Why Data Centers Are the New Oil: Energy, AI and Geopolitics
Key Points
- Global data center electricity consumption reached 565 terawatt-hours in 2026 — a 26% increase from 447 TWh in 2025 — with AI-optimized servers now accounting for 31% of all data center power, according to Gartner's June 2026 forecast
- Amazon, Google, Microsoft, and Meta are collectively spending $725 billion on AI infrastructure in 2026 alone, up 77% from $410 billion in 2025 — more than the GDP of Switzerland
- Google consumed 10.9 billion gallons of water in 2025 — a 34% increase and more than double its 2021 level — while Texas data center water use is projected to rise 714% by 2030
For most of the 20th century, oil determined which nations held power. Today, a different resource is quietly taking its place: electricity dedicated to data centers.
The world's largest technology companies are racing to build facilities capable of training and running advanced AI models. These data centers require enormous amounts of continuous power. In some regions, a single large AI campus now demands as much electricity as a mid-sized city — and they need it every second of every day.
This shift is accelerating rapidly. According to Gartner analyst Linglan Wang, worldwide data center power demand rose 27% in 2026 to reach 132 gigawatts, up from 104 gigawatts in 2025. By 2030, that figure is projected to hit 290 gigawatts — nearly three times today's level. "Surging demand for compute-intensive AI workloads is driving unprecedented data center power growth," Wang said in Gartner's June 2026 report, "while AI capacity is now constrained by power availability, making data center power security the new battleground for scaling."
The result is a new form of geopolitical competition. Nations that can deliver cheap, reliable, and abundant power for data centers are positioning themselves as the energy powers of the AI age.
Power Has Become the Real Constraint
For several years, the limiting factor in AI development was access to advanced chips. That is no longer the full story.
In 2025 and 2026, major technology firms repeatedly delayed or scaled back data center projects not because they could not secure enough chips, but because they could not secure enough electricity. Grid operators in parts of the United States, Ireland, the Netherlands, and Singapore have begun restricting new large-scale connections. Some local governments have imposed temporary freezes on data center approvals.
Ireland provides the starkest example. By 2024, data centers consumed 22% of the country's total electricity — a proportion that has since risen further. The Central Statistics Office Ireland confirmed that figure in its 2025 energy report.
The IEA's April 2026 report Key Questions on Energy and AI found that data center electricity demand soared 17% in 2025, while AI-focused data centers grew even faster — outpacing growth in global electricity demand of just 3%. Between 2020 and 2025, AI server power density increased eleven times. The IEA projects a further fourfold increase by 2027, meaning a single refrigerator-sized server rack could draw power equivalent to 65 households.
The $725 Billion Arms Race
The numbers coming out of Big Tech's 2026 capital expenditure plans are extraordinary in their scale.
Amazon, Google, Microsoft, and Meta are collectively spending approximately $725 billion on capital expenditures in 2026 — up 77% from the already record-breaking $410 billion deployed in 2025, according to earnings disclosures compiled by multiple financial analysts. For context, that exceeds the annual GDP of Switzerland and roughly equals four times the entire annual capital investment of the US energy sector.
The breakdown by company: Amazon leads at approximately $200 billion, Microsoft follows at roughly $190 billion, Google at $175–185 billion, and Meta at $115–135 billion. The overwhelming majority of this spending is going not into chips — which are constrained but available — but into electricity infrastructure and data center construction.
Nvidia is the clearest beneficiary of this spending cycle. Its data center revenue hit a record $75.2 billion in a single quarter in early 2026, up 92% year-over-year, driven entirely by hyperscaler demand for AI accelerators.
The Geopolitical Dimension
The concentration of data center capacity is creating new strategic dependencies between nations.
The United States currently leads in both AI development and data center construction, supported by large domestic energy resources and private investment. China is accelerating its own build-out as part of a broader push for technological self-reliance. The European Union faces a more difficult position: strong regulatory frameworks and ambitious climate targets are colliding with the urgent need for more power capacity.
Singapore's 2026 global risk outlook noted that "disruptive technology will bring new opportunities with AI applications, as well as intensify geopolitical competition for leadership in the technology of the future." Countries with surplus energy — particularly those with nuclear power, large hydroelectric resources, or rapidly expanding renewable generation — are seeing domestic data center capacity as a matter of national security.
Control over physical infrastructure that trains and runs advanced AI systems is becoming as critical as semiconductor manufacturing or critical mineral supply chains.
The Water Nobody Talks About
Data centers don't just consume electricity — they consume vast quantities of water.
Most data centers use evaporative cooling: hot air from servers is cooled by passing it over water, which then evaporates and carries the heat away. The scale of this process, across thousands of facilities worldwide, has become a significant water management challenge.
Google consumed 10.9 billion gallons of water in 2025 — a 34% increase from 2024 and more than double its 2021 level — according to the company's own 2026 Environmental Report. Amazon disclosed its first-ever absolute water figure in 2026: 2.5 billion gallons consumed in 2025. Microsoft has committed to zero-water evaporation cooling across new facilities, with its current fleet averaging 0.27 litres per kilowatt-hour of electricity processed.
A single ChatGPT conversation uses between 10 and 25 millilitres of water in comprehensive accounting, according to a University of California, Riverside study and Sam Altman's own June 2025 disclosure. That sounds small. But when billions of queries happen daily across ChatGPT, Gemini, Grok, Claude, and others, the aggregate becomes significant.
Texas is projected to see data center water consumption increase from 49 billion to 399 billion gallons by 2030 — a 714% increase — according to state infrastructure projections. And 83% of Texas's 341 data centers have not complied with mandatory water reporting requirements, according to a state audit released in 2026.
Environmental and Economic Trade-offs
The rapid expansion of data centers is creating visible tension with climate goals.
Even when powered partly by renewable energy, these facilities require always-available electricity. That often means natural gas plants or nuclear reactors running as backup. In some markets, data center growth is delaying the retirement of fossil fuel plants that would otherwise have closed.
At the same time, the economic incentives are powerful. Data centers create high-value construction jobs, long-term tax revenue, and demand for specialised engineering talent. For regions with excess power generation, they offer a way to monetise electricity that might otherwise go unused. In several US states and European countries, residents have begun questioning why technology companies receive priority access to power while consumer bills climb.
What Happens Next
Three developments will shape this competition over the next two to three years.
First, power availability will increasingly determine where the next generation of AI infrastructure is built. Companies will locate facilities not only near talent and fibre networks, but where electricity can be secured at scale and at predictable cost.
Second, governments will treat electricity as a durable strategic policy tool. Expect more direct state involvement in grid expansion, nuclear project acceleration, and long-term power purchase agreements designed specifically for AI technology firms.
Third, the countries and companies that solve the power bottleneck most effectively will gain a durable advantage in the AI race. Those that fail to expand generation and transmission capacity risk falling behind even if they lead in research or chip design.
Data centers will not replace oil in every respect. But in the emerging economy built around artificial intelligence, control of electricity is becoming the decisive strategic resource. The nations and companies that understand this shift earliest are already acting on it.
Sources: Gartner / IEA / Google Environmental Report and other international news outlets.
Disclaimer: This article was produced with AI assistance based on publicly available news sources. While we strive for accuracy, NewsOracle makes no warranty as to the completeness or accuracy of the information. Errors and omissions may occur. Readers should independently verify all information before acting on it. NewsOracle does not intend to defame any individual or organisation and accepts no liability for any loss or damage arising from reliance on this content. Content is for informational purposes only and does not constitute legal, financial, medical, or professional advice. All rights reserved. Unauthorised reproduction prohibited.
NewsOracle Editorial
The NewsOracle Tech Desk covers breaking technology news including AI, Apple, Google, Tesla, Meta, OpenAI and product launches.
Latest coverage: AI
