What happened?
The artificial intelligence revolution that has been reshaping technology, business, and society is running into a fundamental physical constraint: electricity. AI data centres — the computational infrastructure that powers large language models, image generators, and AI services — are extraordinarily energy hungry. The Iran war's disruption to global energy markets has now added a new dimension to this challenge, raising costs and creating uncertainties for Big Tech companies investing in AI infrastructure.
Key Points
- A single large AI model training run can consume as much electricity as 100 homes use in a year
- Global AI data centre power demand is expected to triple by 2028
- The Iran war has raised electricity costs globally — coal, gas, and oil all impacted
- Microsoft, Google, Amazon, and Meta have all flagged rising energy costs as AI investment headwinds
- Indian data centre expansion plans are also affected — power costs and availability are key issues
- Renewable energy is the long-term answer, but short-term supply is constrained
Background
Artificial intelligence, particularly the large language models (LLMs) like those powering ChatGPT, Claude, and Gemini, requires enormous amounts of computational power. That computation requires electricity — vast amounts of it. Training a single large AI model can consume millions of kilowatt-hours of electricity. Running that model for millions of users adds continuous power demand.
The world's major technology companies — Microsoft, Google, Amazon, Meta — have been on a massive spending spree building AI data centres. They have been signing long-term power purchase agreements, building nuclear power plants, and even reviving coal plants in some cases — all to feed the electricity appetite of AI.
Main Details
The Iran war has disrupted global energy markets significantly. Natural gas prices have surged. Coal prices have risen. Even electricity generated from non-fossil sources has become more expensive in countries with integrated energy markets where wholesale prices track fossil fuel costs.
For Big Tech companies building AI data centres, higher energy costs directly affect the economics of their AI businesses. Microsoft has flagged energy as a key variable in its AI capital expenditure plans. Google has reported rising data centre operating costs. Amazon Web Services has slowed certain data centre expansion projects pending clarity on power availability and cost.
In India, where data centre construction has been booming, energy constraints are a serious challenge. Mumbai and Chennai are the primary Indian data centre markets. Both face power supply constraints at peak demand, and electricity costs have risen due to the Iran war's impact on India's energy mix.
Reactions
Technology analysts say the AI boom is not in danger of reversing — the economic value AI creates for businesses is too large. But the timeline may slow, and the economics of AI services may shift — potentially raising prices for AI tools and services to cover higher energy costs.
Environmental advocates have used this moment to argue more loudly for renewable energy as the solution — solar and wind power are now cheaper per unit in many markets than fossil fuel power, and they are not subject to geopolitical disruption.
Impact Analysis
For India specifically, the AI energy challenge creates both a risk and an opportunity. The risk: Indian AI ambitions — from data localisation to domestic AI models — face power supply constraints. The opportunity: if India accelerates renewable energy development for data centres (as several companies are attempting), it could position itself as a clean-energy AI hub for the global market.
What Happens Next
Energy will remain the defining constraint for AI expansion over the next 3–5 years. Companies that solve their energy challenge — through nuclear, solar, wind, or efficiency improvements — will gain significant competitive advantage. The race to power AI is as important as the race to build it.
FAQ
Q: Why do AI data centres use so much electricity?
A: AI models require millions of processor calculations per second, generating enormous heat that must be cooled — all requiring significant power.
Q: How has the Iran war affected AI investment?
A: Higher global energy costs have raised AI data centre operating expenses, slowing some expansion plans.
Q: What is India's AI energy challenge?
A: India's data centre markets in Mumbai and Chennai face power supply constraints and rising electricity costs.
Q: Is the AI boom going to slow down?
A: Not reverse, but potentially slow — the economics of AI services may shift as energy costs rise.
Q: What is the solution to AI's energy problem?
A: Renewable energy — solar, wind, and nuclear — is the long-term answer being pursued by major tech companies.