Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)
ReleaseInsight · On RadarGuardian→ original

IEA: Data Center Emissions to Rise 70% by 2030 as AI, Coal and Gas Drive Demand

The IEA warns that data centers will consume twice as much electricity by 2030, and their emissions will rise 70%. AI accounts for three quarters of the demand growth, and the US and China are covering it with gas and coal. The additional 350 TWh per year is more than the entire United Kingdom consumes in a year. As a result, energy has become as much of a bottleneck for AI companies as chips.

AI-processed from Guardian; edited by Hamidun News
IEA: Data Center Emissions to Rise 70% by 2030 as AI, Coal and Gas Drive Demand
Source: Guardian. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceGuardian ↗
Editor-in-Chief
Key metric #1In China, about half of the energy for AI data centers will come from coal by 2030
Key metric #2The US will get about 60% of its data center energy from gas, with plans for 180 GW of power plants
Key metric #3In the EU, about 70% of data center electricity will come from renewable sources

Talk that AI is constrained by chips needs an update: increasingly, it is constrained by the power outlet. In its annual report, the International Energy Agency (IEA) said that by 2030, data centers worldwide will consume twice as much electricity as they do today. Carbon dioxide emissions tied to the sector will rise by roughly 70%. The reason is simple: the US and China are building AI capacity faster than they can connect clean power, and they are closing the gap with fossil fuels.

How Much Energy AI Needs

The IEA estimates that three quarters of the growth in data center consumption is attributable to artificial intelligence. AI sites alone will consume an additional 350 TWh per year by 2030. That is more than the entire United Kingdom consumes in a year.

The agency notes that the industry is opaque, so it based its estimates on indirect data. The first is shipments of AI servers, which have grown 18-fold since 2020. The second is satellite imagery, which was used to track construction at 93 large "AI factories" around the world. The capacity of these facilities has grown tenfold over the past 24 months, reaching 17 GW. Once the factories are completed, their combined capacity is expected to reach about 50 GW.

The entire data center fleet, including servers with AI accelerators, is currently estimated at 120 GW, with 80% located in China, the US and the EU. Under the IEA's baseline scenario, another 120 GW will be added by the end of the decade, three quarters of it for AI.

Gas in the US, Coal in China, Green Power in Europe

The differences between regions are enormous. Data center developers have become the largest corporate buyers of renewable energy, but their appetite for fossil fuels is growing even faster.

  • China. By the end of the decade, about half of the electricity for AI data centers will come from coal. The rest will come mainly from renewable sources, with slightly less from nuclear and very little from gas.
  • United States. About 60% of data center electricity will come from gas-fired power plants. In August of this year, the IEA counted plans for 180 GW of new gas capacity tied to data centers. That is nearly the total gas capacity currently installed across the entire European Union.
  • Canada and the rest of the world. Canada plans 11 GW of gas generation for servers, which means US requests are more than 16 times higher. In the rest of the world, AI companies have announced only 4 GW of new gas plants.
  • EU. By 2030, about 70% of data center electricity will come from renewable sources, followed by nuclear, gas and a very small share of coal.

The overall picture is this: the more AI infrastructure is built in the US and China, the higher global emissions will be, even if Europe and other regions remain clean.

The Geopolitical Backdrop and What It Means for the Industry

The report arrives against the backdrop of a war in the Middle East, which the IEA described as the largest disruption to energy markets in history. The agency's director, Fatih Birol, said he sees no reliable path for energy to return to its previous state. Its shape will be determined by government decisions made after the crisis: a growing population, rising demand for air conditioning and an AI-driven economy are all pressing on grids at once.

For practitioners, several conclusions follow. First, the price and availability of electricity will increasingly affect the cost of training and inference, and therefore API pricing. Second, choosing a region for cloud capacity is becoming a question of carbon footprint as well as latency and data-protection laws: companies with climate commitments will face a conflict between speed and cleanliness. Third, model efficiency — quantization, small specialized models, caching — is shifting from an optimization into a direct energy saving.

There are caveats. This is a forecast, not a fact: it depends on how quickly governments and companies build grids, nuclear capacity and renewable generation. Industry data is closed, and the IEA's estimates rely on indirect signals. In addition, the 17 GW and 50 GW figures refer to a specific sample of 93 large sites, not the whole market. But the direction is clear: in the next phase of the AI race, the winners will be those who have not only chips and money, but also enough energy.

Story

Key Questions Answered

How Much Energy AI Needs?

The IEA estimates that three quarters of the growth in data center consumption is attributable to artificial intelligence. AI sites alone will consume an additional 350 TWh per year by 2030. That is more than the entire United Kingdom consumes in a year. The agency notes that the industry is opaque, so it based its estimates on indirect data. The first is shipments of AI servers, which have grown 18-fold since 2020.

What are the key details on Gas in the US, Coal in China, Green Power in Europe?

The differences between regions are enormous. Data center developers have become the largest corporate buyers of renewable energy, but their appetite for fossil fuels is growing even faster. China. By the end of the decade, about half of the electricity for AI data centers will come from coal. The rest will come mainly from renewable sources, with slightly less from nuclear and very little from gas. United States. About 60% of data center electricity will come from gas-fired power plants.

What are the key details on The Geopolitical Backdrop and What It Means for the Industry?

The report arrives against the backdrop of a war in the Middle East, which the IEA described as the largest disruption to energy markets in history. The agency's director, Fatih Birol, said he sees no reliable path for energy to return to its previous state. Its shape will be determined by government decisions made after the crisis: a growing population, rising demand for air conditioning and an AI-driven economy are all pressing on grids at once. For practitioners, several conclusions follow.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Private AI Builders Community · Vibe Club

Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews

Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.

Join Vibe Club →Explore Hamidun Academy
What do you think?
Loading comments…