New Charts Reveal the Surging Energy Demands of AI

Artificial intelligence is getting smarter, but it’s also getting hungrier. A new report from MIT Technology Review, illustrated with four key charts, reveals just how fast AI’s energy consumption is climbing—and the surprising sectors feeling the impact.

1. AI’s Energy Use Is Exploding

The first chart shows the staggering growth in data center energy use driven by AI workloads. Models like ChatGPT and Gemini require vast computational resources to train and run. With more companies deploying AI daily, the total energy footprint is skyrocketing—projected to hit 10 times today’s usage by the end of the decade if unchecked.

2. Water and Power Are Being Stretched Thin

To keep data centers cool, AI systems rely on both electricity and large amounts of water. A second chart highlights spikes in water usage at major AI facilities across the U.S., especially in regions already experiencing drought. In parallel, AI’s growing power demand is putting pressure on aging grids, prompting new urgency around infrastructure upgrades.

3. Big Tech’s Emissions Targets Are at Risk

The third chart focuses on carbon emissions. While companies like Microsoft, Google, and Amazon pledged aggressive climate goals, their AI growth is making those targets harder to meet. Emissions tied to data center operations are now rising again—reversing previous declines and forcing companies to rethink their sustainability strategies.

4. AI Could Soon Rival the Aviation Industry in Emissions

The final chart is the most dramatic: a projection that shows AI’s carbon footprint could rival the entire global airline industry by 2030. Without renewable energy sources or efficiency breakthroughs, AI may become one of the world’s fastest-growing emissions contributors.

Why This Matters

While AI holds transformative promise—from healthcare to education—it comes with real environmental costs. As AI scales up, energy efficiency, sustainable chip design, and green data center practices must scale with it. Policymakers, engineers, and the public need to ask hard questions: Can AI grow responsibly? And who’s accountable for its energy appetite?

Frequently Asked Questions

Q1: Why does AI use so much energy?
Training and running large AI models require millions—even billions—of computations per second. These processes demand powerful servers that consume electricity and need intensive cooling.

Q2: Which companies are most affected?
Major AI developers like Google, Microsoft, Meta, and OpenAI operate massive data centers. As they deploy more AI models, their energy usage increases, straining both infrastructure and climate commitments.

Q3: What can be done to reduce AI’s energy impact?
Solutions include improving model efficiency, using custom AI chips, adopting renewable energy, building greener data centers, and designing algorithms that require fewer resources.

AI is shaping the future—but if we’re not careful, it could also reshape the planet. The challenge now is not just how powerful AI becomes—but how sustainably we power it.

Sources MIT Technology Review