THE CONTEXT: The International Energy Agency (IEA) projects that global electricity demand from data centres could exceed 1,000 TWh by 2026—more than Japan’s current demand. In India, installed IT load has already crossed ~1,000 MW (2024) and is expected to triple by FY-2028. Cooling alone can consume 40-50 % of a facility’s power budget. With AI‐driven workloads and 5 G-enabled data localisation gathering pace, “greening” cooling is emerging as a core question of energy security, water security, and India’s Net-Zero-2070 pathway.
KEY TECHNICAL TERMS:
| Term | What it means | Why it matters |
|---|---|---|
| Power Usage Effectiveness (PUE) | A ratio that tells us how much total electricity a data centre draws compared with the electricity that actually reaches the IT equipment. Formula : PUE = Total Facility Power ÷ IT Equipment Power. • A perfect score is 1.0 (every watt feeds the servers; none goes to cooling, lights, pumps etc.). | Lower PUE → higher energy efficiency, smaller electricity bill, smaller carbon footprint. It is today’s most-quoted yard-stick for benchmarking data-centre efficiency. |
| Water Usage Effectiveness (WUE) | The water counterpart of PUE. Formula : WUE = Total Annual Site Water Use (litres) ÷ IT Equipment Energy Use (kWh). | Highlights hidden “water cost” of evaporative cooling towers and adiabatic systems—critical in water-stressed regions. |
| On-site Membrane Bioreactor (MBR) | A compact waste-water treatment plant that combines biological digestion with ultra-fine membrane filtration. It produces near-potable recycled water directly at the data-centre campus. | Enables “zero-fresh-water” cooling loops or landscape irrigation and cuts sewer-discharge fees—vital for water-neutral data-centre design. |
| Accelerated Depreciation (AD) | A tax policy tool that lets firms write off the value of specified equipment faster than under normal schedules. For example, 40 % per year instead of 15 %. | By front-loading tax savings, AD lowers the effective upfront cost of capital-intensive green gear—liquid-cooling rigs, heat-recovery chillers, photovoltaic panels—thereby improving pay-back periods. |
| Dielectric Fluids | Electrically non-conductive liquids (e.g., synthetic hydrocarbons, perfluoro-ethers, new plant-based esters) used to submerge servers in immersion-cooling tanks. They absorb heat yet prevent short-circuiting. | They allow 100 % direct contact cooling, eliminate server fans, and tolerate higher operating temperatures, boosting efficiency while cutting noise and dust. |
| Liquid Cooling (in data centres) | Any method that replaces or supplements air-flow with a liquid—water–glycol mixes in cold plates, or dielectric baths in immersion—to draw heat away from processors. | Removes 5–10× more heat per unit volume than air; supports dense AI/graphics racks (>50 kW), slashes fan power, and can deliver very low PUE values (~1.1). |
| Green Data-Centre Disclosure Norms | Proposed/voluntary rules that require operators to publish standardised metrics—PUE, WUE, renewable-energy share, e-waste volumes, carbon intensity, etc.—in annual ESG or sustainability reports. | Creates transparency for investors, regulators, and customers; enables green-finance pricing and peer benchmarking; nudges the industry toward measurable sustainability goals. |
TECHNICAL PRIMER: WHY LIQUID COOLING MATTERS
| Parameter | Conventional air cooling | Cold-plate (direct-to-chip) | Two-phase immersion |
|---|---|---|---|
| Heat transfer path | Air → heat exchanger | Coolant micro-channels on CPU/GPU | Dielectric fluid in sealed tank |
| Typical PUE* | 1.7–2.0 | 1.3–1.4 | 1.1–1.2 |
| Water consumption | 25–26 million L / MW / year | Nil–very low | Nil |
| Emission reduction vs air | — | 15–18 % | 18–21 % |
*PUE = Power Usage Effectiveness
-
- A landmark Microsoft–WSP study in Nature (2024) found that cold plates and immersion cut life cycle GHGs by 15-21 %, energy by 15-20 % and water by 31-52 %. ASHRAE’s updated “Liquid-Cooling Guidelines for Datacom Equipment Centres, 2nd Ed.” (2024) now treats liquid cooling as “mainstream” for >50 kW racks.
GLOBAL BENCHMARKS & EMERGING PRACTICE
-
- EU Code of Conduct for Data Centres-2023 mandates voluntary PUE < 1.3 for new builds and pushes Water Usage Effectiveness (WUE) disclosure.
- Singapore’s “Constrained DC” framework (2024) links new licences to MW-hour caps and ≥ 70 % renewable-power purchase.
- Nordic states couple fjord-water immersion cooling with district-heating sale, monetising waste heat.
DRIVERS IN THE INDIAN CONTEXT
-
- Digital Public Infrastructure (DPI): UPI processes ≈ 13 billion tx/month (Mar 2025); latency-sensitive nodes need edge DCs.
- Data localisation mandates under the Digital Personal Data Protection Act 2023.
- Climate commitments: India’s Updated NDC (2022) pledges a 45 % emission-intensity cut by 2030.
- Water stress: In water-scarce hubs such as Bengaluru, a 20 MW facility can draw 500+ million L / year, triggering local opposition.
EXISTING POLICY ARCHITECTURE:
| Layer | Instrument | Salient Provisions |
|---|---|---|
| National | Draft Data Centre Policy-2020 (MeitY) | Data-Centre Economic Zones, single-window clearance, Green DC Rating, indigenous liquid-coolant R&D. |
| Data-Centre Incentive Scheme (under formulation, outlay ₹ 12 000 crore) | 3–4 % CAPEX grant, RE-purchase waiver, real-estate support. | |
| India Cooling Action Plan-2019 links ICT cooling demand to HFC-phase-down roadmap | ||
| Regulatory | Bureau of Energy Efficiency (BEE) DC Guidelines-2019 | Voluntary PUE ≤ 1.5; calls for Liquid Cooling Pilot-Labs. |
| State | UP, TN, Karnataka DC Policies (2021–23) | Land subsidy 30 %, green-power banking, treated-sewage quotas. |
THE CHALLENGES:
-
- Grid-Embedded Carbon: Average grid-emission factor is still ~0.7 kg CO₂e/kWh; merely shifting to efficient cooling without RE uptake risks burden-shifting.
- Water–Energy Trade-off: AI-optimised job scheduling (WaterWise algorithm, 2025) shows carbon and water objectives can be mutually conflicting.
- Hazardous Dielectric Fluids: Per- and poly-fluoro-ether based coolants have undefined end-of-life protocols in India.
- Skill & Standards Gap: Only five ITI curricula currently cover “Liquid Cooling Technician”; no BIS standard for WUE/PUE reporting.
- Jevons Paradox: Efficiency gains may accelerate demand; IEA warns DC power could still double to 945 TWh by 2030 even with best-available tech
THE WAY FORWARD:
| Dimension | Action Agenda | Rationale / Expected Outcome |
|---|---|---|
| Regulatory Mandate | Notify a mandatory Green Data Centre Code, 2025 under the Energy Conservation Act with tiered PUE (<1.3 by 2027) and WUE (<0.4 L/kWh) targets; integrate into ECBC-2025 | Creates enforceable sustainability floor; links to Perform-Achieve-Trade scheme for tradable “DC-ESCO” certificates. |
| Renewable Integration | Make 24×7 RE Power Purchase Agreements compulsory for new hyperscale DCs by 2030; allow virtual net-metering inside Special Economic Zones | Tackles Scope-2 emissions; reduces curtailment risk for DISCOMs. |
| Water Neutrality | Mandate 100 % use of recycled/sea water or dry cooling for >10 MW facilities in stressed basins by 2027; incentivise onsite membrane bioreactors via 5-yr GST rebate | Aligns with Jal Jeevan goals; eases urban potable-water conflict. |
| Circular Heat Economy | Offer Accelerated Depreciation (40 %) for heat-recovery chillers feeding district heating / desalination; pilot in Mumbai-Navi Mumbai cluster | Monetises waste heat; mirrors Copenhagen harbour model. |
| Indigenous Tech & R&D | Launch PLI-Liquid Cooling & Dielectric Fluids (₹ 3 000 crore); partner with IISc-IITs to replace PFPE fluids with glycerol-based bio-coolants | Builds domestic supply chain; lowers import bill; mitigates PFAS risk. |
| Data Transparency | SEBI-aligned Green DC Disclosure Norms: compulsory reporting of PUE, WUE, RE-share and e-waste volumes in Annual Sustainability Report | Increases investor visibility; enables ESG-linked financing. |
| Capacity Building | Up-skill 50 000 technicians via National Skill Qualification Framework (NSQF)-level-5 “Advanced Thermal Management” course, utilising ₹ 200 crore from Skill India Digital platform | Addresses human-resource bottleneck, especially for Tier-2-city edge nodes. |
THE CONCLUSION:
Data centres are the “steel mills” of the digital economy, but without prudent cooling choices they could become the coal plants of the information age. Liquid-cooling paired with 24×7 renewables and closed-loop water systems offers a realistic 40-50 % footprint cut, yet policy architecture must close the standards, disclosure and incentive gaps fast. A Green Data Centre Mission can reconcile India’s trillion-dollar digital ambition with its climate-justice commitments.
UPSC PAST YEAR QUESTION:
Q. Describe the benefits of deriving electric energy from sunlight in contrast to the conventional energy generation. What are the initiatives offered by our government for this purpose? (2020)
MAINS PRACTICE QUESTION:
Q. Analyse the techno-economic and governance challenges posed by data-centre cooling in India. Suggest a comprehensive policy roadmap that aligns digital-infrastructure growth with the country’s climate and water-security goals.
SOURCE:
Spread the Word