India's next strategic asset: Sovereign intelligence

In the age of artificial intelligence, the most valuable asset is no longer simply data. It is the intelligence we can create from it and the ability to put that intelligence to work.

For the last few years, India's AI conversation has rightly centered on data sovereignty. Government policies such as data localization rules across sectors have been an important step in ensuring that Indian data is protected and remains within appropriate national and regulatory boundaries. As AI becomes more deeply embedded in our economy, institutions and everyday lives, the conversation is naturally broadening.

The question is not only where the data sits, but what we can create from it, how that intelligence is governed, and who ultimately benefits from it.

If Indian data is stored in India, but the models that learn from it are built elsewhere and the intelligence they produce is governed elsewhere, an important part of the value chain still sits outside the country.

Data sovereignty provides the foundation for something broader: India's ability to create, govern and benefit from the intelligence derived from its data.

That next layer of digital sovereignty is what I call sovereign intelligence.

From owning the model to building the capability 

Sovereign intelligence is not about India building everything on its own. It is about having the capability to build, adapt and govern AI in ways that reflect India's priorities.

AI leadership is not simply access to the latest models. It is also the ability to understand them, adapt them to local needs and deploy them responsibly at scale.

That requires trusted AI infrastructure, sovereign compute, high-quality datasets, responsible governance, cybersecurity, research and talent. It also requires the ability to evaluate AI systems and continuously improve them.

In short, AI leadership requires more than access to data. It requires the ability to turn data into intelligence and to do so at scale.

Why building for India matters 

This is where indigenous foundational models matter. Their significance is not simply that India can build an AI model. It is that India can build models that work for India.

India's linguistic diversity is immense. Connectivity constraints remain real in parts of the country, while public-service delivery operates at a scale few countries have to manage. Models trained and tuned for Indian languages, speech patterns and local contexts can enable voice-first access, improve citizen services, support vernacular education and bring AI to millions who have historically been less well served by English-centric digital systems.

These are use cases where developing and adapting technology to local context matters because the underlying needs are genuinely different.

Initiatives such as Bhashini, alongside a growing ecosystem of startups, research institutions and government-backed programmes, are already advancing these capabilities. India's scale and diversity create significant opportunities for further innovation.

There is a strategic dimension too. As AI becomes embedded in healthcare, education, finance and public administration, the ability to develop and adapt critical AI capabilities around India's own priorities becomes increasingly valuable.

Why India is well positioned 

India has many of the ingredients needed to build this capability at scale: one of the world's largest technology talent pools, a strong digital foundation and clear policy momentum through the IndiaAI Mission.

The compute backbone is scaling too. Deloitte projects India's data centre capacity to grow from around 1.5 GW in 2025 to 8–10 GW by 2030, driven largely by AI and the wider digital economy.

But AI capability is about more than infrastructure. It depends on talent, research and institutions that can bring these pieces together.

There is an economic opportunity as well. A stronger domestic AI ecosystem can create new companies, skilled jobs and research partnerships while allowing Indian businesses to capture more of the value created by technologies they already consume at scale.

According to EY, sovereign AI could add up to US$1 trillion to India's economy by 2035.

The real test 

Some may argue that sovereignty risks tipping into narrow self-sufficiency. That would be a mistake. The most effective approach is likely to be partnership-led, combining domestic capabilities with global innovation.

Model announcements are only the visible layer. The opportunity beneath them is to strengthen compute access, curate high-quality datasets, build robust evaluation benchmarks, strengthen cybersecurity and maintain the governance needed to make AI safe, trustworthy and accountable.

Several of these building blocks are already emerging. The national AIKosh platform, for example, hosts around 3,000 datasets and more than 200 models, according to EY[3], providing Indian innovators with a growing base to build on.

Indigenous models are therefore one part of a broader ecosystem that includes data governance, digital public infrastructure, research, talent and thoughtful policy design.

The opportunity now is to bring these pieces together and scale them.

From data sovereignty to sovereign intelligence

The next phase is not about choosing between domestic capability and global innovation. India can and should work with the best technology, talent and ideas from around the world.

It is about ensuring that India has the capability to shape, govern and adapt the technologies it increasingly depends on.

Data sovereignty provides greater control over an essential resource. The next layer is the ability to turn that resource into intelligence, innovation and economic value.

The opportunity is not simply to have India's data in India. It is to ensure that more of the intelligence created from that data can be built, trained, managed and valued in India.

That is the real promise of sovereign intelligence.

(Views are personal; Bhawna Agarwal is SVP and Managing Director of HPE, India)