Can Subsidies Alone Deliver Gujarat’s ₹6 Lakh Cr AI Data Centre Plan?

Beyond Mumbai, Chennai, and metro cities, Gujarat is now pitching itself as the next data centre hub in India.

The Gujarat government recently unveiled Viksit Gujarat Data Centre Policy 2026–29, inviting global hyperscalers, cloud service providers, AI companies, digital infrastructure developers, technology enterprises, and institutional investors to establish hyperscale green AI data centres. 

With a target of 7.5 GW of AI-ready data centre capacity and ₹6 lakh crore in fixed capital investment over eight years, the state is banking on generous incentives for hyperscalers to set the ball rolling. It mandates that projects source at least 51% of their operational electricity from renewable energy, reflecting a broader shift in India's AI infrastructure as data centres are poised to be the next opportunity driver but also raise concerns about their environmental impact.

The state government wants water to be sourced through dedicated captive desalination plants. To facilitate this, it will cover 20% of the eligible capital expenditure up to ₹2 crore per million litres per day. 

Meanwhile, other fiscal incentives include a capital subsidy of 2.5% of investment, up to 4% interest subsidy up to ₹25 crore per year, a subsidised power tariff rate of ₹1 per unit for 20 years, 100% exemption on stamp duty, as well as SGST exemption on plant machinery, building infrastructure, and eligible services.

The state believes it already possesses many of the ingredients hyperscalers seek. Gujarat has about 69 GW of installed power capacity, of which 47 GW comes from renewable sources. It boasts industrial clusters such as Dholera and GIFT City, as well as a strong manufacturing ecosystem built around semiconductors and electronics.

But while Gujarat's ambitions are high, industry experts argue that building AI infrastructure requires much more than policy incentives. The bigger challenge lies in creating the ecosystem that can sustain long-term demand.

More Than a Data Centre Policy

For Gujarat, the policy attempts to move beyond attracting manufacturing investments into becoming a foundational layer of India's AI economy.

Kartik Narayan, CEO of jobs platform, Apna.co, tells AIM that the rationale extends far beyond simply hosting more server farms.

"The larger shift underway is that data centres are becoming foundational digital infrastructure, much like highways, ports, or power networks. As AI adoption accelerates, cloud computing expands, and digital services continue to scale, the demand for secure data storage and computing capacity is growing rapidly. States that invest early in this infrastructure are positioning themselves to attract long-term technology investments," he states.

According to him, Gujarat already enjoys several structural advantages, including reliable power availability, industrial land, strong connectivity, and a policy environment that has historically supported manufacturing and large-scale investments.

"What makes this particularly significant is that data centres generate demand beyond technology professionals. They create opportunities across facility management, electrical systems, cooling infrastructure, security, and maintenance, making them an important source of employment across multiple skill levels," he states.

Not Shifting Away From Semiconductors

The announcement comes as Gujarat continues to aggressively court semiconductor investments through projects in Dholera and Sanand.

Tata Electronics is building a semiconductor fabrication plant in Dholera, in partnership with Taiwan’s PSMC. Meanwhile, Sanand is serving as an OSAT (outsourced semiconductor assembly and test) hub, with Micron Technology, CG, and Kaynes Semicon plants already operational.

However, experts caution against interpreting the policy as a pivot away from chips.

"Rather than viewing this as a shift from semiconductors to data centres, it is more accurate to see them as complementary parts of the same digital infrastructure ecosystem," Narayan states. “Semiconductor manufacturing builds the hardware that powers the digital economy, while data centres provide the infrastructure where that computing power is deployed and managed.”

The overlap extends into workforce requirements as well.

"Electrical technicians, HVAC and cooling specialists, instrumentation engineers, and facility operators are essential to both semiconductor facilities and data centre operations," he adds.

Forrester Principal Analyst Biswajeet Mahapatra believes that Gujarat is trying to extend the semiconductor bet into the next layer of digital infrastructure. AI-ready data centres can offer Gujarat a faster route to monetise investments already made in power, industrial corridors, and renewable energy.

"The timing also reflects a shift in enterprise demand: AI workloads need high-density compute, predictable power, low-latency connectivity, and sovereign infrastructure," Mahapatra explains.

The Real Challenge

While Gujarat has announced one of the country's most aggressive incentive packages, experts warn that incentives alone rarely determine where hyperscalers invest.

Mahapatra argues that semiconductor investments improve Gujarat's credibility but do not automatically attract cloud providers. Instead, companies evaluate multiple operational factors.

"Hyperscalers choose locations based on different priorities: grid reliability, fibre density, cloud demand, latency zones, renewable energy access, operating risk, and long-term policy stability," Mahapatra adds.

He believes that Gujarat's credibility will ultimately depend on execution rather than announcements.

"Hyperscale data centre demand will materialise only if the state can prove 24x7 power resilience, scalable connectivity, cooling economics, and speed of approvals," he says.

Is India Entering the Data Centre Race?

Nearly every major state, including Maharashtra, Uttar Pradesh, Tamil Nadu, Telangana, and Karnataka, is either updating or drafting dedicated data centre policies. As states compete to attract AI infrastructure, another question is emerging: could India replicate the healthy competition already seen in semiconductors?

Mahapatra believes so. "India is seeing the early signs of a state-level incentive race in data centres, though it is less intense than semiconductors today," he mentions.

States are increasingly competing through capital subsidies, electricity concessions, tax reimbursements, and single-window approvals. But he warns against using incentives without measuring long-term economic value.

"The fiscal risk is not the incentive itself, but poorly targeted incentives that subsidise land- and power-heavy assets without securing local cloud demand, skilled jobs, ecosystem spillovers, renewable capacity, or measurable digital-economy outcomes," Mahapatra states.

Large industrial groups are also increasingly investing simultaneously in semiconductors, renewable energy, and AI infrastructure.

Mahapatra believes that it reflects a much larger strategic shift. "Large industrial groups are seeing semiconductors and data centres as connected infrastructure plays, not separate bets. Chips, AI compute, cloud infrastructure, renewable power, industrial land, and digital services are converging into one strategic stack."

Rather than isolated investments, conglomerates are attempting to control multiple layers of future AI infrastructure.

"Owning or influencing multiple layers of that stack gives conglomerates scale advantages in energy procurement, land aggregation, infrastructure execution, public-sector alignment, and long-term participation in AI-led demand," he explains.

The Talent Question

Despite the focus on billion-dollar infrastructure projects, Narayan believes the real bottleneck will be workforce readiness. He argues that discussions around AI infrastructure often overlook the technicians needed to operate it.

Instead, future hiring will increasingly revolve around operations. Data centres require professionals across electrical systems, power management, cooling technologies, facility operations, network support, server maintenance, and physical infrastructure management.

"At Apna, we are already seeing strong hiring momentum across infrastructure, operations and facility management roles, with employer demand consistently outpacing talent availability," Narayan states.

That demand is expected to create new opportunities for diploma holders and ITI graduates.

Narayan argues that workforce planning must evolve alongside infrastructure investments.

Landscape is Still Evolving

Gujarat hopes projects across Dholera and GIFT City can eventually evolve into a digital corridor serving hyperscale AI infrastructure. But Mahapatra believes that the outcome remains uncertain.

"They can become a genuine digital corridor only if Gujarat connects physical infrastructure with demand-side ecosystem development."

He says infrastructure alone cannot create a technology ecosystem.

"A corridor needs more than anchor projects. It needs cloud regions, AI startups, enterprise adoption, academic talent pipelines, data governance capability, cyber resilience, subsea or terrestrial fibre strength, and repeatable demand from banking, manufacturing, public sector, and digital-native firms," Mahapatra states.

Although Gujarat is investing heavily in AI-ready infrastructure, the state's AI ecosystem remains in its early stages. "The state has momentum on compute infrastructure, but the higher-value AI ecosystem is still emerging," Mahapatra comments.

Gujarat is also unlikely to compete directly with India's established technology hubs on software talent or startup density. Instead, Mahapatra believes it should pursue a different strategy.

"Gujarat can compete on cost, scale, land availability, renewable power, industrial infrastructure, and policy incentives, not primarily on software talent depth," Mahapatra says.

Its opportunity lies in becoming a preferred destination for AI infrastructure rather than application development.

"Gujarat's best path is to position itself as an AI infrastructure and industrial cloud hub for high-density compute, sovereign workloads, manufacturing AI, financial-services infrastructure, and energy-efficient data centres," he explains.

Ultimately, however, policy announcements alone will not determine success.

As Mahapatra concludes, "To win, it must reduce execution risk, prove uptime and connectivity, avoid overdependence on subsidies, and convert infrastructure announcements into a real customer ecosystem."