Big Tech Cloud Infra is Breaking the Illusion of India's Sovereign AI Ambitions

In the recent Amazon Prime docuseries, Make in India: A Titan Story, a narrative about entrepreneurial optimism and indigenous capability building, when Tata brought advanced quartz watchmaking technology to Tamil Nadu, the objective was not merely to manufacture watches in India. It was to transfer know-how, create a skilled workforce, build local supplier ecosystems, and ensure that India could eventually own a greater share of the value chain. Over time, what began as technology adoption evolved into technological capability.

Looks like we are at the same crossroads again.

India’s AI and digital infrastructure ecosystem is expanding through large-scale collaborations between global technology companies and the Indian government. The government is looking at such collaborations through a sovereignty lens, aiming to strengthen cloud infrastructure, AI capability building, and digital public services.

India’s AI buildout is shaped by partnerships with global hyperscalers, including Google Cloud, Microsoft, Amazon Web Services, and IBM. These collaborations span AI-powered governance systems, cloud migration for public sector workloads, digital skilling programmes, and startup acceleration initiatives.

However, the idea of attaining technological independence through Big Tech collaborations itself is a paradox, and the goal of creating sovereign AI seems steep and distant.

Public Sector Engagement

Among the Big Tech collaborations, Google Cloud has been actively working with central and state agencies across areas such as AI-powered governance, digital skilling initiatives, and cloud adoption for public sector workloads, while also supporting startup acceleration and data infrastructure programmes through its India-focused digital transformation efforts. 

Earlier this year, Madhya Pradesh Chief Minister Mohan Yadav met senior officials from Google Cloud India to explore collaboration opportunities focused on deploying artificial intelligence-driven solutions to improve administrative efficiency and support preparations for Simhastha 2028, a sacred Hindu congregation. According to Neeraj Mandloi, the Additional Chief Secretary to the Chief Minister, the Madhya Pradesh government is also preparing to finalise a significant agreement with Google Cloud India to advance these digital transformation initiatives.

Similarly, Microsoft is working with government agencies on AI skilling, cloud-first governance frameworks, and data residency-compliant infrastructure, while Amazon Web Services is supporting secure cloud migration, AI-driven citizen service platforms, and infrastructure modernisation efforts. 

IBM, meanwhile, continues to contribute through AI research partnerships, hybrid cloud deployments, and public sector skill development programmes.

Sovereignty Debate

Strategically, India faces a build-versus-buy dilemma. Sovereign infrastructure offers control over data and strategic autonomy but demands massive capital, and risks slower iteration cycles and underutilisation. On the other hand, reliance on hyperscalers enables rapid scaling and access to cutting-edge technology but increases vendor lock-in and external pricing control. 

In practice, India is currently optimising for speed and scale rather than full sovereignty.

While India’s public cloud market is projected to cross $25–30 billion by 2030, per International Data Corporation, hyperscalers already control nearly 70–80% of enterprise cloud workloads in the country. At the same time, India hosts 2,800 global capability centres (GCCs) and faces an estimated 40–50% GPU supply-demand gap in the 2026–2028 cycle, highlighting significant infrastructure constraints.

Against this backdrop, experts argue that sovereignty in AI must go beyond physical infrastructure ownership.

Srinivas Varadarajan, Founder and CEO of deeptech company Vigyanlabs, notes that India’s AI infrastructure is still largely dependent on Big Tech. “This could impact sovereign capability unless we build our own AI infrastructure. A few of us are already working on developing sovereign AI infrastructure in India. What is needed now is stronger support for Indian startups from the government, including procurement of technology from domestic vendors,” he tells AIM.

Meanwhile, geographies, including the European Union, are strategically shrugging off overdependence on US Big Tech. Recently, the EU laid out plans in the European Technological Sovereignty Package to enhance digital autonomy with homegrown cloud services and accelerate the construction of data centres.

Public-Private Partnership as the Emerging Model

However, building AI infrastructure is inherently complex and requires deep global expertise and significant capital intensity. 

“Hyperscalers such as Google are actively investing in this space primarily to serve their own enterprise clients, and therefore play a central role in large-scale infrastructure development,” remarks Yugal Joshi, Partner at Everest Group.

However, “these global players typically work in collaboration with Indian partners such as Adani and Bharti that provide critical supporting infrastructure, including land, connectivity, and co-location facilities,” he adds.

The middleground can be a public-private partnership (PPP) model, where the private sector focuses on building applications, enterprise tools, and domain-specific AI models, while government departments take charge when it comes to AI-led public delivery systems. 

Such PPP models have borne fruit with initiatives such as Aadhaar, UPI, DigiLocker, CoWIN, and ONDC building interoperable national digital public goods while enabling startups, enterprises, and institutions to innovate on top of these foundational layers.

However, this also reflects a broader contradiction within the Atmanirbhar Bharat framework, where AI development is inherently global, with chips, models, and frameworks concentrated across a few geographies, making sovereignty more about negotiating within ecosystems than complete technological independence.

Structural Contradictions

India stands at a fork in the road with three possible AI pursuits. 

Currently, the ecosystem is operating on a managed dependency model dominated by global hyperscalers, with India focusing on regulation and applications. There is an emerging hybrid sovereign stack combining government compute pools and Indian neoclouds alongside global partners. In the long term, for India to attain complete autonomy, developing a full domestic AI infrastructure capability would require significant state and private investment. 

However, even then, the government must manage strategic dependencies. Hyperscalers and infrastructure partners are the clear near-term winners, application-layer startups and GCCs benefit from expansion, while domestic startups and legacy IT models face structural pressures.

Joshi also highlights that startups are largely excluded from such capital-heavy sovereign infrastructure buildouts, as these initiatives do not align with their business models or strategic priorities. Instead, they tend to focus on AI applications rather than foundational infrastructure. While companies like Yotta Data Services and NxtGen Cloud Technologies are attempting to build capabilities, they are facing challenges related to capital access and procurement of advanced GPU infrastructure.

On the broader policy perspective, he emphasises that this trend should not be viewed as contradicting the idea of Atmanirbhar Bharat. “India didn't want to isolate itself from global innovation. It was more about self-reliance and determination and capability to build, run, and use technology without hindrance from external entities,” Joshi says.

Joshi further notes that dependence will always exist in global technology supply chains, but the key is ensuring that such dependence does not translate into external control over India’s technology direction. He also stresses the need for greater investment in startups to enable them to scale and compete globally, drawing parallels with how early policy support helped build India’s IT services industry.

Sovereignty Reimagined

Rahul Agarwalla, Managing Partner at SenseAI Ventures, further notes that AI infrastructure development requires substantial capital and operational expertise, which is currently concentrated among a small set of global technology companies, making their involvement necessary in the near term.

On sovereign AI, he explains that ownership of data centres alone does not define sovereignty. “The more important questions are who controls the data, who owns the applications, and which laws govern both.” 

While infrastructure located in India and governed by Indian regulations represents a meaningful step forward, true sovereignty will require a gradual expansion of ownership across the AI stack. Addressing the constraints faced by domestic firms, including capital intensity and technical complexity, remains central to the sovereign push.

Building large-scale AI infrastructure demands sustained investment, access to advanced technologies, and experience in managing high-end compute environments. While Indian firms are actively investing in this space, Joshi observes that scaling to global levels will take time.

India’s AI strategy is not about owning the entire stack—it is about controlling outcomes while sharing infrastructure dependency. This creates a hybrid sovereignty model, rather than attaining absolute autonomy.