Open Source Securities has released a research report stating that the global commercialization of AI computing power is accelerating, driving global and US hyperscale data centers into a high-speed construction phase. The rapid deployment of large-scale AI and supercomputing centers is boosting infrastructure investment, which directly increases electricity demand for computing power, creating a solid foundation for growth in the gas turbine industry.
Traditional US demand is at a "replacement cycle" inflection point, with the 25-30 year replacement cycle for the 2000 installation peak converging with AI-driven demand growth in 2026, forming a "demand resonance" window. The firm expects supply shortages to persist until at least 2031. Here are the key points from Open Source Securities.
Gas turbines and diesel generators are gaining market attention.
As global AI computing power commercialization accelerates, the world's and particularly the US's hyperscale data centers are entering a rapid construction phase. To meet the core needs of adapting to fluctuating computing loads, ensuring high power supply stability, and enabling fast deployment, large US AI data centers commonly adopt gas turbine-based "energy island" self-supply power solutions. Meanwhile, diesel generator sets, with their easy deployment, fast response, and wide adaptability, complement gas turbines, forming a synergistic development pattern.
Data centers become the core driver of gas turbine demand.
AI infrastructure investment is becoming the primary catalyst for gas turbine demand growth. The large-scale deployment of global AI and supercomputing centers is accelerating, with rising infrastructure investment directly driving up electricity demand for computing power, thereby creating a solid incremental demand base for the gas turbine industry. Currently, power supply demand for AI data centers accounts for 20%-25% of new global gas turbine demand, and this share is rapidly increasing. Global AI capital expenditure is projected to grow from USD 232 billion in 2024 to USD 1,416 billion in 2028, with a compound annual growth rate exceeding 50%. By 2030, global data center electricity consumption is expected to reach about 945 terawatt-hours, roughly 3% of total global electricity consumption, approximately double the 2024 level, with annual growth of about 15%. Additionally, traditional US demand is at a "replacement cycle" node, with the 25-30 year replacement cycle for the 2000 installation peak converging with AI-driven demand growth in 2026, creating a "demand resonance" window. The firm expects supply shortages to persist until at least 2031.
Market competition is highly concentrated, with leading companies expanding production cautiously.
GEV, Siemens, and Mitsubishi are the top three leaders. GEV announced plans to reach 20 GW capacity by Q3 2026 and 30 GW by 2030; Mitsubishi plans to double its future capacity; Siemens Energy aims to increase its gas turbine capacity from 17 GW in 2024 to 30 GW by 2030. The combined capacity of these three companies is expected to be around 80-100 GW by 2030. The firm predicts supply shortages will persist until at least 2031. GEVernova reports that its gas turbine backlog and slot reservations have jumped from 100 GW to 116 GW, while Siemens has raised its annual gas turbine demand forecast to 110-120 GW for the coming years, indicating demand growth outpaces supply.
Investment recommendations and beneficiary targets.
For the gas turbine direction, the firm recommends: first, system integration, with 杰瑞股份 deeply tied to OEMs like Siemens Energy and Baker Hughes, holding over USD 2.5 billion in orders for North American data center generator sets; second, core components, with 应流股份 as a supplier of Siemens F/H-class heavy-duty gas turbine blades and 万泽股份 securing orders for multiple Siemens Energy gas turbine blade types; third, supporting equipment, with 西子洁能's gas turbine waste heat boilers being essential for combined cycle operations; fourth, parts and thermal management, with 联德股份 supplying large engine blocks for Caterpillar and 银轮股份's generator set cooling business already in mass production. For the diesel generator direction, 潍柴动力 has broken the foreign monopoly on large-bore engines, with data center power generation product sales increasing about 2.6 times year-on-year in 2025, and its SOFC layout provides long-term investment potential. Beneficiary targets include: 杰瑞股份, 应流股份, 万泽股份, 西子洁能, 联德股份, 银轮股份, and 潍柴动力.
Risk factors: Global macroeconomic volatility, insufficient AI investment, geopolitical risks, industry substitution risks, and US tariff risks.
Comments