Data Centers, Natural Gas, and the New Energy Geopolitics

Luis Augusto Medeiros Rutledge

Energy Geopolitics

When the growth of data centers began to materialize on a large scale, I conducted an analysis on the impacts this expansion would have on energy demand. At that time, much of the discussion was focused on the high water consumption associated with cooling systems. While this concern remains relevant, it has become increasingly evident that the main structural challenge will be ensuring sufficient, continuous, competitive, and reliable electricity to power a digital infrastructure that is growing at a rate far exceeding the traditional expansion of global electricity demand.

Today, we can better gauge the scale of this transformation. Relevant studies, particularly from the International Energy Agency, indicate that global electricity consumption by data centers is expected to nearly double between 2025 and 2030, rising from approximately 485 TWh to around 950 TWh, while consumption specifically tied to artificial intelligence data centers is projected to grow even faster. This means that digital infrastructure is becoming a new structural driver of global energy demand.

The accelerated expansion of data centers and artificial intelligence is therefore creating a specific sector with high energy demand, in which natural gas tends to occupy an increasingly strategic position. This occurs because digital infrastructures operate 24 hours a day and demand high levels of reliability, frequency stability, and power availability. Renewable sources such as solar and wind will be fundamental to meeting part of this growth, but their variability requires dispatchable generation, storage, and significantly more robust power grids.

In this scenario, natural gas also plays a structural role from an economic perspective by offering flexible, dispatchable generation capable of responding quickly to fluctuations between electricity supply and demand. This characteristic reduces outage risks, limits costs associated with grid instability, and contributes to the economic viability of energy-intensive projects like data centers. Technical studies project that gas-fired generation dedicated to serving this segment could grow by approximately 175 TWh by 2035, with an especially significant impact in markets where digital infrastructure expansion is occurring at a rapid pace, such as in the United States.

Currently, natural gas already accounts for about 26% of the electricity physically consumed by data centers worldwide and over 40% in the United States. This increase tends to produce ripple effects extending beyond the power sector, putting pressure on gas demand, spurring investment in generation, pipelines, and liquefied natural gas (LNG) terminals, and expanding the interaction among electricity, natural gas, and digital infrastructure markets. Project proposals designed to build pipelines exclusively to supply massive AI data centers are already materializing.

This movement introduces an important shift in the energy transition debate. The growth of renewables does not necessarily eliminate natural gas in the short and medium term. On the contrary: in power systems undergoing simultaneous expansion of solar and wind energy and accelerated growth of continuous digital loads, gas can assume the role of a safety and flexibility fuel for the new digital economy.

Undoubtedly, we can already conclude that the coming decade will witness the formation of a global market focused specifically on data centers. Moreover, this transformation is occurring precisely as the global natural gas market undergoes another historic shift. In past decades, gas trading depended mainly on pipelines directly connecting producers and consumers. Russia–Europe, Canada–United States, and North Africa–Europe were classic examples of this structure. Currently, the expansion of global liquefaction, maritime transport, and regasification capacity has profoundly altered this logic.

LNG has progressively transformed gas into an internationally tradable commodity, bringing its commercial dynamics closer to those observed in crude oil. A cargo produced in the United States, Qatar, or Africa can be directed to Europe, Asia, or Latin America depending on price differentials, terminal availability, logistical costs, and geopolitical conditions. This means that markets that were once relatively isolated are now competing for the exact same cargo.

An increase in electricity demand driven by data centers in the United States could, for example, raise domestic US gas consumption, impact internal prices, and alter the economic availability of LNG for export. Likewise, a cold spell in Europe or Asia can redirect LNG tankers, applying upward pressure on prices in other regions. Natural gas is thus shifting from a predominantly regional commodity to an interdependent global market subject to international arbitrage.

Data centers are consolidating as one of the primary drivers of global electricity demand growth, propelled by the rapid expansion of artificial intelligence, cloud computing, and digital services. These activities require high, reliable, and uninterrupted energy supply operating 24 hours a day, opening new prospects for the natural gas industry.

As the scale of these facilities grows, so does the need for power sources capable of guaranteeing security of supply, stability, and operational flexibility. In this context, natural gas reinforces its importance as a firm, dispatchable generation source capable of supporting power grids and responding rapidly to load variations while complementing the expansion of renewable energy sources. This combination tends to expand the strategic role of gas in the energy transition associated with the digitalization of the economy and the advancement of artificial intelligence.

Gas turbines. The accelerated expansion of data centers, driven mainly by artificial intelligence, is sparking a sharp resurgence in demand for gas turbines as a solution to secure large volumes of continuous, dispatchable power generation available on timelines compatible with the deployment speed of new digital infrastructure.

This movement is already beginning to transform the power generation equipment market. Global gas turbine orders reached approximately 51 GW in 2025, the highest level since 2000, while major manufacturers amass growing order backlogs and manufacturing capacity reservations. Equipment shortages, combined with long manufacturing lead times and the global concentration of large turbine manufacturing among a few suppliers, are starting to drive up the cost of building new plants and turning power equipment availability into one of the main bottlenecks facing data center expansion.

From an economic perspective, however, two distinct effects must be distinguished. The growth in turbine orders initially puts upward pressure on equipment prices, construction costs, and the capital investment required to expand generation capacity. In contrast, pressure on natural gas prices occurs later, as these new thermal power plants come online and structurally increase fuel consumption.

An economic transmission chain is thus established: more artificial intelligence generates greater demand for data centers; more data centers demand greater electrical capacity; the need for firm generation increases demand for gas turbines; shortages of these equipments elevate the cost of new power plants; and once this capacity comes online, natural gas consumption rises, sharpening competition for the fuel.

This process tends to be more acute in regions where electricity demand growth outpaces the expansion of power generation, transmission, and gas infrastructure. In these markets, simultaneous surges in data center loads and thermal power generation can put upward pressure on both electricity rates and regional natural gas prices, while also heightening the need for investments in pipelines, storage, and LNG infrastructure.

In the United States, this dynamic takes on an even larger scale. The country holds one of the world's highest concentrations of data centers, features a gas share exceeding 40% in the electricity currently consumed by these facilities, and is simultaneously expanding its LNG export capacity. The expansion of artificial intelligence can therefore introduce a new structural source of competition for American gas among power generation, industry, residential consumers, and LNG exports.

The true economic risk therefore lies in the mismatch between the speed of demand growth and the capacity for expanding energy supply. Infrastructure needs to keep pace with the advance of artificial intelligence so that the impact on electricity prices remains relatively limited. However, under grid and generation constraints, this impact could become substantially larger. In this scenario, gas turbines cease to be mere power generation equipment and become one of the main physical and economic bottlenecks in the global race for artificial intelligence infrastructure.

In conclusion, the expansion of artificial intelligence tends to deepen the connection among technology, energy, economics, and geopolitics. The growth of data centers increases demand for firm, continuous, and flexible electricity, and in this process, natural gas gains relevance not only as a power generation source but also as a strategic asset for energy security. Consequently, LNG is taking up an increasingly prominent position in foreign policies, supply strategies, and trade relations between producing and consuming nations and major tech hubs.

In the coming years, expanding AI capacity could bolster global demand for natural gas and LNG, especially in economies where electrical load growth outstrips grid expansion, renewable generation, energy storage, and other firm power sources. This dynamic may intensify international competition for gas molecules, liquefaction capacity, regasification terminals, LNG carriers, and long-term contracts, heightening importing countries' exposure to price swings, logistical constraints, and geopolitical tensions.

For Brazil, this transformation demands particular attention. Although the country possesses a predominantly renewable power mix and substantial natural gas reserves, limitations remain in the infrastructure for processing, transporting, evacuating, and integrating domestically produced gas. During periods of heightened thermal power generation needs, the Brazilian system continues to rely on imported LNG, leaving it exposed to international prices, exchange rates, freight costs, and competition from major consumers in Europe and Asia.

This picture could become even more critical as Brazil advances in deploying data centers and power-intensive digital infrastructure. The availability of competitive, reliable energy will increasingly become a pivotal factor in attracting tech investments. Thus, the conversation surrounding artificial intelligence ceases to be merely about computing power and directly involves energy policy, gas infrastructure, power transmission, security of supply, and industrial competitiveness.

From an economic perspective, excessive reliance on imported LNG during high-demand periods can translate into a direct transmission of international volatility into domestic power generation costs. Gas price shocks, logistical bottlenecks, or international disputes over LNG cargoes can drive up thermal power costs and consequently increase the Brazilian power system's exposure to rate spikes and inflationary pressures. On the other hand, greater integration among domestic gas production, transport infrastructure, and the power market could mitigate this vulnerability and expand the economic utilization of national resources.

A Brazilian paradox thus emerges. The country possesses significant advantages to participate in the new energy economy of artificial intelligence—an abundance of renewable resources, growing oil and gas production, expansion potential for generation, and a favorable geographic location for digital infrastructure deployment—yet it still needs to overcome bottlenecks in transmission, pipelines, processing, storage, and supply-demand integration.

In this new paradigm, energy security becomes digital security and economic competitiveness as well. Countries capable of offering abundant, stable, and competitive electricity will enjoy growing advantages in attracting data centers, AI investments, and tech-intensive industrial activities. Meanwhile, those overly dependent on imported energy will remain more vulnerable to geopolitical swings and international price cycles.

The new geopolitics of artificial intelligence will therefore also be a geopolitics of electricity and natural gas. For Brazil, the challenge will be to translate its wealth of energy resources into effective infrastructure, reducing external vulnerabilities and utilizing domestic gas to complement renewable expansion. Otherwise, the country risks entering the age of artificial intelligence bound by a structural contradiction: possessing abundant energy resources while remaining exposed to the international market precisely when it most needs firm generation and security of supply.

Reference

IEA – Agência Internacional de Energia

Luis Augusto Medeiros Rutledge is a Petroleum Engineer and Energy Geopolitics Analyst. He holds an Executive MBA in Petroleum and Gas Economics from the Federal University of Rio de Janeiro (UFRJ) and a postgraduate degree in International Relations and Diplomacy from IBMEC. He serves as CEO of BFI Republic, researcher at UFRJ, Consulting Member of the Islamic World Observatory of Portugal, and author of numerous published articles on Geopolitics and Energy.