors noted that currently, some top new energy power projects are recovering their valuations very quickly amid the sector logic shift. In their view, current AI development faces four bottlenecks: computing power, transmission capacity, storage capacity, and power. Among these, power is the most core constraining bottleneck.
"It's hard to imagine that building a data center in the US takes only 1-2 years, but the supporting power and grid infrastructure take 3-7 years," the investor said. Therefore, when AI data centers meet high base loads, combined with extreme peak demands, power must shift from being 'cheap' to being 'reliable.'"
Against this backdrop, capabilities within the green power system, such as power grids, energy storage, and adjustable power sources, will usher in a new round of investment opportunities, and industry value will be revalued.
"Opportunity Favors the Prepared"
"Having gone through cycles, looking back now, the most difficult two years might have been the best sowing period for new energy power investment," one investor remarked with emotion.
Take the energy storage sector as an example. From 2024 to early 2025, energy storage cell prices hit rock bottom, system integration gross margins were compressed to single digits, and a large number of small and medium-sized manufacturers were forced out. At that time, daring investors could acquire stakes in leading companies at prices less than one-third of the peak period. But by 2026, with the implementation of the "Computing-Power Synergy" policy and the significantly increased certainty of the green power upturn cycle, energy storage has been upgraded from an "optional backup power" to a core infrastructure for AI's active power supply, leading to a massive explosion in demand.
The latest Morgan Stanley research report shows that driven by AI demand, capital expenditure forecasts for hyperscale cloud service providers in 2026 have been significantly revised upwards, with data centers expected to face a power gap of 55GW.
What does 55GW mean? It's equivalent to the installed capacity of about 27 large nuclear power units, or exceeding the electricity demand of over 100 million households. Predictions from the China Academy of Information and Communications Technology are even more substantial: by 2030, electricity consumption by China's computing power centers may exceed 700 billion kWh, accounting for 5.3% of the total societal electricity consumption, with AI demand being the main growth driver.
The sharp increase in power demand directly boosts the demand and incremental space for power investment, grid upgrades, energy storage installation, and green power usage, elevating data center electricity consumption from being a "scrap" load on the grid in the past to becoming the main engine driving investment.
Against this backdrop, the primary market has also become "lively." "Currently, liquid cooling and temperature control, large-scale energy storage integration, virtual power plants, UHV distribution networks, nuclear power base loads, and green power supporting facilities are the most worthwhile sectors to invest in," one investor said. Another investor, prioritizing investment areas, stated they would prioritize investing in power management, megawatt-level distributed energy, and data center backup power sources.
Furthermore, some investors are optimistic about AI computing power and data center supporting projects, such as insulating fibers for liquid cooling and immersion cooling fluids. This is akin to investing in the "water sellers" of AI infrastructure first.
Acute-scented PE giants have already acted. Previously, private equity giant KKR had sensed this historic opportunity and began early layout in the energy infrastructure sector underlying AI computing power.
Unlike crowding into popular targets like large models and chips, KKR focused on hard-core foundational assets like green power, energy storage, data centers, and liquid cooling and heat dissipation. They deployed hundreds of billions of funds globally in wind and solar energy storage bases, AI-specific data center supporting power stations, and high-end liquid cooling technology assets.
The latest scene is that KKR, by investing in the liquid cooling project CoolIT Systems, achieved an investment return of approximately 15 times in less than three years, becoming one of the highest-return investments in KKR's history, truly a case of "turning stone into gold."
Under the new narrative, for new energy investors, this turnaround battle might have just begun; and those investors who persevered through the industry's low period will ultimately welcome their own harvest season.
This article is from the WeChat public account "Touzhong Wang," author: Chen Mei
