Why Aren’t the Data Centres Paying Us?

Here is a sample of what is on offer across Indian states as they compete fiercely to attract data centre investment:

• 20-year power tariff subsidy of ₹1 per unit

• Full electricity-duty reimbursement

• 100% stamp-duty exemption

• 25-50% land subsidies

• Halving the mandatory green power requirement (Maharashtra: 100% to 51%)

• Deemed Distribution Licence (bypassing state power companies)

The entire discussion is about how to structure deals, navigate approvals, and maximise "returns", presumably in employment or ancillary industrial activity. But what those returns actually amount to, and who they accrue to, is rarely examined.

What do the communities or the state actually receive in exchange? 

Temporary construction jobs, a handful of permanent positions, and the risk of water shortages and ecological damage - as in Visakhapatnam, where the site sits just 860 metres from the Kambalakonda Wildlife Sanctuary, home to leopards and pangolins. 

Official documents for the Google-Adani project there show only 2,415 jobs during the construction and operational phases combined (90 permanent and 400 temporary during construction, and 1,725 permanent and 200 temporary during operations). Even these figures are contested. Industry data suggests a hyperscale facility may employ just 30-60 permanent on-site staff; the bulk of the job numbers are temporary construction roles or indirect ecosystem jobs, not permanent, local employment.

The detailed legal and policy guidance aimed at foreign investors makes no mention of community benefits, water use, or local impacts. While environmental activists have gone to court against the Google Vizag project, the current policy conversation in India is still about attracting such investment, not about managing its consequences - consequences which have seen communities blocking or delaying data centre projects in the US and elsewhere.

Notwithstanding the finer differences in the nature of local opposition, one question in this global backlash is simple, sharp, and growing louder: why aren’t the data centres paying us?

What Data Centres Really Make
To understand why the question matters, we need to grasp the scale of what’s at stake: three companies alone - Meta, Alphabet, and Amazon - generated nearly £1.3 trillion in revenue in 2024-25, virtually all derived from monetising our personal data. Just Alphabet, by itself, reached £406.2 billion, with the vast majority of it coming from advertising powered by our search histories, email metadata, location tracking, and browsing behaviour.

A recent study found that a single UK internet user’s lifetime data contribution generates up to £194,000 in value for big tech and AI companies.

The global digital economy is now estimated at approximately £16 trillion, roughly 15% of world GDP.

This is our data. We generate it. They monetise it. We have no voice in how it is valued or shared - and neither, it appears, do the governments that are meant to represent us. 

The state negotiates incentives, not our interests.

Meanwhile, communities that host data centres bear the costs: water stress, grid strain, and environmental disruption. Data centres are turning into local businesses with local impacts, similar to highways, industrial corridors, and renewable energy parks. But unlike those forms of infrastructure (which create lasting local employment, supply chains, and tax bases), data centres generate almost no permanent local economic activity. A hyperscale facility runs on a skeleton crew of specialised technicians; the promised job bonanza is largely a mirage of temporary construction work and inflated “ecosystem” estimates.

The Johor, Malaysia Case: A Model for Community Engagement
In Malaysia’s Johor state, a different story is unfolding. Data centre capacity is projected to triple by 2030, but the state has already halted approvals for less-efficient Tier 1 and Tier 2 facilities (these lower tiers are less reliable, have minimal backup systems, and are far more water-intensive).

In Gelang Patah, a data centre developer faced a protest led by a local resident. The issue was water. In response, the data centre engaged directly with the community and demonstrated its solution: it was building its own water treatment plant to process wastewater for cooling. The company told residents it was using 100% reclaimed water from a sewage treatment facility, not competing for drinking water.

The protest leader stated then that his fears had been allayed as the developer had committed to using recycled water or wastewater, and that he had also sought opportunities for local youth, either immediately or long term.

In Ulu Tiram, another developer has built a water reclamation plant that currently processes 5 million litres of treated wastewater daily, with plans to expand to 20 million litres. The water is treated through membrane bioreactor and reverse osmosis systems, then used for cooling before being discharged into rivers at room temperature.

The Johor experience offers three vital lessons: investment in water infrastructure is feasible when demanded; direct community engagement can transform opposition into support; and the model is replicable in water-stressed regions like India - but it requires political will and regulatory mandates.

Water Concerns Addressable
The resource concerns that fuel community resistance are not insurmountable, as most of the technology to address them already exists. The old way of cooling data centres, i.e., air conditioning and evaporative cooling towers, is reaching its physical limits, but the solution is at hand. Nvidia’s new Rubin platform uses a closed-loop liquid cooling system that achieves near-zero water consumption in suitable climates, saving over $4 million annually per 50 MW facility.

If Johor can insist on reclaimed water and closed-loop systems, so can states in India. The obstacle is not engineering but the absence of mandates requiring its use. The question for Indian states is whether they will demand what is technologically feasible, or continue accepting projects that externalise their water costs onto local communities.

Other Models Emerging: The US Push for Community Benefit Agreements
The community backlash in the US is generating its own policy responses. In Illinois, the proposed POWER Act would mandate annual payments equal to at least 10% of the property tax that would have been owed, with not less than 50% of those funds used to offset residential tax burdens. The payments would be deposited into a segregated fund and cannot be commingled with general corporate funds. This ensures that communities see tangible, recurring benefits.

At the federal level, the proposed US Local Control Protection Act would make community benefit agreements a condition for federal tax incentives for data centres, with terms covering local tax liabilities, infrastructure mitigation, environmental monitoring, and local workforce utilisation. These are early signals, not yet a trend, but a model that could gain traction. They show that communities are beginning to demand a seat at the table, and legislators are beginning to listen.

The Vizag Case Goes to Court
In August 2026, the Vizag community’s concerns reached the Andhra Pradesh High Court. The petition raised the very issues at the heart of this article: the use of (temple) land, the adequacy of environmental clearances, and the project’s massive water and electricity demands in a water-stressed region.

The court last week refused an interim stay on construction, allowing work to proceed. However, it delivered a strong message: all construction must be carried out only after obtaining necessary legal and environmental permissions, and “environmental clearances cannot be given a go-by.” The state’s defence - that as “data centre” is not a listed activity under the 2006 EIA Notification, the clearance was granted based on the facility’s built-up area rather than its operations - exposed a critical regulatory loophole. The court has demanded a consolidated response from all approving departments, signalling deeper scrutiny to come.

The case is a reminder of what’s at stake when incentives outpace accountability, and also a test of whether India’s courts will fill the policy gap that states have left open.

A Better Model: The Auction + Community Benefit Framework
The Accelerated Community Energy (ACE) model, developed by energy software company Piclo, offers a proven alternative. The model is already live across multiple regions in the UK and is being rolled out in the US and Australia. Here is how it works:

• Data centres face long delays to connect to the grid.

• Under ACE, they can bid for priority connection slots in an auction.

• The money from those bids goes into a fund that supports community energy projects like local batteries, solar, or demand-response programs.

• Communities benefit directly; data centres get faster access; the grid gains flexible capacity.

The economics are compelling. Set aside just 3% of a 100 MW data centre's $1.275 billion build cost - $38 million - and communities could receive incentives comparable to California's successful Demand Side Grid Support program. That program pays customers $60-82/kW-year for reducing load during grid emergencies, effectively creating a distributed "virtual power plant" that helps prevent blackouts. At national scale, this could unlock $40 billion for virtual power plants and community energy programs, adding 100 GW+ of flexible capacity to the grid. Data centre operators, losing an estimated $18.4 million in revenue for every month stuck in queues, could achieve return on investment in just two months.

This is the direction India should be moving in: not competing on subsidies, but competing on the benefits data centres can offer back to the communities that host them.

A Community Benefit Framework for India: State-Level Solutions
All the solutions proposed here operate at the state level in India. Land, electricity, water, and local approvals are state subjects. The incentives described in this article (power tariff subsidies, electricity duty exemptions, stamp duty exemptions, land subsidies) have all been offered by state governments. The Deemed Distribution Licence was granted by Andhra Pradesh. The Johor moratorium was imposed by the Johor state government. The Illinois POWER Act is a state bill.

So this is not a federal-level problem but a state-level policy challenge, and it is at the state level that solutions need to be designed and implemented. A Community Benefit approach would require state governments to mandate that data centres:

1. Pay for Power Infrastructure: This would entail funding all new substations, transmission lines, and grid upgrades themselves- no passing of costs to local consumers. Google’s Monrovia, Indiana agreement already does this, delivering over $770 million in customer savings over 15 years.

2. Provide Firm Energy Commitments: This would mean they become grid assets, not liabilities, building in the ability to ramp down during peak hours using on-site batteries or generators.

3. Adopt Closed-Loop or Reclaimed Water Cooling: Ensure zero or near-zero freshwater consumption using existing technology, or invest in wastewater reclamation as Johor has done.

4. Auction Site Permissions: Have bids for priority interconnection slots, replacing the current scramble with an orderly, transparent process.

5. Make Direct Community Payments via Local Government/ULBs: Make annual payments to host communities, ring-fenced for infrastructure and education (something that is needed for many social infrastructure projects). The Illinois model of annual payments of at least 10% of the property tax that would have been owed provides a useful template.

6. Ensure Transparency: There Should be no NDAs on Environmental Data: Mandate periodic public disclosure of water and power consumption and independent monitoring of all binding commitments.

These requirements are not theoretical, as each one is already being implemented somewhere in the world, whether in Indiana and Illinois in the US, Piclo in the UK, or Johor in Malaysia.

Can India adopt them? India’s data centre boom offers a genuine opportunity, but it also risks repeating the mistakes of the West- offering massive subsidies in a race to the bottom while communities bear the costs. An analysis noted that “India cannot just lure investment to become an AI leader. It also has to earn what is called a ‘social license to operate’.” Trust cannot be legislated or certified but is earned through transparency, meaningful consultation, and a fair sharing of benefits and costs.

The question at the heart of the global backlash is simple: why aren’t the data centres paying us? In India, this question has a particular urgency - the state is offering huge concessions by way of subsidies, exemptions, and even the power to bypass the grid itself. Yet it captures almost none of the value in return through firm commitments. 

Why isn’t the State being Paid?
India can continue offering ever more generous incentives in a race to the bottom - or it can shift the competition. Instead of competing on subsidies, states compete on the benefits they can extract for their communities. They can demand that data centres pay their fair share through auction models, community benefit agreements, and closed-loop technology. They can use this moment to establish a framework that ensures the AI revolution benefits the communities that host it.

The Johor, Malaysia case proves that data centres can be built in water-stressed regions without depleting local resources, provided the political will exists. The US proposals show that communities are beginning to organise and legislators are beginning to listen. The Andhra Pradesh High Court's recent directive signals that India's judiciary is beginning to scrutinise what states have rushed to approve. Will India's policymakers act before the backlash forces their hand, or will they wait for the courts, and the communities, to do it for them?