“Can India develop a large AI and data-centre economy without creating unacceptable pressure on electricity, water and the environment?”

 What is a Data Centre?

A data centre is a specialised facility containing servers, storage systems, networking equipment, power systems and cooling systems that store and process enormous quantities of digital information.

A simple chain is:

User → Internet → Data Centre → Computing/Storage → Response → User

For example, when you use an AI service, search engine, cloud storage or digital banking, the required computing may happen inside a data centre.

2. Why are Data Centres expanding so rapidly?

AI is now a major driver.

Globally, the IEA projects data-centre electricity consumption to more than double to around 945 TWh by 2030. AI-oriented computing is growing particularly rapidly. �

IEA +1

India is also expanding quickly. Government data says India's data-centre capacity increased from approximately 375 MW in 2020 to around 1,500 MW in 2025. �

Press Information Bureau

One 2026 Wood Mackenzie projection is even more ambitious, forecasting India's operational capacity could reach 12 GW by 2030. �

Wood Mackenzie

3. 🌱 GOOD IMPACTS

A. Digital transformation

Data centres make possible:

AI and machine learning

Digital education

Telemedicine

Digital banking and UPI

E-governance

Cloud computing

Online businesses

Scientific research

Satellite and weather-data processing

B. Economic development

They attract large investments in:

Land → buildings → electricity → telecom → cooling → hardware → software → skilled employment

India is positioning itself as an important digital and AI infrastructure hub.

C. Research and innovation

This is particularly interesting for education.

Universities, startups and students can obtain access to powerful computing without purchasing their own expensive supercomputers.

India's AI compute programme, for example, had onboarded about 38,231 GPUs through 14 empanelled providers/data centres, according to the Government of India. �

Press Information Bureau

D. Better utilisation of renewable energy

Data centres can potentially become large customers for:

☀️ Solar

🌬️ Wind

💧 Hydro

⚛️ Nuclear

🔋 Battery storage

The IEA expects renewables to provide a substantial share of the additional electricity needed for global data-centre growth. �

IEA

4. ⚠️ BAD / CHALLENGING IMPACTS

This is where our research becomes really interesting.

① Electricity consumption

Data centres operate 24 × 7.

Servers produce heat → cooling is required → cooling requires electricity → more electricity is required.

So:

More computing → more electricity → more heat → more cooling → additional electricity

This creates an important sustainability challenge.

② Water consumption

Some cooling systems consume significant quantities of water.

This becomes particularly important when a data centre is located in a water-stressed region.

A 2026 CEEW analysis estimates India's data centres used approximately 150 billion litres of water annually in 2025, with demand expected to more than double by 2030. �

CEEW

However, this figure should be interpreted carefully because water consumption varies enormously depending on cooling technology, climate and electricity source.

India's government says newer approaches such as direct-to-chip liquid cooling, immersion cooling and adiabatic cooling are being adopted to improve water efficiency. �

Press Information Bureau

③ Carbon emissions

If the electricity comes from coal or other fossil fuels:

Data centre → electricity demand → fossil-fuel generation → CO₂ emissions

Therefore, the environmental impact depends not only on the data centre's efficiency but also on where its electricity comes from.

④ Local environmental pressure

Large facilities can require:

considerable land

transmission infrastructure

backup power systems

cooling infrastructure

water supply

construction materials

Therefore, a data centre may have relatively small global impacts compared with the whole economy, but large local impacts if concentrated in one area.

⑤ E-waste

Servers and electronic equipment eventually become obsolete.

This produces:

Servers → replacement → electronic waste → recycling/disposal challenge

Responsible recycling and recovery of metals therefore become important.

Global data-centre distribution

The 2026 Stanford AI Index, using Cloudscene data for 2025, lists 11,700+ operational data centres worldwide. The United States dominates with 5,427 facilities, followed by Germany, UK and China. The report explicitly cautions that a facility count does not indicate its size or computing capacity. �

Stanford HAI +1

Rank

Country

Centres

1

🇺🇸 USA

5,427

2

🇩🇪 Germany

529

3

🇬🇧 UK

523

4

🇨🇳 China

449

5

🇨🇦 Canada

337

6

🇫🇷 France

322

7

🇦🇺 Australia

314

8

🇳🇱 Netherlands

298

9

🇷🇺 Russia

251

10

🇯🇵 Japan

222

11

🇧🇷 Brazil

197

12

🇲🇽 Mexico

173

13

🇮🇹 Italy

168

14

🇮🇳 India

153

15

🇵🇱 Poland

144

2. 🌍 Global electricity

The IEA estimates that data centres consumed approximately 415 TWh of electricity in 2024, about 1.5% of global electricity consumption. Its base case projects approximately 945 TWh by 2030. �

IEA +1

If we use:

11,700 centres + 415 TWh/year

then the simple arithmetic average is:

≈35.5 GWh/year per centre

or approximately:

3. 🔥 Heat generation

 The hypothetical global-average centre is approximately:

4 MW of continuous heat generation

Annual heat energy:

≈35.5 GWh/year

or approximately:

128 TJ/year

That is why cooling is such a fundamental part of data-centre engineering.

The IEA notes that cooling can represent around 7% of electricity consumption in efficient hyperscale facilities but over 30% in less-efficient enterprise facilities. �

4. 💧 Water calculation

Here we need to be particularly careful.

Water consumption cannot reliably be calculated simply from the number of centres.

The industry uses:

WUE = Water Usage Effectiveness

measured in:

litres/kWh

As an illustrative scenario, let's use 0.27 L/kWh, approximately the FY2025 global WUE reported by Microsoft for its data-centre.

5. Country-wise MODEL

hypothetical 35.5 GWh/centre/year to every country. This is deliberately a normalised model, not a claim that each country's centres actually consume this amount.

Country

Centres

Model electricity/year

Model continuous load

Model heat

Water @0.27 L/kWh

🇺🇸 USA

5,427

192.7 TWh

22.0 GW

22.0 GW

52.0 billion L

🇩🇪 Germany

529

18.8 TWh

2.14 GW

2.14 GW

5.07 B L

🇬🇧 UK

523

18.6 TWh

2.12 GW

2.12 GW

5.01 B L

🇨🇳 China

449

15.9 TWh

1.82 GW

1.82 GW

4.30 B L

🇨🇦 Canada

337

12.0 TWh

1.37 GW

1.37 GW

3.23 B L

🇫🇷 France

322

11.4 TWh

1.30 GW

1.30 GW

3.09 B L

🇦🇺 Australia

314

11.1 TWh

1.27 GW

1.27 GW

3.01 B L

🇳🇱 Netherlands

298

10.6 TWh

1.21 GW

1.21 GW

2.86 B L

🇷🇺 Russia

251

8.91 TWh

1.02 GW

1.02 GW

2.41 B L

🇯🇵 Japan

222

7.88 TWh

0.90 GW

0.90 GW

2.13 B L

🇧🇷 Brazil

197

6.99 TWh

0.80 GW

0.80 GW

1.89 B L

🇲🇽 Mexico

173

6.14 TWh

0.70 GW

0.70 GW

1.66 B L

🇮🇹 Italy

168

5.96 TWh

0.68 GW

0.68 GW

1.61 B L

🇮🇳 India

153

5.43 TWh

0.62 GW

0.62 GW

1.47 B L

🇵🇱 Poland

144

5.11 TWh

0.58 GW

0.58 GW

1.38 B L

The IEA gives actual regional electricity consumption. In 2024:

🇺🇸 USA = 45%

🇨🇳 China = 25%

🇪🇺 Europe = 15%

of global data-centre electricity consumption. �

IEA

Therefore:

USA ≈ 186.8 TWh/year

China ≈ 103.8 TWh/year

Europe ≈ 62.3 TWh/year