Karnataka's New Data-Centre Policy Makes Power, Water and Location Part of AI Infrastructure Planning

Karnataka's Sustainable Data Centre Policy 2026–2031 makes electricity, water, cooling, renewable energy and location explicit parts of the state's plan to expand digital infrastructure. Approved by the state cabinet on September 18, 2026, the five-year framework targets up to 1 GW of cumulative data-centre IT load by 2031 and aims for Karnataka to account for 10–12% of India's AI-ready data-centre capacity.

The important signal is not only the capacity target. The policy treats data-centre growth as a resource-allocation problem: where facilities are built, how much power and water they need, how they cool equipment, and how their electricity demand fits into the wider infrastructure system.

What Karnataka approved

The cabinet-approved policy covers more than conventional hyperscale facilities.

The IT/BT department's reported framework includes data-centre units and parks, edge data centres, cable-landing infrastructure and related digital infrastructure. It also proposes incentives connected to capital investment, power, connectivity, renewable energy, research and development, and skills.

Implementation is expected to involve the Department of Electronics, Information Technology, Biotechnology and Science & Technology, with Karnataka Innovation and Technology Society handling registration and policy implementation and Karnataka Udyoga Mitra serving as a single-window agency for clearances.

The policy's proposed budget framework is about ₹1,300 crore.

The 1 GW target is a capacity goal, not delivered capacity

Karnataka is targeting 1 GW of cumulative data-centre IT load by 2031.

That number describes an intended future capacity level. It is not evidence that 1 GW of new capacity has already been financed, built or connected.

The policy also aims for Karnataka to represent 10–12% of India's AI-ready data-centre capacity by 2031. Again, that is a policy objective rather than a measured market share today.

Keeping the distinction clear matters because data-centre announcements often mix investment intentions, planned capacity and operational capacity.

The Observatory treats the policy targets as policy signals until deployment data shows what was actually delivered.

Why power and water appear beside compute

A modern data centre is simultaneously a computing facility and a large physical utility customer.

Its operating constraints can be represented as:

compute demand
     ↓
electricity + cooling
     ↓
power / water / land constraints
     ↓
location and grid decisions

Karnataka's policy explicitly calls for resource-availability mapping to identify where electricity, water, land and connectivity can support new facilities.

The state also plans resource heat mapping to identify locations suitable for development.

That is a notable planning shift. Instead of assuming that data centres can be placed wherever land is available, the policy proposes matching projects to the infrastructure resources they require.

Sustainability is being made an infrastructure requirement

The policy emphasizes renewable power, energy efficiency, advanced cooling, efficient water use, wastewater recycling and reduction of carbon emissions.

The exact effectiveness of those measures will depend on implementation and measurement.

For example, saying that a facility uses renewable electricity does not by itself establish its total water footprint, hourly grid impact or lifecycle emissions. Likewise, more efficient cooling reduces resource use per unit of compute but does not necessarily reduce total resource demand if overall capacity grows much faster.

The policy therefore creates a framework for measurement; it does not yet provide evidence of the eventual environmental outcome.

Why Mangaluru matters

The policy explicitly seeks to expand the data-centre ecosystem beyond Bengaluru.

Mangaluru is identified as a location with potential for new facilities, supported by coastal connectivity and the possibility of future submarine cable-landing infrastructure. Mysuru is also mentioned as an expansion location.

This matters because data-centre concentration creates its own constraints.

If every large facility competes for the same power, land and network resources around one metropolitan region, geographic diversification can become an infrastructure strategy rather than simply a regional-development goal.

Whether Mangaluru develops into a significant data-centre hub remains uncertain. The policy establishes an opportunity, not an already-built cluster.

The policy connects AI growth to physical infrastructure

The Observatory has already tracked AI data centres becoming flexible grid loads.

That article focuses on a different question: how large AI facilities could adjust their electricity demand in response to grid conditions.

Karnataka's policy approaches the problem from the planning side.

Together, the signals point to a broader shift:

AI model growth
      ↓
more compute
      ↓
more data-centre capacity
      ↓
power + water + network + land requirements
      ↓
regional infrastructure policy

AI infrastructure is therefore increasingly becoming a government planning issue, not only a technology-company investment decision.

What the policy could change for developers

For developers, the most consequential part may be the move toward infrastructure-aware site selection.

A project could increasingly be evaluated not only on land and construction economics, but also on:

• available electricity and grid connectivity;
• renewable-energy access;
• cooling and water availability;
• network connectivity and potential cable landing;
• proximity to skilled workers and research institutions; and
• whether the site fits designated data-centre zones or parks.

The policy also proposes a Centre of Excellence for Data Centres, intended to support research and innovation around the sector.

These mechanisms could make Karnataka more attractive to infrastructure investors, but their effect will depend on how quickly incentives, clearances and physical infrastructure are actually delivered.

What the policy does not prove

The September 18 cabinet approval does not prove that Karnataka will reach 1 GW by 2031.

It also does not establish that every proposed data centre will operate on renewable energy, achieve a particular water-efficiency level, or locate outside Bengaluru.

The 10–12% AI-ready capacity figure is a policy target, not a forecast that can be treated as guaranteed.

The strongest current evidence is therefore about intent and policy design, not future execution.

Why this is meaningfully new

The Observatory has covered data-centre energy flexibility at the technology level, but not Karnataka's newly approved regional infrastructure framework.

This article adds a different system layer: state-level planning for the physical resources required by AI and cloud infrastructure.

The strongest conclusion is modest: Karnataka's 2026–2031 policy makes data-centre expansion explicitly dependent on power, water, cooling, connectivity and location planning, while setting a 1 GW IT-load target. Its significance will ultimately be measured by how much of that planned infrastructure is actually financed, connected and operated sustainably.

Sources and further reading

• Economic Times — Karnataka approves data centre policy
• The Indian Express — Karnataka data centre policy targets 1 GW
• Times of India — Karnataka Sustainable Data Centre Policy
• Moneycontrol — Karnataka Sustainable Data Centre Policy