India’s ‘7GW by 2030’: Forecast, Pipeline, or Live Capacity?

A 2030 forecast is not today’s operating inventory; label the date, scope and capacity basis.

Gridvik Research · Public-source check through 12 October 2026 (IST) · Long-form market guide

Quick answer: The clearest “7GW” figure in the supplied July 2026 KPMG report is a forecast: KPMG projects about 7.0–7.5 GW by FY30. It is not India’s live capacity today, and it is not the same thing as KPMG’s separate ~4.5 GW development pipeline. Around H1/FY26, published operating or installed estimates sit near 1.8–1.9 GW, depending on source and basis. There is an important wrinkle: a July 2026 media report attributed an “over 7 GW IT” 2030 forecast to Savills, while Savills’ official H1 2026 report says 6 GW IT and its city table totals 6,065 MW IT. Those public statements are not reconciled.

Bottom line: “7GW by 2030” can describe a forward-looking estimate in a named report. It must not be presented as today’s operating inventory, as a verified project pipeline, or as a settled industry consensus. State the source, publication date, forecast period, geography, stage and capacity basis every time.

Scope note: This article uses KPMG’s July 2026 report as its main source, alongside Savills’ original H1 2026 report, Cushman & Wakefield’s H1 2026 update and other published research. These are provider estimates, not one harmonised government census. Where a source leaves the IT-load versus facility-power basis unclear, the article says so rather than inferring one.

In this guide

1. The answer in one minute
2. What KPMG’s 7–7.5 GW figure actually is
3. Three scoreboards: operating stock, pipeline and forecast
4. What the H1 2026 operational baseline says
5. The Savills 6 GW versus “over 7 GW” discrepancy
6. Other 2030 estimates—and why they are not one range
7. Savills’ city-by-city 2030 projection
8. Why IT load, facility power and grid capacity matter
9. How to read a data-centre pipeline
10. What would have to happen for a forecast to become operating stock?
11. A reusable checklist for testing a 7GW headline
12. What the numbers mean for investors, operators and suppliers
13. Glossary: capacity, stages and forecast language
14. Frequently asked questions
15. Sources and further reading
16. Related Gridvik guides

The answer in one minute

There are three different questions hiding inside the phrase “India’s data-centre capacity will reach 7GW by 2030.”

• What is operating now? Savills puts H1 2026 operator-based stock at 1,819 MW IT. Cushman & Wakefield (C&W) reports 1,789 MW of operational capacity in its H1 2026 update. KPMG gives a rounded ~1.9 GW installed-capacity estimate for FY26. These are nearby estimates, but their dates, scopes and capacity labels are not identical.
• What is in the pipeline? KPMG says a development pipeline is expected to add about 4.5 GW over the next five years. C&W reports 3,860 MW of upcoming supply through 2030, split into 509 MW under construction and 3,351 MW planned. The reports do not provide a project-level crosswalk showing that these are the same pipeline.
• What is forecast? KPMG projects 7.0–7.5 GW by FY30. This is a forward projection, not an operating stock count. Savills’ detailed H1 report projects roughly 6 GW IT by 2030, while an earlier July news report attributes over 7 GW IT to Savills. The discrepancy has not been explained in the public material reviewed.

The key is not to pick the largest or smallest number and call it “the answer.” It is to identify which source is answering which question. “Operational,” “installed,” “pipeline,” “planned,” “under construction” and “projected” are not interchangeable status words.

What KPMG’s 7–7.5 GW figure actually is

The supplied KPMG report, India data centre opportunity: From emerging demand hub to integrated data centre powerhouse is dated 1 July 2026; see also the official KPMG insight page. It presents India’s data-centre market as moving from about 1.9 GW of installed capacity in FY26 toward a 7.0–7.5 GW projected-capacity figure for FY30. In accompanying text, KPMG says the country’s installed capacity has more than tripled since FY19 and that a development pipeline is expected to add roughly 4.5 GW over the next five years.

Those three KPMG numbers answer different questions:

1. ~1.9 GW in FY26: a reported installed-capacity baseline for the report’s current period.
2. ~4.5 GW: expected additions associated with a development pipeline over a five-year window.
3. 7.0–7.5 GW in FY30: the forward projection / target-period capacity in KPMG’s outlook.

The 7.0–7.5 GW figure is therefore a forecast, not the pipeline by another name and not an assertion that this capacity is operating now. KPMG’s report uses FY26/FY30 labels in its charts and also uses “by 2030” in the narrative. Preserve the source’s FY30P wording where precision matters; do not silently recast it as a 31 December 2030 inventory date.

What KPMG’s public figure does—and does not—define

KPMG calls the baseline installed data-centre capacity and presents a projected capacity figure. In the pages reviewed, the 7.0–7.5 GW headline is not explicitly cross-walked to IT load versus gross facility power or a utility/grid-connection measure. It would be unsafe to relabel it “7GW IT load” or “7GW of power demand” without a definition from KPMG.

That does not make the forecast useless. It means the correct citation needs a qualifier: “KPMG projects 7.0–7.5 GW of data-centre capacity for FY30; the report’s headline does not make the IT-load versus gross-power basis explicit in the reviewed material.” A data-centre market forecast is only comparable with another forecast after the underlying measurement basis is clear.

Why 1.9 + 4.5 does not automatically equal 7–7.5

The headline arithmetic can tempt a reader to add KPMG’s ~1.9 GW baseline to the ~4.5 GW pipeline and ask why the result is not exactly 7–7.5 GW. That addition is not justified by the published summary alone.

The report does not provide, in the reviewed pages, a project-by-project bridge showing whether the 4.5 GW pipeline is a net addition, whether it includes every future phase, whether capacity has been counted at the same IT/facility basis as the forecast, or how projects that are delayed, redesigned or cancelled are treated. Nor does the source label the pipeline’s complete stage mix in the headline.

The figures could reflect different scopes, cut-off dates, stages, forecast assumptions or capacity definitions. Those are possible explanations—not established facts. The right editorial treatment is “KPMG reports both figures; the relationship between the pipeline and the forecast is not reconciled in the public material reviewed,” not “1.9 + 4.5 proves the forecast” or “the figures do not add up, so one must be wrong.”

Three scoreboards: operating stock, pipeline and forecast

A compact way to interpret the India data-centre capacity story is to keep three separate scoreboards. They should not be drawn on one common scale unless the source confirms comparable dates and units.

Scoreboard 1: operational or installed capacity

This is the closest measure to what exists as operating infrastructure now, subject to each researcher’s definition. It is a stock measured at a stated date, not a running sum of future plans. “Operational” still does not mean every megawatt is occupied, sold, available to a new customer or running at full load. It means the source has classified the facility capacity as operating / in stock.

Scoreboard 2: development pipeline

A pipeline is a list or estimate of projects at different stages, often including projects that are only planned. The farther a project is from construction and commissioning, the more delivery conditions remain. Pipeline capacity is not a guarantee that all projects will be built by a specific date.

Scoreboard 3: forecast capacity

A forecast is an estimate of a future total or trajectory. It can draw on expected demand, announced projects, historical growth, market assumptions and analyst judgment. It may or may not publish the underlying project list. A forecast should carry a horizon, baseline, geography and capacity basis; it is not itself a live inventory or a count of secured projects.

The short rule: operational stock tells you what the source counts as operating at the measurement date; pipeline tells you what projects the source tracks as upcoming; forecast tells you what total the source expects or models in a future period. Even where a provider uses all three, do not assume that the source has published a complete bridge between them.

What the H1 2026 operational baseline says

The reports reviewed place the current Indian market around 1.8–1.9 GW in H1/FY26, not 7 GW. The values are close, but the figures should remain source-attributed instead of being merged into a single supposedly exact number.

Savills’ 1,819 MW IT and C&W’s 1,789 MW differ by 30 MW, but numerical closeness is not a methodology note. KPMG’s ~1.9 GW is rounded and labelled FY26, while the other two report H1 2026. Use the figures as separate reference points, not as a triangulated official census.

The practical conclusion is robust even without forcing a reconciliation: the operating/installed base reported for H1/FY26 is roughly 1.8–1.9 GW on these sources’ own terms. A 7 GW statement is about the future, unless a source explicitly says it is describing current stock—which none of the reports above does.

Operational stock is not utilisation

A facility can be operational while some halls are unleased, unoccupied or reserved for a customer ramp. Some market estimates count designed IT capacity, including occupied and unoccupied space; others may use a different stock definition. An operational-capacity number is not automatically a measure of current electricity draw, server utilisation, sellable capacity, leased capacity or revenue-generating load.

For that reason, a market-size brief should keep separate fields for:

• commissioned / operational capacity;
• customer-occupied or contracted capacity, if disclosed;
• available or vacant capacity, with the provider’s denominator;
• actual electricity consumption or load, if measured; and
• future planned capacity.

The Savills 6 GW versus “over 7 GW” discrepancy

There is a material publication difference that should be visible in any article repeating a “Savills forecasts over 7 GW IT by 2030” line.

• On 22 July 2026, ETDatacenters reported that Savills forecast India’s capacity to reach over 7 GW IT by 2030. The article describes the figure as a nearly fourfold expansion and attributes it to a Savills report/statement.
• Savills’ official H1 2026 Data Centre Market Watch, listed as published on 30 July 2026, says India’s operational stock was 1.8 GW IT and that it was “well on its way to reach 6 GW IT by 2030.”
• The detailed Savills PDF repeats the 6 GW framing. Its city-wise table is headed “Projected Stock in 2030 MW IT”; the nine city / Tier-II entries sum to 6,065 MW IT, or about 6.1 GW IT.

This is not a basis for accusing Savills of an error. The July 21/22 coverage and July 30 report may reflect a forecast update, differing scope, a press-summary paraphrase or another explanation. The material reviewed does not explain which. Until Savills clarifies, write the publication-specific facts and state that the difference is unresolved. Do not present “over 7 GW IT” as the definitive conclusion of the later H1 report.

A fair pre-publication question for Savills would be: Did the July statement of over 7 GW IT by 2030 refer to a different forecast version or scope from the 6 GW IT in the H1 2026 report and its 6,065 MW city table? If so, what changed and which figure should readers use? No request for comment has been sent for this draft.

Other 2030 estimates—and why they are not one range

A reader searching for India data-centre capacity by 2030 will find several figures. A spread of numbers is not automatically evidence that a provider is wrong: the date, geography, customer segment, capacity basis and project stage may differ. But the variation is a reason to label every estimate, not to average them.

Two tempting but incorrect shortcuts follow from this table. First, do not call these figures a 4–7.5 GW consensus range: some are national projections, one is top-seven-city colocation IT load, another is colocation-only, and C&W’s number is pipeline, not forecast stock. Second, do not calculate an average. Without a harmonised basis, an average would have no clear meaning.

The defensible summary is more specific: KPMG publishes a 7.0–7.5 GW FY30 projection; Savills’ later H1 report publishes roughly 6 GW IT; CEEW cites a 4.5–6.5 GW IT-related range from other research; and narrower city/colocation sources publish different totals. The estimate used should match the decision being made.

Savills’ city-by-city 2030 projection

Savills’ H1 2026 report provides an unusually useful detail for reading a forecast: a city-wise table that labels both current stock and 2030 projected stock in MW IT. The report’s current city rows total 1,819 MW IT; its 2030 rows total 6,065 MW IT.

This table helps answer what Savills’ 6GW means: it is a projected operator-based stock expressed as IT capacity, assembled from named city categories. It does not say that 6,065 MW has been financed, built, connected to the grid, contracted by customers or commissioned. Nor does it identify a project-by-project delivery schedule for each projected city total.

Hyderabad’s projected rise is particularly large in the table, from 90 MW IT in H1 2026 to 2,004 MW IT in 2030. That is a forecast signal about expected geographic growth—not proof that 1,914 MW of Hyderabad projects are already under construction. A city forecast cannot substitute for a dated inventory of named projects and their stages.

Why IT load, facility power and grid capacity matter

A data centre can be described with several power-related measures. The same number of megawatts can mean different things depending on whether it refers to computing equipment, the full building, a utility connection or a planned campus at full buildout.

• IT load / IT capacity: Power for servers, storage and networking equipment, subject to the source’s definition. Savills and Colliers explicitly label relevant figures as IT capacity / IT load.
• Facility or gross power: A wider measure that can include IT equipment plus cooling, electrical conversion, lighting and other building systems. The exact definition varies by source.
• Grid connection / contracted demand: The amount the facility can draw from a utility or power system under the relevant connection or contract. It is not automatically the same as the IT load.
• Energy consumption: Electricity used over time, usually measured in MWh, GWh or TWh. A capacity figure in MW or GW is power, not annual energy.

For an illustrative example, a source might report 100 MW IT load for a facility while the site’s total power requirement is higher because supporting systems also consume electricity. The difference depends on design and operating conditions; there is no safe universal multiplier to apply to an India market forecast.

That is why KPMG’s unqualified 7.0–7.5 GW capacity number should stay in the source’s own language unless KPMG confirms the basis. Do not silently transform it into “7GW IT load,” “7GW grid demand” or “7GW of power consumption.” Those expressions answer different questions.

Capacity basis changes comparisons

Colliers’ 2025 report is unusually explicit about one important boundary: its total represents colocation data centres in seven cities and total designed IT load, including occupied and unoccupied capacity. A reader comparing it with a national forecast that includes hyperscaler self-builds, enterprise and edge facilities may be comparing different universes.

Savills’ H1 report describes operator-based stock, reports IT capacity, and breaks its stock into dedicated hyperscaler, enterprise, edge and mixed categories. C&W reports operational capacity and pipeline, but its H1 report’s headline metric does not provide the same explicit IT-load wording in the pages reviewed. KPMG uses installed/projected capacity without a clear IT-versus-gross crosswalk for its 7GW target. These differences belong beside the numbers, not hidden in a footnote.

How to read a data-centre pipeline

“Pipeline” can refer to anything from a publicly announced campus concept to a project already being built. A serious comparison should ask whether the source separates at least these stages:

1. Announced / proposed: A company or government has publicly described an intention, target, MoU or proposed development. This is not proof of land, power, financing or construction.
2. Planned: The source includes the project in an expected supply list, but construction has not started. The project may still depend on site, approvals, utility arrangements, financing, customer demand or other conditions.
3. Under construction: Physical building work is reported as underway. This is a stronger delivery signal, but does not establish that the full stated capacity is complete or powered.
4. Commissioned / operational: A facility or phase has passed the source’s operational threshold. A campus can still be phased, partly occupied or ramping.
5. Commercially serving customers: Services are available to customers. This is a distinct commercial milestone and may occur after testing or commissioning.

KPMG’s report says its development pipeline is expected to add about 4.5 GW over five years, but the cited headline does not publish the same 509 MW / 3,351 MW stage split as C&W. C&W’s H1 2026 update reports 3,860 MW of upcoming supply through 2030, of which 509 MW is under construction and 3,351 MW is planned.

The two pipeline numbers—4.5 GW and 3.86 GW—should not be treated as a clean discrepancy or as alternative forecasts. They may differ in report date, stage threshold, geography, facility coverage, project inclusion rules or measurement basis. The public material reviewed does not provide a shared project register to reconcile them.

Why a pipeline is not a 2030 operating total

A project can appear in a pipeline without being complete by the forecast horizon. A pipeline may include projects with different probability of delivery; it may contain overlapping phases or parent-company targets; it may not include all potential future projects; and some planned projects can be delayed or cancelled. The pipeline can also describe additional supply, while a forecast describes total stock at a future date.

Do not simply add current capacity to the pipeline and call the result a 2030 forecast. That arithmetic would require a shared capacity basis, geography, cut-off date, inclusion rule and assurance that the pipeline is net new and not overlapping with already counted stock or forecast additions. Those conditions are not established by the headline figures reviewed here.

What would have to happen for a forecast to become operating stock?

A 2030 forecast is a statement about an expected future market, not evidence that every input has already been secured. A campus’s planned MW must pass through multiple dependencies before it is counted as operating capacity.

1. Site and land control

A project needs a location, rights to use the land, relevant approvals and a design that fits the parcel. A land announcement, allotment approval or option is not the same as final possession or a powered facility. Project-level land facts are important because city-wide forecast capacity does not show which sites are firm.

2. Power and network readiness

The project needs an electricity supply arrangement, a feasible connection, adequate transmission/distribution infrastructure and a schedule aligned with construction and commissioning. A state’s generation capacity or a policy target does not prove that a specific data-centre phase has a firm connection at the required time.

3. Financing and customer demand

A development may need equity, debt, a customer commitment, a lease or a hyperscaler’s internal approval. A broad company investment pledge is not necessarily a funded project budget. Colocation facilities also depend on customer demand, leasing and the timing of fit-out.

4. Construction, equipment and commissioning

The campus must be built and equipped. Electrical systems, cooling, backup power, controls, fire protection and network links all need testing. A headline construction milestone does not make every planned phase operational.

5. Customer service and ramp

Even when a building is ready, customer migration, server installation, acceptance testing and service activation can occur in phases. A market report may count capacity as operational before it is fully utilised. Actual load or electricity consumption is a separate measurement.

KPMG and Savills both discuss demand drivers such as cloud, AI and hyperscale workloads. Savills also identifies power and suitable land as important development challenges. Those factors make the 2030 estimates worth tracking; they do not independently prove that a specific forecast will be achieved or missed.

A reusable checklist for testing a 7GW headline

Use this checklist before repeating any India data-centre 7GW forecast in a presentation, news story, investment note or procurement plan.

1. Who published it? KPMG, Savills, a broker, a ratings agency, government or a company? Link to the original report, not only a headline.
2. What is the source date? A forecast published in July 2026 may differ from a year-end 2025 outlook. Preserve the report date and any later update date.
3. What is the exact period? FY30, “by 2030,” end-2030 or a five-year horizon from publication? Do not silently convert fiscal-year labels into calendar-year inventory dates.
4. What geography is included? All India, top seven cities, selected metros, one state or one operator’s portfolio?
5. What capacity basis is named? MW IT, total facility power, contracted grid capacity, site load or an undefined installed-capacity number?
6. What project stage is counted? Operating, under construction, planned, announced, permitted, contracted or some combination?
7. Is the figure a stock or an addition? “Total capacity by 2030” differs from “pipeline additions over five years.”
8. Are phases and parent totals separated? Do not add a company-wide target to its campus phases if they may describe the same capacity.
9. Is a current baseline supplied? A forecast without a dated baseline is difficult to interpret as a growth rate or delivery path.
10. Does the report publish a bridge? Look for the assumed additions, retirements, cancellations, ramp timing and capacity conversion. If absent, say it is absent.
11. Does another primary source disagree? The Savills H1 report is a reminder to check the report itself, not just a media paraphrase.
12. What does the number not say? A forecast does not prove financing, construction, customer occupancy, energy use or capital flows.

A compact data card for research teams

This kind of card is more useful than a spreadsheet cell containing only “India: 7 GW.” It makes the source’s definition auditable and gives the next researcher a clear question to resolve.

What the numbers mean for investors, operators and suppliers

Different users should use different scoreboards.

For investors and lenders

Use a forecast to frame a scenario, not to underwrite a guaranteed future market. Ask which demand segments and project phases the model includes; which power and land assumptions are embedded; how customer commitments are treated; and what the downside case looks like if announced projects are delayed. At the asset level, verify permits, power arrangements, financing and phase timing rather than substituting a national 2030 estimate for due diligence.

For operators and hyperscalers

A national capacity forecast can help with broad market entry, but site choices depend on city-level demand, power availability, connectivity, water and permitting. Savills’ city table makes clear that one provider’s model expects materially different growth by market. Treat those rows as a published outlook, then test them against operator announcements and verified project stages.

For equipment vendors and contractors

A pipeline is more relevant than a national forecast for estimating procurement windows—but only after projects are screened by stage, scope, package and timeline. A planned 500 MW campus is not a confirmed purchase order. The actionable unit is usually a named project phase with a credible construction / procurement schedule and a defined technical scope.

For market analysts and journalists

Keep the 1.8–1.9 GW current baseline, the 3.86–4.5 GW reported pipeline measures and the 6–7.5 GW forecast estimates in separate fields. Do not average provider totals or present a national forecast as an inventory. The Savills discrepancy is worth asking about, but it is not a reason to imply misconduct or error before the provider responds.

For Singapore-based readers

A consistent India capacity forecast may be relevant to regional market sizing, operator strategy and cross-border investment analysis. It is not evidence that capital has moved from Singapore to India, that a Singapore project has been displaced, or that any named fund or operator is responding to the figure. Those are separate claims requiring project-level funding and decision evidence.

Glossary: capacity, stages and forecast language

Frequently asked questions

Is India already operating 7 GW of data-centre capacity?

No published baseline reviewed here says that. Savills reports 1,819 MW IT of H1 2026 operator-based stock, C&W reports 1,789 MW operational capacity in its H1 2026 update, and KPMG gives about 1.9 GW installed for FY26. The 7.0–7.5 GW figure is KPMG’s forward projection for FY30, not the current operating inventory.

What does KPMG mean by “7GW by 2030”?

KPMG’s report projects 7.0–7.5 GW of capacity for FY30. The report separately gives a ~1.9 GW installed baseline for FY26 and says a ~4.5 GW development pipeline is expected to add capacity over five years. In the reviewed pages, the 7GW headline is not explicitly cross-walked to IT load versus gross facility power.

Is the 4.5 GW KPMG pipeline the same as the 7.0–7.5 GW forecast?

No. KPMG presents the pipeline and future-capacity projection as separate measures. The public pages reviewed do not show a project-level reconciliation between them. Do not treat pipeline additions as the future total or force the numbers to add up.

Why does Savills say both 6 GW and over 7 GW?

The July 22 ETDatacenters article reports an “over 7 GW IT” forecast attributed to Savills. Savills’ official H1 2026 report page and detailed PDF say 6 GW IT; the city-wise table sums to 6,065 MW IT. The public material reviewed does not reconcile the difference. It could reflect a change or a different scope, but that has not been established; seek clarification and attribute each figure to its specific publication.

Is the Savills 6 GW figure IT load?

Yes. The H1 2026 PDF labels current and projected city stock in MW IT and says the country is on track to reach 6 GW data-centre capacity by 2030. Its city projections total 6,065 MW IT.

Is C&W’s 3.86 GW figure a forecast?

It is C&W’s reported upcoming supply / pipeline through 2030, split into 509 MW under construction and 3,351 MW planned. It is not the same as an estimated total operating stock in 2030.

Why can two providers give different operational-capacity totals?

They may use different reporting dates, facility lists, city coverage, inclusion rules, capacity definitions or treatment of hyperscaler and enterprise facilities. A near match—such as 1,819 MW IT versus 1,789 MW—does not prove that the methodologies are identical. Compare the source notes before using either as a universal census.

Can I call KPMG’s 7.0–7.5 GW an IT-load forecast?

Not from the reviewed material alone. KPMG describes installed/projected data-centre capacity, but its headline does not clearly state the IT-load versus gross-power basis. Use KPMG’s own wording until the provider confirms the measurement.

Does Colliers’ 4.5 GW projection cover all of India?

Colliers’ May 2025 source describes capacity across the top seven cities and identifies the figure as colocation data-centre IT load, including occupied and unoccupied capacity. It should not be presented as a directly comparable all-India forecast without qualification.

Should I average the published 2030 forecasts?

No. They have different scopes and units, and some refer to a pipeline or a segment rather than total national capacity. Averaging them would imply a shared statistical basis that the sources do not provide.

What would resolve the main open questions?

A KPMG note defining the 7.0–7.5 GW capacity basis and explaining how its 4.5 GW pipeline relates to the FY30 projection; a Savills clarification on its July “over 7 GW IT” statement versus the later 6 GW report; and a provider-level crosswalk of project stages, geography and MW basis would materially improve comparability.

Sources and further reading

1. KPMG, India data centre opportunity: From emerging demand hub to integrated data centre powerhouse, 1 July 2026 — official report page and full PDF — FY26 installed baseline, five-year development pipeline and FY30 projected capacity.
2. Savills India, Data Centre Market Watch H1 2026 — official report page and full PDF — operational stock, city-wise H1 2026 baseline and explicit MW IT projection for 2030.
3. ETDatacenters, “India’s Data Centre Capacity Surges 59% in H1 2026, Forecasts 7 GW by 2030,” 22 July 2026 — secondary report attributing an “over 7 GW IT” forecast to Savills; compared with the later official report above.
4. Cushman & Wakefield, India Data Centre H1 2026 Update — official report page and H1 2026 report PDF — operational capacity, city distribution and upcoming supply split.
5. Cushman & Wakefield, “India’s Data Centre development Pipeline through 2030 Reaches 3,860 MW,” September 2026 — official summary of 509 MW under construction and 3,351 MW planned.
6. CEEW, Scaling India’s Data Centre Ecosystem, February 2026 — publication page and white paper PDF — cites a 4.5–6.5 GW 2030 outlook drawing on Colliers and S&P Global.
7. Colliers, The Digital Backbone: Data Centre Growth Prospects in India, 28 May 2025 — report page, official release and report extract — top-seven-city colocation IT-load scope and capacity outlook.
8. S&P Global, “Will India become a leading global datacenter market?”, 17 September 2025 — reports India’s IT-load baseline and additional pipeline-based capacity expected by 2030.
9. CareEdge Ratings, India’s Data Centre Capacity to Grow 4x to ~4 GW by 2030, 25 March 2026 — co-location-only capacity outlook.
10. Gridvik, India’s Data Centre Boom and Expansion (2026–2030): Capacity, Cities, AI, Power and What Comes Next — related live guide on market definitions, forecast interpretation and delivery constraints.

Related Gridvik guides

These public Gridvik guides were checked as live pages on 12 October 2026 (IST). They provide broader context; they are not the unpublished local draft linked nowhere in this article.

• India Data Centre Investment Map 2026: Big-Tech Commitments, Operators, Finance and State Incentives — separates investment announcements and policy targets from deployed capital and operating assets.
• India’s Data Centre Boom and Expansion (2026–2030): Capacity, Cities, AI, Power and What Comes Next — wider market guide covering capacity, cities, forecasts and delivery constraints.
• India’s Data Centre Power Bottleneck: Grid Readiness, State Policy and the Pre-Tender Window — explains why state-level power or policy signals do not by themselves establish project-level grid readiness.
• CoreWeave’s 240 MW India AI Data Centre Plan: What the AdaniConneX Taloja Campus Means — a project-specific example of planned capacity, future phases and what an announcement does not establish about present operation.

Editorial status: Draft only. The KPMG capacity basis and pipeline-to-forecast bridge, and the Savills July “over 7 GW IT” versus H1 report “6 GW IT” difference, require clarification or prominent unresolved-status treatment before publication. No company, research provider or executive was contacted or tagged for this draft; seek a fair opportunity to respond before publication.