From computing power to electricity, China's new paradigm for green energy going global

On March 30, the green power sector saw a significant pullback, making many investors wonder: at this point, can it still be allocated? We believe the current adjustment is not the end of the trend, because the core investment logic of the green power sector has not changed.

Upgrading the Energy Security Strategy. Since the end of February 2026, geopolitical risk events in the Middle East have continued to escalate, and we have yet to see clear signals of easing. The Strait of Hormuz— the global “throat” for oil and gas transportation— has been disrupted, and the risk of supply interruptions remains persistently high. Brent crude oil prices have surged from $70 per barrel to above $100, hitting a new high since 2022. This means that geopolitical risks to energy supply have not been resolved, and elevated oil and gas prices may become the norm. Against this backdrop, the importance of energy security strategy has increased.

The fragility of traditional energy accelerates a repricing of the value of green power. The daily average crude oil transportation volume through the Strait of Hormuz accounts for about 20% of global oil trade, and LNG trade’s 20% also transits through it. For countries worldwide, geopolitical conflicts have exposed structural shortcomings in the supply of traditional energy. In this context, green power such as solar, wind power, and hydropower is no longer limited to environmental protection topics, but has become a core component of an energy security strategy.

With energy security becoming the main line, the investment logic for the green power sector also becomes clear. From a strategic position, China’s power industry—whether measured by generation volume or installed capacity—stands far ahead globally, with clear advantages in both scale and systems. From the market-space perspective, the surge in AI computing power demand has opened up incremental space for green power. From the value perspective, exporting computing power provides an entirely new path to monetize power value. From the valuation perspective, China’s green power leaders are currently generally valued at 15–20x. Therefore, amid increasing uncertainty in traditional energy supply, the green power sector—which combines “strategic rigid demand + market space + valuation advantages”—is at the starting point of a value repricing.

Within the energy security theme, which industries are China’s advantage sectors? Prefer power assets: simultaneously meeting scale leadership + system leadership

China’s green power industry has already formed a globally leading scale advantage. As of the end of 2025, the country’s photovoltaic power installed capacity reached 1.2 billion kW, wind power installed capacity reached 640 million kW, and total renewable energy installed capacity exceeded 2.3 billion kW, accounting for about 60% of the country’s total power installed capacity. China has built the world’s largest clean energy power generation system.

China has globally unique power grid infrastructure. Unlike the United States’ power grid, which is divided among many private companies and regional operators—and whose new transmission line construction cycle faces a structural difficulty lasting 5–10 years—China has a unified national-level grid dispatching hub, enabling optimized allocation of power resources across regions and across time periods. The dual advantages of “hardware + system” provide the foundation for China’s green power to make a leap from scale expansion to value creation.

Is the market space for green power big enough? In the AI computing power wave, green power becomes a new rigid demand, and the market space is accelerating open

The end of computing power is electricity. The China Academy of Information and Communications Technology (CAICT) conducted multi-scenario forecasts for China’s computing power electricity demand. Based on the development trajectory of artificial intelligence technology, it constructed three differentiated development scenarios: high, medium, and low. In the high scenario, AI grows explosively; by 2030, electricity consumption of computing power centers in China may exceed 7000 billion kWh, accounting for 5.3% of total electricity consumption across society. Computing power centers are both core AI facilities and major electricity-consuming entities, so the end of AI computing power is energy. This means that a rapidly expanding computing economy needs an electricity base of an equivalent magnitude to support it.

Policy has already clearly supported the development of green power. In 2026, “computing-power and power coordination” was written into the Government Work Report for the first time. The National Data Bureau clearly requires that the share of green power applications in newly built computing infrastructure at national hub nodes reaches 80% or more. The deep integration of green power and computing power also has the potential to bring entirely new business model exploration for green power operators. The continued release of green electricity demand from domestic data centers is locking in long-term, stable utilization capacity for green power operators.

How will China’s electricity go global? Through computing-power and power coordination, exporting via Token, and successfully monetizing power advantages

Electricity is difficult to transport across borders, but computing power can be. For a long time, although China’s clean energy is massive in scale, it has been constrained by the boundaries of the physical power grid, making it hard to directly participate in global trade. With the explosion in demand for AI large models, a brand-new export pathway is taking shape—turning green electricity into computing power, then transmitting it across borders in the form of Tokens via fiber-optic networks, achieving “power value going global” while “power does not leave the country.”

Computing power exports are becoming a new growth pole. According to OpenRouter data, in February 2026, China’s models surpassed the United States for the first time on the core metric of Token call volume. Over the past year (from February 2025 to February 2026), China’s Token consumption share increased by 421%. Chinese models such as MiniMax M2.5 and DeepSeek V3.2 are among the leaders globally in call volume. China’s low-cost advantage in exporting computing power is huge. For example, the input price for one million Tokens for MiniMax is only $0.3, which is 1/20 of that for overseas models in the same tier.

Behind “exporting computing power” is a large increase in the value of electricity. In new energy-rich regions such as Guizhou and Yunnan, the feed-in tariff for wind and solar power is about 0.3 yuan per kWh. Consuming electricity of about 15–20 degrees (kWh) to generate 1 million Tokens means the electricity cost is only in the single-digit renminbi. Meanwhile, the international market’s pricing for the output of similar Tokens is about $60–168 per million Tokens, enabling an order-of-magnitude increase in export value realization. Under the “electricity-to-computing-power” model, the digital value supported by each kWh can reach several times to dozens of times that of the traditional mode.

As AI interactions evolve from “simple Q&A” to “autonomous Agents,” the Token consumption amount grows at a geometric rate; this expansion on the demand side opens up new space for electricity utilization.

Why choose the CSI Green Power Index?—Low valuation, thematic catalysts, and broad coverage

Low valuation: green power may be poised for a value reassessment

China’s green power companies are currently still in a valuation trough, with valuations around the 50th percentile over the past five years, and there is still logic to further raise the valuation center. Under the new paradigm of “computing-power and power coordination,” green power operators have deeply integrated into the AI industrial chain. The valuation framework for the green power sector is shifting from “utilities” to “a digital power base.”

Green power operators directly benefit from catalysts driven by increased computing power demand

When the three main lines of energy security, AI computing power, and going global intersect, the value of green power is rising significantly, and green power operators will directly benefit from the incremental computing power demand. Computing power demand locks in long-term growth space for green power, while exporting computing power opens up a new path for monetizing power value.

The index covers generation assets more comprehensively

In terms of constituent stocks, the CSI Green Power Index mainly focuses on green power areas such as hydropower, wind power, and nuclear power. The top ten constituents include green power leaders such as Yangtze Power and China National Nuclear Power, and it also includes enterprises such as Guodian Power, which are transitioning from thermal power.

Currently, among the indices tracking green power in the market, the main ones are the CSI Green Power Index and the Guozheng Green Power Index. Compared with comparable indices, the CSI Green Power Index provides broader coverage of the core power theme. E Fund CSI Green Power ETF (562960, OTC fund connection A/C: 019058/019059) tracks the CSI Green Power Index. One-click packaging of wind/solar, hydropower, and thermal power transition leaders can serve as a high-quality tool to capture the beta of the transition to a new power system.

Table: Comparison of the two green power indices

Table: Top ten constituents of the CSI Green Power Index

Source: Wind; data as of February 27, 2026.