Artificial intelligence is no longer just a tool for chatbots, coding assistants or online shopping suggestions. It has transformed into a matter of national strength, economic survival and strategic control.
The phrase increasingly heard in policy circles is sovereign AI or sovereign cloud. It may sound technical, but the idea behind it is simple: countries want control over their own artificial intelligence systems, data and computing infrastructure.
Across the world, nations are asking the same question -- who controls the intelligence that runs modern society?
What is sovereign AI?
Sovereign AI refers to artificial intelligence systems that are developed, trained and governed within a country’s own legal and regulatory framework. It usually involves local data storage, domestic control of cloud infrastructure, and AI models that are subject to national laws rather than foreign oversight.
A related term -- sovereign cloud -- means that critical data and computing systems are hosted within national borders and controlled by domestic entities.
Debjani Ghosh, Distinguished Fellow at NITI Aayog, described the shift clearly. “Earlier, we used to say technology was an enabler of power. Today, AI is an axis of power,” she said. In other words, AI is no longer just supporting governments and businesses. It is shaping power itself.
But sovereignty does not simply mean where servers are placed. Ghosh explained: “It’s not about where data is stored. It’s about who controls the data, workflows and supply chains.” Control, rather than geography, is at the centre of the debate.
Why the urgency now
Artificial intelligence now influences defence systems, financial markets, healthcare diagnostics, education platforms and public services. As its role expands, so do the risks of dependency. Global tensions between major powers, especially the United States and China have exposed vulnerabilities in supply chains, semiconductor production and cloud infrastructure. Countries fear that in a crisis, access to critical AI tools could be restricted. “You have to ensure that no one can cut off supply because of geopolitical tensions,” Ghosh said.
The US-China technology rivalry, chip export controls and the dominance of a handful of global cloud companies have intensified these concerns. Artificial intelligence depends heavily on advanced semiconductors and large-scale computing power, both concentrated in a few countries. This concentration has triggered what some describe as a sovereignty instinct.
Privacy or power?
At first glance, sovereign AI may appear to be about protecting citizens’ data. Data localisation laws and domestic hosting are often framed as privacy measures. But the issue goes beyond privacy. It is also about strategic autonomy.
Mukesh D Ambani, Chairman of Reliance Industries, said at the India AI Impact Summit 2026: “India cannot afford to rent intelligence.” His warning reflects a wider concern that countries relying entirely on imported AI systems may lose control over critical decisions. If AI drives economic output, defence systems and governance tools, then dependence on foreign platforms becomes a strategic weakness.
Arthur Mensch, CEO of Mistral AI, also cautioned against excessive concentration. “In a world where large portions of GDP will be produced by AI, those running AI workloads must have access to the on-and-off button,” he said. “Ownership is not about isolation… Partnerships are essential. But dependency is not.”
Who is pushing sovereign models
The push for AI sovereignty is not limited to one country. Across continents, the message is similar that AI is too important to be left entirely in foreign hands. The European Union has emphasised digital sovereignty through regulation and domestic capacity building. France and Germany have invested in local AI labs and cloud infrastructure to reduce reliance on American technology giants.
China has pursued technological self-reliance for years, building domestic chip production and AI platforms to reduce exposure to foreign sanctions. India has taken a different path, seeking to build domestic capabilities while staying connected to global markets.
Even the US, often seen as the dominant force in AI, is framing its partnerships in sovereign terms. At the summit in New Delhi, Assistant to the President and Director of the White House Office of Science and Technology Policy Michael Kratsios said: “Real AI sovereignty means owning and using best-in-class technology for the benefit of your people, and charting your national destiny in the midst of global transformations.” He encouraged countries to pursue autonomy without full technological isolation.
Britain has also entered the debate. The UK established a sovereign AI unit backed by 500 million euros. Former prime minister Rishi Sunak argued that technological revolutions do not always reward the inventors most. Regulatory and institutional choices matter.
The trigger: Chips and cloud dominance
Globally, artificial intelligence funding has crossed $473 billion. A large share of this capital is concentrated in a few firms. OpenAI, Anthropic and xAI together account for around $170 billion which is about 36% of global funding. Such concentration worries policymakers.
AI systems require vast computing power, often delivered through global cloud providers. Semiconductor manufacturing is even more concentrated, with advanced chips produced in limited locations.
When the US restricted exports of advanced chips to China, it showed how quickly access could be curtailed. For many governments, this was proof that digital infrastructure can become a geopolitical weapon.
The economic stakes
Control over AI means control over data, infrastructure and economic growth. AI is expected to reshape industries from agriculture to pharmaceuticals.
India’s AI sector has raised more than $5.5 billion across over 1,700 firms, according to a report by Tracxn titled India and the Sovereign AI Shift. As of 2026, India is home to more than 1,700 AI-native companies working across enterprise software, consumer platforms, sector-specific tools and infrastructure.
Funding peaked at $1.1 billion in 2022, slowed during a global downturn, then recovered to $856 million in 2025. In 2026 so far, it has already reached $626 million.
Large infrastructure rounds are emerging. Neysa AI raised $600 million at an enterprise valuation of around $1.4 billion, reflecting growing interest in AI cloud and GPU-backed systems.
The Indian government’s Rs 10,372 crore IndiaAI Mission has allocated GPU computing power to 12 companies building foundational models. More than Rs 100 crore has already been disbursed as subsidies for high-performance GPUs.
Sarvam AI received 4,096 NVIDIA H100 GPUs and around Rs 99 crore in compute subsidies. It launched two large language models — a 30-billion-parameter model and a 105-billion-parameter model — both trained from scratch in India. “Sovereignty matters much more in AI than building the biggest models,” said Sarvam co-founder Vivek Raghavan.
Beyond startups, major corporations are investing heavily. The Adani Group has committed $100 billion towards infrastructure, including energy and cloud capacity. Jeet Adani outlined three pillars: energy, compute and cloud, and services. “AI is written in code. But it runs on electricity... energy security is intelligence security,” he said.
Dell Technologies unveiled its ‘AI India Blueprint’, calling for a national compute strategy, energy-efficient data centres and workforce development. Industry estimates suggest AI workloads in India could grow at 30 per cent annually until 2030, with compute demand reaching 12–15 exaflops.
Can countries build fully sovereign AI?
Despite ambition, building a fully sovereign AI system is extremely difficult. AI requires advanced chips, massive data centres, skilled researchers and billions in capital. Even wealthy nations struggle to produce cutting-edge semiconductors independently. Kratsios suggested full self-sufficiency may not be realistic. He encouraged countries to integrate components of the American AI stack rather than aim for complete independence.
Ghosh also stressed alliances. “It’s not just about building everything indigenously. It’s also about alliances and trusted supply chains.” The reality is that AI supply chains are global. Chips may be designed in one country, manufactured in another and deployed through cloud servers in a third.
Risks of going fully sovereign
Pursuing strict digital sovereignty carries risks. First is fragmentation. If countries build closed AI ecosystems with incompatible standards, the global internet could split into separate blocs.
Second is higher costs. Domestic production of chips and infrastructure may be more expensive than relying on global
markets. Third is slower innovation. Open research collaboration has driven AI progress. Excessive restrictions could limit knowledge exchange. And fourth is regulatory overreach. Overly rigid governance may discourage startups and private investment.
Arthur Mensch warned of “excessive leverage” if only three or four companies dominate AI. Yet, replacing one concentration with multiple isolated systems may not solve the problem.
Where does India stand
The country is building domestic AI capacity while maintaining global partnerships. Global firms are expanding in India. Google and Adani announced a $15 billion data centre plan. Amazon Web Services committed $8.4 billion towards infrastructure.
At the same time, India is promoting open-source development and multilingual AI tools. Bengaluru-based Sarvam AI, Gnani.ai and the IIT-Bombay-led BharatGen consortium unveiled new models at the India AI Impact Summit. BharatGen’s 17-billion-parameter multilingual model will be released open-source, enabling developers to build India-centric applications.
India’s digital public infrastructure, Aadhaar, UPI, DigiLocker, ONDC and Bhashini, provides a strong base. These platforms handle identity, payments, commerce and language services at population scale.
India is also positioning itself as a voice for the Global South. Lammy praised India for “centering the Global South” in AI discussions, noting the opportunities in agriculture and rural communities.
Future or political slogan?
Funding flows, government missions and infrastructure commitments show serious intent. Yet, complete technological independence is unlikely. Hybrid models are emerging, domestic data control combined with international collaboration.
AI may develop into technology blocs aligned with geopolitical alliances, rather than a single open global system. The coming decade will test whether sovereignty can coexist with innovation or whether the world’s intelligence systems will fragment along political lines.
One thing is certain, artificial intelligence is no longer just about algorithms. It is about who holds the switch.