.png)
Gurumurthy, ex-central banker and a Wharton alum, managed the rupee and forex reserves, government debt and played a key role in drafting India's Financial Stability Reports.
March 10, 2026 at 1:03 PM IST
For much of the internet age, technology companies promoted a powerful idea: the digital world had no borders. Data could flow seamlessly across continents, cloud computing made location appear irrelevant and software companies could operate globally without being anchored to geography.
Artificial intelligence is rapidly challenging that assumption.
As AI systems move into finance, defence and public infrastructure, governments are asking uncomfortable questions: Where does the intelligence actually reside? Who controls the data that trains it? And in moments of crisis, whose rules prevail?
Recent debates involving companies such as Anthropic and OpenAI over defence collaborations with the US government illustrate how AI is increasingly entangled with national security. Meanwhile, financial regulators are discovering the limits of digital sovereignty. Following security concerns and drone attacks in the Gulf, the Central Bank of the United Arab Emirates reportedly allowed temporary flexibility in strict data-residency rules for financial institutions to maintain operational continuity.
These developments signal a broader shift: artificial intelligence is no longer merely a technological innovation. It is becoming strategic infrastructure.
And like all infrastructure, it is shaped by geography, politics and increasingly, energy.
Modern AI systems rely on enormous computational power and vast datasets. Training advanced models can require thousands of processors running continuously for weeks inside hyperscale data centres owned by a handful of global technology firms.
These facilities have quietly become the industrial backbone of the digital economy.
For governments, the implications are significant. When critical data and AI capabilities are hosted abroad, questions arise about regulatory oversight, national security and economic dependence.
Many countries have responded with data localisation policies, requiring sensitive information, particularly financial, health or strategic data, to be stored within national borders. Initially framed as privacy protections, these rules increasingly reflect a broader ambition: digital sovereignty.
Yet the global cloud system was designed to distribute computing across borders for efficiency and resilience. Restricting data flows may strengthen sovereignty but can reduce flexibility during crises.
The UAE episode illustrates this tension. When infrastructure risks disruption, regulators may temporarily relax localisation rules to keep financial systems functioning, even if that means relying on overseas data centres.
Digital sovereignty, in practice, can collide with operational necessity.
Behind debates about AI regulation lies a more fundamental issue: where the computing power actually resides.
Data centres concentrate the core resources of the AI economy i.e., computing capacity, vast datasets and the ability to train and deploy advanced algorithms. Because of this concentration, they are increasingly viewed as strategic assets.
Governments are competing to attract data-centre investments through tax incentives, regulatory support and energy subsidies. Hosting AI infrastructure promises economic benefits, from high-skilled jobs to wider digital ecosystems.
But this competition faces a critical constraint: Energy.
Artificial intelligence is not only computationally complex; it is extremely energy-intensive.
Training large models requires immense electricity, while data centres must also run sophisticated cooling systems to prevent overheating. Hyperscale facilities can consume hundreds of megawatts of power—comparable to the electricity demand of a medium-sized city.
As AI adoption expands, this energy footprint is expected to grow rapidly.
This creates an unexpected intersection between digital policy and energy policy. Countries aspiring to become AI hubs must ensure reliable electricity supplies even as they pursue ambitious climate goals.
Communities increasingly question why scarce electricity or water resources should be devoted to massive server facilities, while environmental groups warn that AI’s energy demands could complicate efforts to decarbonise power systems.
The cloud, despite its ethereal name, carries a very tangible carbon footprint.
These pressures are reshaping the global technology landscape.
Energy availability, grid stability and access to renewable power are becoming decisive factors in determining where new AI infrastructure is built. Regions with abundant energy, whether hydroelectric power, natural gas or renewable capacity, may gain an advantage in hosting the next generation of data centres.
This adds a new dimension to geopolitical competition around artificial intelligence.
The twentieth century saw technological rivalries centred on oil, nuclear power and semiconductors. In the twenty-first century, competition may increasingly revolve around computing infrastructure and the electricity required to sustain it.
AI systems may appear intangible, but their existence rests on physical foundations: land, hardware, energy and legal jurisdiction.
The End of a Borderless Dream
The early internet fostered the belief that digital technologies would transcend national boundaries. Artificial intelligence is exposing the limits of that vision.
Technology companies still operate globally, but their data centres sit within specific jurisdictions and depend on local energy grids. Their operations are shaped by national regulation and increasingly by national security concerns.
Artificial intelligence may look like pure software. Data is boasted as the future business for many IT companies that were at the receiving end recently.
More From BasisPoint
