combined-cycle-gas-turbines
The Thermodynamic Constraint of Computing Power
The global digital infrastructure is not immaterial; it rests on physical foundations that consume electricity, generate heat, and require fossil fuels to operate continuously. In Southeast Asia, the demand for electricity from hyperscale data centers is expected to quadruple, rising from 2.6 gigawatts (GW) in 2025 to 10.7 GW by 2035. This increase will transform data center consumption from the current 1% to 3-4% of regional peak demand. The growth is not linear but exponential, with a total energy needs increase from 17 terawatt-hours (TWh) to 57 TWh over the same period. These figures, reported by Wood Mackenzie, indicate that computing power is becoming the primary driver of structural energy demand.
The very nature of artificial intelligence generates a continuous and unpredictable load profile that traditional renewable sources cannot satisfy with the necessary reliability. Servers require power 24 hours a day, 7 days a week. The intermittency of solar and wind, combined with the lack of technological maturity in industrial-scale battery storage solutions in the region, creates an energy gap. This gap cannot be filled by future promises; it must be filled today by a dispatchable source, that is, capable of being turned on or off on demand to balance the grid.
Node Engineering: Combined Cycle Gas Turbines and LNG
The engineering response to this thermodynamic constraint is the installation of combined cycle gas turbines (CCGT). These power plants offer superior thermal efficiency compared to older coal technologies, but require a liquid or gaseous fuel that can be transported and stored. In Southeast Asia, the natural gas pipeline network is fragmented and insufficient to meet the impact of this new massive demand. Consequently, the market relies on Liquefied Natural Gas (LNG), which can be transported by ship and regasified locally.
The operating mechanism is clear: hyperscalers are building data centers near LNG regasification terminals or stable gas-powered electricity grids. The transition is not driven by ideological preferences but by the physical necessity of avoiding blackouts that would disrupt cloud services and data processing. LNG thus becomes the energy backbone of AI, replacing coal not for environmental reasons, but for operational reliability and supply flexibility compared to immature renewable alternatives.
Microeconomic Mapping: Who Bears the Costs?
This infrastructural transition is redefining regional trade balances. Countries exporting LNG are seeing a structural increase in demand, while local governments are finding themselves having to guarantee the stability of the electricity grid in order to attract technological investments. The construction of new combined cycle gas turbine (CCGT) power plants requires intensive capital and long lead times, creating bottlenecks in the availability of generation capacity by 2030.
Wood Mackenzie highlights how this dynamic is specific to Southeast Asia; Southern Asia, for example, remains excluded from this model due to the predominance of renewables in India’s electricity mix. This regional divergence underscores how energy geopolitics are not uniform but depend on local infrastructure and resource conditions. The costs of the transition fall on final energy consumers and investors in energy infrastructure, while the benefits in terms of digital competitiveness go to hyperscalers and nations that can provide stable power.
Trajectory and Structural Limit
The future trajectory of energy in Southeast Asia is now constrained by the availability of LNG. The structural limit is not technological, but logistical: import, storage, and regasification capacity must expand rapidly to support the 57 TWh projected. If renewable energy sources do not achieve battery maturity by 2030, natural gas will remain dominant in the region’s electricity mix.
For energy and logistics analysts, the key indicator to monitor is the actual growth of long-term LNG contracts in Southeast Asia, which will reflect real investments in data centers. The green transition has not been abandoned, but it has been subordinated to the operational need for artificial intelligence. The infrastructure cost of this choice will be measured in the continued dependence on gas imports and the resilience of regional electricity grids.
Photo by Brett Jordan on Unsplash
⎈ Content generated by multi-agent AI under Human-in-Command protocol
in an Epistemic Safety regime. Read the Operational Disclaimer.
> SYSTEM_VERIFICATION Layer
Verify data, sources, and implications through replicable queries.