Telangana, Maharashtra, U.P. lead in data centre policy attractiveness: S&P Global

1. At a Glance

• S&P Global Energy has released a Data Centre Policy Attractiveness India (DCPAI) index. It scores the incentive packages that Indian states offer to data centres. Telangana is ranked as the most attractive [1].
• Maharashtra and Uttar Pradesh are noted for balanced frameworks. Gujarat offers the most lucrative electricity and ease-of-doing-business measures, aimed at hyperscalers. All four leading states score around 50/100 [1].
• Why it matters for UPSC: data centres underpin cloud, AI and digital public infrastructure. The topic links competitive federalism, industrial policy, energy demand and data governance (GS-II and GS-III).
• Central policy context: the draft National Data Centre & Cloud Policy (MeitY) [2], the TRAI recommendations on data centres [4], and the Union Budget 2026–27 tax holiday up to 2047 for cloud services provided from Indian data centres [5].

2. Why in the News

• The DCPAI index appears in S&P Global's report "India Forward: Reimagining Growth", covered in The Hindu (print edition of 3 October 2026, p. 11, PTI) [1].
• It ranks states on how attractive their data centre incentive packages are, and groups them into score clusters [1].
• Backdrop: the Union Budget 2026–27 announced measures to make India a "global hub for cloud and AI infrastructure" [5].

3. Background & Evolution

• Rationale: to grow India's domestic data centre capacity, promote indigenous platforms and solutions in data centre and cloud operations, and encourage energy-efficient, sustainable, green data centres [2].
• Milestones (chronological):
• 2016 onwards: states begin enacting IT, ITeS, data centre and GCC policies. These are the policies the DCPAI index evaluates [1].
• Draft National Data Centre & Cloud Policy: MeitY held an industry consultation on it. PIB material on the draft cites a target of 2,000 MW of capacity addition by 2027 [S2, S3].
• PIB document "India to be a Cloud Computing and Data Centre hub" (December 2022 upload) [3].
• TRAI recommendations on a "Regulatory Framework for Promoting Data Economy Through Establishment of Data Centres, Content Delivery Networks, and Interconnect Exchanges in India". TRAI proposed a Data Centre Incentivization Scheme (DCIS) [4].
• Union Budget 2026–27: a tax holiday up to 2047 for foreign companies that provide cloud services globally using data centre services located in India [5].
• October 2026: S&P Global's DCPAI index is published [1].

4. Core Static Facts

Item	Fact
Index name	Data Centre Policy Attractiveness India (DCPAI) [1]
Publisher	S&P Global Energy (S&P Global), a private agency, not a government body [1]
Report	"India Forward: Reimagining Growth" [1]
Coverage	19 state policies across 11 states, enacted since 2016 [1]
Policy types assessed	IT, ITeS, data centre and Global Capability Centre (GCC) policies [1]
Four incentive categories	(1) Electricity; (2) Land and capital; (3) Operational and technical support; (4) Regulatory and EoDB frameworks [1]
Cluster 1 (~50/100)	Telangana (top; strong electricity and regulatory incentives), Maharashtra, Uttar Pradesh (balanced packages), Gujarat (best electricity and EoDB measures, aimed at hyperscalers) [1]
Cluster 2 (42–45)	Odisha, Haryana, Andhra Pradesh, Karnataka [1]
Cluster 2 notes	Andhra Pradesh leads on land and capital support. Karnataka has notable EoDB measures, including an Essential Services Maintenance Act (ESMA) cushion. Odisha and Haryana offer incentives across the whole project life cycle [1]
Cluster 3	Scores from about 30 upward; the upper bound and member states were cut off in the available excerpt [1]
Central nodal ministry	MeitY (draft National Data Centre & Cloud Policy) [2]
Draft policy target	2,000 MW capacity addition by 2027 [S2, S3]
TRAI proposal	DCIS with two lists: (a) central fiscal and non-fiscal incentives; (b) guidelines that give states flexibility to set their own fiscal incentives [4]
Budget 2026–27	Tax holiday up to 2047 for foreign cloud-service companies using data centres located in India [5]

5. Multi-Dimensional Analysis

Economic

• Incentives on electricity, land and capital target the largest cost items for a data centre. Power cost is the main reason Gujarat's package appeals to hyperscalers [1].
• The index covers GCC and ITeS policies too, so data centres are being folded into the wider services-led investment story [1].
• The central tax holiday to 2047 aims to make India an export base for global cloud services [5].

Governance / Federalism (Competitive Federalism)

• With 11 states and 19 policies, states are competing for investment. The index makes their incentive packages directly comparable [1].
• TRAI's two-list DCIS model divides the work: the Centre provides baseline incentives, and states have flexibility on fiscal incentives. This is a cooperative-federalism template [4].
• Karnataka's ESMA cushion shows states using labour and essential-services law as a tool to promise uninterrupted operations [1].

Environmental / Energy

• The draft national policy stresses energy-efficient, sustainable, green data centres [2].
• Analysis: electricity is the dominant incentive category, so state packages could lock in high power demand. That creates tension with renewable-energy and water-stress goals.

Scientific / Technological

• The draft policy promotes indigenous platforms and solutions in data centre and cloud operations [2].
• The Budget 2026–27 framing ties data centres to AI infrastructure capacity [5].

Strategic / Data Sovereignty

• Analysis: domestic data centre capacity supports data localisation, compliance with the Digital Personal Data Protection regime, and resilience of critical digital infrastructure.

Administrative

• Single-window and EoDB measures form a separate scoring pillar, which shows that clearances are a known bottleneck [1].
• Policies in some states are fragmented across IT, ITeS, data centre and GCC instruments (19 policies across 11 states) [1].

6. Recent Developments (last 12–18 months)

• Union Budget 2026–27: tax holiday up to 2047 for foreign companies providing cloud services globally from data centres in India [5].
• 3 October 2026 (print): S&P Global's DCPAI index is reported. Telangana is first, and Telangana, Maharashtra, Uttar Pradesh and Gujarat all score around 50/100 [1].
• Second cluster (42–45): Odisha, Haryana, Andhra Pradesh, Karnataka [1].

7. Prelims Hooks

• The DCPAI index is published by S&P Global Energy, not by NITI Aayog or MeitY [1].
• It appears in the report "India Forward: Reimagining Growth" [1].
• Telangana has the most attractive data centre policy package [1].
• Gujarat has the most lucrative electricity and EoDB measures, aimed at hyperscalers [1].
• The top four (Telangana, Maharashtra, U.P., Gujarat) each score about 50 out of 100 [1].
• Coverage: 19 policies, 11 states, enacted since 2016 [1].
• Four pillars: electricity; land and capital; operational and technical support; regulatory/EoDB [1].
• Andhra Pradesh leads on land and capital support [1].
• Karnataka offers an ESMA cushion for data centres [1].
• Second cluster scores 42–45: Odisha, Haryana, A.P., Karnataka [1].
• Draft National Data Centre policy target: 2,000 MW by 2027 [S2, S3].
• TRAI recommended a Data Centre Incentivization Scheme (DCIS) for data centres and Data Centre Parks [4].
• Budget 2026–27: tax holiday up to 2047 for foreign companies serving global cloud customers from Indian data centres [5].

8. What the Index Rewards, and What It Leaves Out

• The index measures what a state offers, not what it achieves
• It scores incentive packages under four pillars: electricity, land and capital, operational support, and regulatory/EoDB (Ease of Doing Business) measures [1].
• It does not ask how much capacity was actually built, how many jobs came, or how much state revenue was given up. (Analysis)

• So a state can rank high just by promising more.

• A bigger subsidy means a better score

• Electricity is a full pillar, and Gujarat is praised for the most generous power measures for hyperscalers (very large cloud companies such as global tech firms) [1].

• When the ranking rewards cheaper power and cheaper land, each state is pushed to give away more than its neighbour. (Analysis)

• No pillar for water or green power

• None of the four pillars scores sustainability [1].
• Yet the draft national policy says it wants energy-efficient, green data centres [2].

• So the index and the national goal pull in different directions.

• Even the best states score only about 50/100

• The top four all sit around 50 [1].
• This means no state yet offers a complete package. The best state covers only about half of what the index looks for.

9. Where Data Centres Are Actually Being Built

• Investment goes to a few cities, not spread across many states
• India's installed data centre capacity was 1.57 GW (gigawatt, a unit of power) in August 2026 [7].

• Mumbai and Navi Mumbai alone hold over 25% of live capacity. Chennai holds about 13% [7].

• A good policy ranking does not decide where companies build

• Tamil Nadu (Chennai) is not in the index's top two clusters [1], yet Chennai is the second-largest hub [7].

• So existing networks, demand and a city's head start matter as much as state incentives. (Analysis)

• What this means for balanced regional development

• If capital follows the hubs that already exist, generous incentives in newer states may just raise the cost of each project without moving it there. (Analysis)

• The growth is fast, so today's choices last a long time

• Capacity grew from about 375 MW in 2020 to about 1,500 MW in 2025 [6].
• It is projected to reach nearly 8 GW by 2030 [7].
• Plants built under today's incentive rules will set power and water demand for decades. (Analysis)

10. The Water Cost No State Policy Asks About

• Data centres need a lot of water for cooling
• A single 20 MW data centre uses roughly 1.4 million litres of water a day [6].

• Data centres across India use about 150 billion litres a year, and this is expected to more than double by 2030 [6].

• Many are being built where water is already short

• Among Asian countries, India is second only to Indonesia in data centres located in "extremely high" water-stress zones [6].

• Bengaluru needs up to 3,000 MLD (million litres a day) but gets only 2,000 MLD from the Cauvery. About 6,900 government borewells have dried up [6].

• No one is required to report water use

• None of the major state data centre policies require operators to disclose water use through a central reporting system [6].

• Without data, a state cannot tell whether a new project will compete with households for water.

• Karnataka's ESMA cushion makes this sharper

• ESMA (Essential Services Maintenance Act) lets the state treat a service as essential so that strikes cannot stop it. Karnataka offers this cushion to data centres [1].
• If data centres are treated as essential, their water and power supply may get priority in a shortage, possibly ahead of local users. (Analysis)

11. Is Competition Between States a Good Thing Here?

• The case for competition (strongest version)
• States that compete must clear files faster. That is why EoDB is a separate pillar and Karnataka offers measures like the ESMA cushion [1].
• Data centres are now listed as infrastructure, which helps them get loans more easily [7]. States that make land and power easy help India reach the nearly 8 GW target sooner [7].

• This is the logic behind competitive federalism, and it is partly right: competition has pushed states to write clear policies. There are 19 of them since 2016 [1].

• Where the argument breaks

• The Centre already gives a tax holiday up to 2047 [5]. When states add their own power and land subsidies on top, the public pays twice for the same project. (Analysis)

• Competition on speed of clearance is useful. Competition on who gives away the cheapest power and water is not, because the cost falls on the state's grid and local water supply [6].

• Honest verdict

• Competition should continue, but on how fast and how clearly states work, not on who gives the biggest subsidy. (Analysis)

12. Who Should Fix What

• MeitY: make the final national policy require water and power disclosure
• The draft policy already aims for green data centres [2].

• No state policy requires water reporting today [6]. A national rule would fill that gap for every state at once.

• The Centre: set conditions on state incentives using TRAI's two-list model

• TRAI's DCIS (Data Centre Incentivization Scheme) has a second list of guidelines within which states set their own fiscal incentives [4].

• The Centre can use this list to set basic conditions, for example that a state may subsidise power only if the operator reports water use and efficiency. (Analysis)

• States: link power subsidies to renewable energy

• Electricity is the biggest pillar [1]. Tying cheap power to green power would turn the main incentive towards the national green goal [2].

• Index makers and policy watchers: add a sustainability pillar

• If water and energy efficiency were scored, states would compete on them too. (Analysis)

13. Anchors for Answers

• Data: India's installed data centre capacity was 1.57 GW in August 2026, projected to reach nearly 8 GW by 2030 [7]
• Data: Capacity grew from about 375 MW (2020) to about 1,500 MW (2025) [6]
• Data: One 20 MW data centre uses about 1.4 million litres of water a day. India's data centres use about 150 billion litres a year [6]
• Data: Mumbai and Navi Mumbai hold over 25% of live capacity, Chennai about 13% [7]
• Report/Committee: TRAI recommendations on a regulatory framework for data centres, CDNs and interconnect exchanges, which proposed the DCIS with a Centre–state two-list design [4]
• Law/Case: Essential Services Maintenance Act, used by Karnataka as a cushion for data centres [1]
• Scheme: Data centres were added to the Harmonized List of Infrastructure in Union Budget 2022–23, which gave them infrastructure status and easier credit [7]. Budget 2026–27 added a tax holiday up to 2047 for cloud services provided from Indian data centres [5]

14. Mains Relevance

• GS-III: Indian economy and investment models; infrastructure (energy); science and technology (IT, computers); effects of liberalisation and industrial policy.
• GS-II: federalism (competitive and cooperative); government policies and interventions.
• Model questions: 1. "Data centres are the new industrial infrastructure of the digital economy." Examine how state-level incentive competition shapes India's data centre landscape, and its implications for balanced regional development. (250 words) 2. Discuss the energy and environmental challenges of India's rapid data centre expansion. How can policy reconcile hyperscaler demand with India's climate commitments? (150 words) 3. Evaluate the role of competitive federalism in attracting digital infrastructure investment, with reference to recent rankings of state data centre policies. (150 words)

15. Related Topics to Study Next

• Digital Personal Data Protection Act, 2023: shapes where data is stored and processed (data localisation).
• IndiaAI Mission and GPU compute infrastructure: AI demand is a major driver of data centre capacity.
• Global Capability Centres (GCCs): assessed in the same index; a key services-export theme.
• TRAI's role beyond telecom: its data centre, CDN and interconnect-exchange recommendations [4].
• Green energy open access and RE-RTC power: electricity is the top incentive pillar.
• Competitive federalism indices (BRAP/EoDB rankings, Export Preparedness Index): similar logic of ranking states.
• Essential Services Maintenance Act: its use to classify data centres as essential services.
• Union Budget 2026–27 tax measures for cloud and AI infrastructure [5].

16. Common Errors / Trap Areas

• Wrong publisher: the DCPAI index is from S&P Global Energy, a private agency, not NITI Aayog, MeitY or TRAI [1].
• Telangana vs Gujarat: Telangana is the most attractive overall. Gujarat is best only on electricity and EoDB [1].
• Andhra Pradesh vs Telangana: A.P. is in the second cluster but leads on land and capital [1].
• Counting: the index covers 19 policies across 11 states, not 19 states [1].
• TRAI vs MeitY: DCIS was proposed by TRAI [4]. The draft National Data Centre & Cloud Policy is MeitY's [2]. The tax holiday to 2047 is a Budget 2026–27 measure [5].