State Grid Corporation of China published its 2026 half-year report on July 24, revealing fixed asset investments exceeding 310 billion yuan in the first six months, a year-on-year increase of 12.6%. Simultaneously, China Southern Power Grid completed nearly 89.3 billion yuan in investments during the same period, up 14.79% year-on-year, also setting a new historical high for the comparable period. The two power grid companies together spent nearly 400 billion yuan in half a year. These figures are already astonishing, but the more intriguing aspect lies in the structural changes hidden behind the numbers. While the total capital remains substantial, the way it is being spent has completely diverged from the past five years.

Previous Investment Logic: Expand Lines and Build More Stations

Looking back at the "14th Five-Year Plan" period, the core logic of power grid investment boiled down to one word: expansion. The explosive growth of new energy installations saw wind and solar power transition from "supporting roles" to "mainstays," with installed wind and solar capacity reaching 1.84 billion kilowatts by the end of 2025, accounting for over 47% of total power generation capacity. However, the grid remained the same old infrastructure, with many lines built ten or even twenty years ago. Consequently, the primary investment theme during those years was clear: where new energy was built, the grid was extended. As large-scale bases in the Gobi and desert regions were established, ultra-high voltage (UHV) lines were stretched to connect them. The logic was straightforward: the more new energy installed, the more the grid expanded. The results were evident. However, problems soon emerged. The utilization rate of solar power dropped from 98% in 2023 to approximately 91% in the first quarter of 2026, meaning nearly 9% of every 100 kilowatt-hours of solar power could not be transmitted. The curtailment of wind and solar power is resurfacing, not because the grid is not "large enough," but because it is not "smart enough"—it lacks the flexibility to adapt to the fluctuating output patterns of new energy sources. Expanding scale solves the "availability" issue but fails to address the "adaptability" problem.

Current Investment Logic: Shifting from 'Expansion' to 'Adaptation'

The phrasing of the "15th Five-Year Plan" for power grids has quietly changed. State Grid plans 4 trillion yuan in investments, a 40% increase over the "14th Five-Year Plan" period; China Southern Power Grid aims to approach the trillion-yuan mark, scheduling 180 billion yuan for 2026. The National Development and Reform Commission stated at a press conference on May 22 that total grid investment during the "15th Five-Year Plan" period will exceed 5 trillion yuan. However, where this money is being spent now has a completely different direction.

First, adapting to the instability of new energy. Solar power generates during the day but drops to zero at night; wind power surges when the wind blows and stops when it is calm. This output pattern severely mismatches traditional electricity load curves. The grid needs to shift from "I build what you use" to "I accept whatever you generate." This requires substantial investment in flexible regulation capabilities—flexible DC transmission, energy storage integration, intelligent dispatch systems, and regional interconnection channels. State Grid plans to put 15 UHV DC projects into operation during the "15th Five-Year Plan" period. The core function of these projects is no longer simply "transmitting power from west to east," but using flexible technology to smooth out the unstable wind and solar power from the western regions before delivering it to load centers in the east. The technology road has changed, and so has the way capital is spent.

Second, adapting to residential electricity consumption habits. An often-overlooked change is that residential electricity use has become one of the fastest-growing segments. The average annual growth rate during the "14th Five-Year Plan" period reached 7.7%, with residential electricity consumption hitting 1.59 trillion kilowatt-hours in 2025. The most significant growth driver is air conditioning load. When air conditioners are turned on in summer, the load instantly spikes, accounting for 30% of total social electricity consumption, and over 40% in some provinces. This "weather-dependent" peak load poses a critical impact on the distribution grid—not because of insufficient total capacity, but because the peak arrives too suddenly and too concentratedly. Therefore, distribution grid transformation has become an investment priority. The goal is not to build larger substations, but to enable the existing distribution grid to withstand short-term extreme loads and absorb the reverse power flow from distributed solar photovoltaics.

Third, adapting to the electricity demand of AI computing power. This is perhaps the most forward-looking change. Electricity consumption related to computing power in the eight national hubs has grown at a rate of 39.5% over the past three years. In 2025, national computing power electricity consumption was about 170 billion kilowatt-hours, which is expected to surge to 800 billion kilowatt-hours by 2030—an increase of nearly five times. A more challenging issue is the severe mismatch between computing power load and new energy output time curves. Solar power peaks at noon, but AI data centers require round-the-clock, uninterrupted electricity. This means that simply "powering computing with green electricity" is insufficient. There must be "computing-power coordination" at the grid level, where computing loads dynamically adjust to the pace of power supply, or matching is achieved through energy storage and time shifting. The grid needs to transition from "passive response" to "active adaptation," which is an entirely new capability dimension.

How to Understand This Change?

Simply put, over the past five years, grid investment addressed the "availability" issue—if new energy was built but couldn't be transmitted, the solution was to build more UHV lines and extend more power lines. For the next five years, the focus is on the "adaptability" issue. With the unstable output of new energy, the peak load from residential electricity use, and the round-the-clock high load of AI computing power all converging simultaneously, the grid does not need to be larger; it needs to be more flexible, more intelligent, and better able to handle uncertainty. The 5 trillion yuan investment scale is indeed staggering, but what is more noteworthy is where and how this money is being spent. Shifting from "expanding scale" to "improving efficiency," and from "passive response" to "active adaptation"—this is the true character of this cycle of power grid investment.

Want to Participate in the Dividends of Power Grid Investment? Two ETF Portfolios

For ordinary investors, it is understandable that the 5 trillion yuan investment will bring significant dividends to many listed companies in the industry. The theme of power grid investment is clear, but the industry chain is long, with many links and high technical barriers, making stock selection difficult. Investing via ETFs may be a more practical approach. The Power Grid Equipment ETF (159326) is currently one of the few ETF products on the market closely tracking the CSI Power Grid Equipment Thematic Index. It has a size of approximately 19 billion yuan, with over 75% exposure to power grid equipment and a 75% weighting in UHV, covering core leaders like NARI Technology, TBEA, and Sieyuan Electric. It is a "pure" choice for investing in the power grid equipment sector. The Green Power ETF (562550) tracks the CSI Green Power Index, with a size of about 2 billion yuan—the largest among similar products. It has over 96% exposure to power stocks, with "wind, solar, hydro, and nuclear" content exceeding 55%, and includes many stocks related to the computing-power coordination concept. If you are more optimistic about the intersection of "new energy + computing power," this ETF covers that area more precisely. The 5 trillion yuan investment cycle is just beginning, with the first half of this year being only a start. Signal of MACD golden cross forming, these stocks are showing promising upward momentum.