Electricity will be the star over the next five years.

As the ancient saying goes, 'While living in peace, one must think ahead to potential dangers.' Although China’s power system is already world-leading, it remains insufficient to fully meet the demands of emerging fields and new development paradigms—hence continued advancement is imperative. Specifically, three key areas require focused investment: First, upgrading and expanding ultra-high-voltage (UHV) transmission networks—the ‘highways’ of electricity, foundational to everything else. Second, modernizing urban and rural distribution grids—the ‘last mile’ of the power network and the ultimate point of power consumption. Third, digitizing and intelligentizing the grid—the core driver for improving grid efficiency. Together, these three domains constitute China’s future-oriented, next-generation power grid—a massive undertaking still in progress, with preliminary completion targeted for 2030. Accordingly, average annual investment in grid projects over the next five years is projected to reach 1 trillion RMB—far exceeding the 553.6 billion RMB average during the 14th Five-Year Plan period. This represents a huge, high-growth market—and an entirely incremental one. Beyond this next-generation grid, another much-discussed concept is ‘computing–electricity synergy’—a field poised for explosive growth. When it comes to infrastructure, China has always prioritized overcapacity over insufficiency. ‘Computing–electricity synergy’ recently gained widespread attention amid the global AI wave. Any technology comprises two essential components: infrastructure and applications. For AI, electricity *is* the infrastructure—while hardware components like GPUs and CPUs are merely individual parts among many. Without reliable, abundant power, even the most advanced chips are rendered useless. Computing power is the soul of AI—and its sole consumable resource is electricity. In fact, AI is an enormous electricity guzzler: a single data center requires vast amounts of power—not only for computation but also for cooling and other operations. Given China’s existing advantages in power generation and grid infrastructure, global leadership in computing power is destined to rest here—not something solvable by Elon Musk selling tens of billions of dollars’ worth of solar panels or launching data centers into space. Electricity is a massive, long-accumulated system. Since China holds this advantage, it will inevitably leverage it to the fullest—put simply, to make computing power as affordable as possible. This is precisely the logic behind the long-promoted ‘Eastern Data, Western Computing’ initiative: western China boasts abundant power resources—whether hydropower or new energy sources—especially Xinjiang, where new-energy capacity is expanding rapidly year after year, resulting in extremely low electricity costs. The ‘Eastern Data, Western Computing’ project continues advancing. Of course, cost reduction isn’t limited to lower electricity tariffs alone; reducing power consumption itself is equally effective. Recently, authorities promoted underwater data centers—China built the world’s first such facility in the East China Sea, housing several thousand servers. Though modest in scale, its successful operation paves the way for countless similar deployments. Data centers are not only highly energy-intensive but also water-intensive. Typically, around 40% of their total electricity consumption goes toward cooling. Lower ambient temperatures dramatically reduce cooling-related power use—making the ocean an ideal location. Nearshore seawater maintains an average temperature of roughly 15°C—a major advantage. On land, typical data centers operate at a Power Usage Effectiveness (PUE) of about 1.5. PUE measures energy efficiency: the closer to 1.0, the higher the efficiency—and thus the lower the operating cost. Underwater data centers can achieve a PUE below 1.15. Beyond superior energy efficiency, these underwater facilities run on green electricity—powered directly by offshore wind farms—hence the term ‘computing–electricity synergy.’

Tech giants are scrambling for electricity—but no one sees the real battlefield.

GPUs can be stacked infinitely; model parameters can grow exponentially. Yet once a single AI rack exceeds 250 kW in power demand—equivalent to the simultaneous electricity consumption of 300 ordinary households—the bottleneck shifts subtly away from algorithms and chips—and squarely onto the supporting power system itself. On May 26, the National Energy Administration convened the national ‘AI+ Energy’ Field Promotion Conference in Shenzhen. The attendee list carried profound symbolic weight: three state-owned energy giants—State Grid Corporation of China, China National Petroleum Corporation (CNPC), and China Energy Investment Corporation—shared the stage with three private tech firms—Envision Group, Alibaba, and Tencent. This marks a historic convergence at the crossroads of two previously distinct, trillion-dollar industries: AI and energy. Zhang Lei, Chairman of Envision Group, delivered a paradigm-shifting statement at the event: ‘The power system is becoming the principal engineering backbone of artificial intelligence—not merely a supporting utility.’ When Moore’s Law governing chips and models collides with the slow, incremental evolution of traditional power systems, the AI industry’s fundamental bottleneck is shifting—from algorithms and chips—to ‘end-to-end energy management across the entire intelligence production chain.’ The breakthrough lies in the organic integration of energy, chips, and computing centers—and AI-powered electricity systems are emerging as the new foundational infrastructure enabling that integration. After AI’s breakneck acceleration, traditional grids are falling behind. In recent years, AI technology has evolved into a multi-trillion-dollar industry commanding massive venture capital investment. Since 2022, the combined market capitalization of AI-related companies in the S&P 500 has surged by approximately $12 trillion. Especially over the past two years, global computing demand has grown significantly. According to the ‘China AI+ Energy Development Report 2026,’ compiled under the guidance of the National Energy Administration, ‘AI + Energy’ has become a critical strategic priority for major countries vying for leadership in future development. AI’s rapid expansion is driving sustained growth in global power demand from computing infrastructure. The International Energy Agency (IEA) forecasts in its 2026 report ‘Key Issues in Energy and Artificial Intelligence’ that data center electricity demand will surge further post-2030. The U.S. is expected to account for the largest share of this growth, followed by China. Data center electricity consumption is projected to rise from 485 terawatt-hours (TWh) in 2025 to 950 TWh in 2030—exceeding Japan’s current total national electricity consumption. The core driver behind surging computing demand is the Artificial Intelligence Data Center (AIDC). As the foundational infrastructure for storing, processing, and deploying AI’s massive datasets, AIDCs have seen electricity demand skyrocket due to the explosive growth of generative AI—far outpacing traditional data centers. For years, AIDCs competed fiercely for GPUs and high-bandwidth memory (HBM). But recently, as AI has transformed into an ‘electricity-hungry monster,’ a new bottleneck has emerged: traditional power systems simply cannot keep up. Tech giants are buying electricity—but they’re not addressing the fundamental question.

IEA chief says electrification remains the way forward despite AI investment jitters

4 days ago

IEA Chief Fatih Birol reaffirmed global electrification priorities despite AI investment concerns, noting 61% of energy investment now targets electricity. He launched a new Asian energy security partnership to address grid capacity hurdles and LNG market volatility amid Middle East instability.

reuters

2026-05-11

Iran War Undermines Trust in Hormuz Transit, IEA’s Birol Says

2026-07-10

IEA Chief Urges EU to Reconsider Arctic Drilling Moratorium

2026-07-10

IEA Head Urges EU to Reconsider Opposition to New Arctic Oil Drilling

2026-07-15

IEA Executive Director Warns Global Economy in Peril If Hormuz Crisis Persists

2026-07-17

IEA Warns EU Methane Regulations Could Undermine Energy Security