Guosheng: AI computing power creates electricity gap; under the gas turbine boom cycle, the domestic industrial chain welcomes development opportunities.

Guosheng released a research report stating that order backlogs for leading gas turbine manufacturers have now extended to around 2030, with unit delivery cycles and construction costs rising sharply. Under this supply-demand mismatch, industry prosperity continues to trend upward. China's domestic gas turbine industry chain has achieved breakthroughs at multiple points, with manufacturers across various segments deeply integrating into the global supply system. Under this global gas turbine demand expansion cycle, we are bullish that domestic manufacturers across the industry chain with technological and cost advantages are expected to fully benefit from this round of industry prosperity dividends. Guosheng's main viewpoints are as follows: AI computing power drives accelerated global data center expansion, making North America's power supply bottleneck prominent, with gas turbines becoming the core on-site power supply solution The AI infrastructure investment super-cycle has arrived, with North American cloud capital expenditure continuing to surge; as hyperscale data centers enter the GW-level construction phase, electricity demand is rising rapidly in tandem. Constrained by factors such as aging existing transmission infrastructure and lengthy grid connection approval cycles, data center grid connection challenges in North America have become prominent, with grid expansion and computing power construction rhythms significantly mismatched. Against this backdrop, developers are gradually turning to on-site power generation solutions. Gas turbines, with advantages including short construction cycles, flexible dispatch, and strong power supply reliability, have become the mainstream choice for data center backup power. New AI data centers in the United States are successively locking in gas power, with U.S. gas power generation capacity under construction reaching 52GW in 26H1, up 76% year-over-year. The global gas turbine industry has entered a high-prosperity cycle, with significant supply-demand mismatch against the backdrop of long-term industry downturn The global gas turbine market landscape is highly concentrated, with the three major manufacturersGE Vernova, Siemens Energy, and Mitsubishi Heavy Industriesholding 85% of the heavy-duty gas turbine market share in 2024 (measured by GW). Benefiting from demand driven by AI data centers and other factors, global gas turbine orders are growing rapidly, with leading manufacturers' orders and performance improving in tandem. In 2025, total new orders for gas power generation equipment manufacturers reached $115 billion, up 43% year-over-year. However, the industry supply system has significantly contracted over the past decade-plus downturn cycle. Total annual manufacturing capacity for the three major producers is estimated at approximately 50-60GW in 2026, with capacity expansion remaining relatively cautious. Current order backlogs for leading manufacturers have extended to around 2030, with unit delivery cycles and construction costs rising sharply. Under this supply-demand mismatch, industry prosperity continues to trend upward. China's domestic gas turbine industry chain has achieved breakthroughs at multiple points, with manufacturers across various segments deeply integrating into the global supply system On the complete unit end, Dongfang Electric Corporation's self-developed G50 completed fully domestically-produced F-class 50MW gas turbine commercial operation and achieved its first complete unit export; Shanghai Electric Group's new orders for gas power generation equipment in 26H1 reached 5.89 billion yuan, up 158.7% year-over-year. On the packaged system end, Yantai Jereh Oilfield Services Group entered the North American data center market with mobile power generation units, accumulating gas turbine orders exceeding $3.1 billion. On the component end, Anhui Yingliu Electromechanical supplies hot-end blades in batch to Siemens Energy, Baker Hughes, and Ansaldo, with its two-machine business in-hand orders at approximately 2.2 billion yuan in 26H1. Himile Mechanical Science and Technology, Xizi Clean Energy Equipment Manufacturing, Hunan SUND Technological Corporation, and others are simultaneously entering the supply chains of leading domestic and international OEMs. Risk warnings: Risk of AI computing power demand falling short of expectations, risk of supply chain and delivery falling short of expectations, risk of technological domestic substitution breakthroughs falling short of expectations, risk of raw material price fluctuations, and policy and geopolitical risks.