For most people, artificial intelligence still appears as a tool. It writes emails, generates images, translates languages, summarizes documents, recommends products and answers questions. But for governments, militaries, corporations and intelligence agencies, AI is no longer merely a productivity tool. It is becoming a new infrastructure of power.
The country that controls advanced AI will not only build faster software. It may design stronger weapons, discover new medicines earlier, process intelligence faster, automate surveillance, dominate financial markets, shape public opinion, influence elections, command cyber operations and control the next layer of the global economy. This is why AI has moved from technology conferences to national security rooms.
The old symbols of power were territory, oil, aircraft carriers, nuclear weapons, industrial capacity and reserve currencies. Those symbols have not disappeared. But a new layer is being added above them: compute, chips, data, algorithms, cloud infrastructure, energy access and technical standards. A state may have a large population and a large economy, but if it depends completely on foreign chips, foreign cloud platforms, foreign AI models and foreign digital rules, its sovereignty becomes incomplete.
This is the central contradiction of AI geopolitics. Artificial intelligence is advertised as borderless, but its real power is highly concentrated. It looks like software, but it depends on hardware. It promises openness, but it is increasingly controlled by a handful of countries and corporations. It speaks the language of innovation, but it is now being treated as a strategic asset.
The world has entered an AI power race, and India cannot afford to watch it from the sidelines.
The Real AI Race Is Not About Chatbots
The public face of AI is generative AI: chatbots, image tools, coding assistants and voice systems. But the deeper contest is not about who builds the most impressive consumer app. It is about who controls the full AI stack.
That stack begins with advanced semiconductors. Modern AI systems require high-performance chips, especially GPUs and accelerators, to train and run large models. It then moves to data centres, cloud infrastructure, electricity supply, cooling systems, undersea cables, datasets, talent, frontier research labs, model architectures, security testing and deployment platforms. At the top are applications used in defence, healthcare, finance, education, logistics, governance and intelligence.
This is why AI has become a geopolitical issue. The United States, China, Europe, Japan, South Korea, Taiwan, the Gulf states and India are not only discussing AI ethics. They are building compute capacity, subsidising semiconductor ecosystems, regulating AI systems, funding national models and securing supply chains.
The scale of investment shows how serious this race has become. Stanford’s 2026 AI Index reports that US private AI investment reached $285.9 billion in 2025, more than 23 times China’s private AI investment of $12.4 billion. The same report notes that the United States had 1,953 newly funded AI companies in 2025, more than ten times the next closest country.
These numbers do not mean China is weak in AI. Private investment figures may understate China’s total AI push because state-directed funds, military-linked research, public procurement and industrial policy do not always appear in the same way as venture capital. But the numbers do show one clear fact: AI is now absorbing strategic capital at a scale normally associated with national infrastructure.
The AI race is therefore not just a business story. It is a power story.
Compute Is the New Oil
Every major technological era creates a new strategic resource. In the industrial age, coal and steel mattered. In the twentieth century, oil shaped wars, alliances and foreign policy. In the digital age, data became valuable. In the AI age, compute is becoming one of the most important foundations of power.
Compute means the processing capacity required to train and run AI systems. Without advanced chips and large-scale data centres, AI ambition remains a slogan. A country can have talented engineers, large datasets and strong policy documents, but without compute, it cannot compete at the frontier.
This is why semiconductors now sit at the centre of US-China rivalry. The US Commerce Department’s Bureau of Industry and Security has repeatedly tightened export controls to restrict China’s access to advanced computing chips and semiconductor manufacturing equipment. The stated purpose is not commercial protection alone; it is to prevent technologies such as AI from strengthening military applications.
This shift is historic. For decades, globalisation encouraged the free movement of technology across borders. Chips were designed in one country, manufactured in another, assembled elsewhere and sold globally. But AI has changed the political meaning of the semiconductor supply chain. Advanced chips are no longer seen as ordinary commercial goods. They are treated as strategic assets.
This has turned companies into geopolitical actors. A chip designer, a cloud provider, a foundry, a lithography-equipment supplier or an AI lab may now matter as much as a traditional defence contractor. Governments no longer ask only, “Who sells the product?” They ask: Who controls the architecture? Who controls the supply chain? Who can deny access during a crisis? Who writes the standards? Who owns the training data? Who can shut down the model?
That is why compute is the new oil. Not because it replaces energy, but because it becomes a gatekeeper of national capability.
AI Turns Data Into Strategic Leverage
Artificial intelligence needs data, but data is not merely a technical input. It is social knowledge converted into machine-readable form. It includes language, behaviour, transactions, location, health records, agricultural patterns, consumer habits, legal documents, satellite images, government records and cultural expressions.
This is where the fear of digital colonialism begins.
During the colonial era, raw materials moved from colonies to imperial centres. Value was added elsewhere, and the finished product returned at a higher price. In the AI era, a similar structure can emerge if developing countries provide users, data and markets while foreign companies control models, cloud infrastructure and monetisation.
The issue is not that foreign AI tools are automatically harmful. Many of them increase productivity and access. The problem is dependency. If a country’s schools, businesses, courts, hospitals, farms, media and government departments begin using AI systems trained elsewhere, governed elsewhere and priced elsewhere, then its public life becomes dependent on external digital infrastructure.
This dependency can shape language, culture and policy. AI models trained mainly on Western datasets may not understand Indian languages, local dialects, regional histories, caste realities, agricultural practices, legal context or social nuance with sufficient accuracy. If these systems become default decision-making tools, they may quietly reproduce foreign assumptions inside domestic governance.
AI sovereignty therefore requires more than data localisation. It requires domestic capability across compute, models, datasets, safety testing, regulation and public-sector deployment. It also requires linguistic inclusion. A country like India cannot treat AI merely as an English-speaking elite productivity layer. If AI does not work for Indian languages and ordinary citizens, it will deepen inequality instead of reducing it.
AI and National Security
Artificial intelligence is now entering the security domain in at least five ways.
First, AI improves intelligence processing. Modern states collect enormous volumes of satellite imagery, signals intelligence, cyber data, social media content and battlefield information. AI can help identify patterns faster than human analysts. In a crisis, speed itself becomes power.
Second, AI changes cyber warfare. It can help attackers write malicious code, identify vulnerabilities, generate phishing campaigns and automate reconnaissance. It can also help defenders detect anomalies and respond faster. The cyber battlefield is becoming more automated, and this raises the tempo of conflict.
Third, AI affects autonomous and semi-autonomous weapons. Drones, loitering munitions, robotic systems and sensor networks can become more effective when paired with AI. The danger is not only a future of fully autonomous weapons. Even partial automation can compress decision time and increase escalation risks.
Fourth, AI strengthens information warfare. Deepfakes, synthetic audio, fake images, automated propaganda and targeted persuasion can damage trust in institutions. In democracies, this is especially dangerous because public consent depends on shared reality. If citizens cannot distinguish fact from fabrication, elections become vulnerable to manipulation.
Fifth, AI can influence military logistics and planning. Wars are not won only by weapons. They are won by supply chains, maintenance, targeting, battlefield awareness and decision support. AI can help optimise these functions, giving technologically advanced militaries a significant edge.
This explains why AI governance cannot be left only to technology companies. The Bletchley Declaration of 2023 recognised that frontier AI creates safety risks requiring international cooperation, while the Council of Europe’s Framework Convention became the first legally binding international treaty seeking to align AI with human rights, democracy and the rule of law.
The challenge is that law moves slowly while technology moves quickly. States want safety, but they also want advantage. This tension will define AI geopolitics for the next decade.
Regulation Becomes a Form of Power
In the AI era, countries compete not only through technology but also through rules. Whoever writes the rules can shape markets.
The European Union has tried to position itself as a regulatory superpower. The EU AI Act entered into force on 1 August 2024 and becomes fully applicable on 2 August 2026, with some provisions applying earlier and some high-risk system rules following later timelines.
This matters globally because large companies often adapt their systems to comply with major markets. Just as Europe’s data protection rules influenced global privacy practices, its AI Act may influence how AI systems are classified, documented, audited and deployed.
The OECD AI Principles, adopted in 2019 and updated in 2024, also show how AI governance is becoming a diplomatic field. They promote trustworthy AI aligned with human rights, democratic values and practical policy guidance.
But regulation can cut both ways. Too little regulation may create unsafe systems, misinformation, discrimination and public distrust. Too much regulation may slow innovation and push companies to more permissive jurisdictions. This is the central policy dilemma: how to govern AI without surrendering the future to either corporate recklessness or bureaucratic paralysis.
For developing countries, the challenge is even sharper. If they simply copy Western frameworks, they may import rules designed for different economies. If they regulate too weakly, their citizens may become test subjects for powerful platforms. If they regulate too strongly, domestic startups may fail before they scale.
India needs a middle path: innovation with accountability, openness with sovereignty, and deployment with safeguards.
India’s AI Moment
India enters the AI age with important advantages. It has a large digital population, a strong software services base, deep engineering talent, a successful digital public infrastructure model, a growing startup ecosystem and massive public-sector use cases. India also has something many countries lack: scale. If AI can work in India’s complexity, it can work almost anywhere.
The IndiaAI Mission reflects this ambition. According to the Press Information Bureau, the mission involves more than ₹10,300 crore over five years and includes deployment of 38,000 GPUs. The same official note states that India’s tech and AI ecosystem employs around 6 million people, and that AI could add $1.7 trillion to India’s economy by 2035.
These figures show that India is no longer treating AI as a niche technology. It is becoming part of economic strategy, public service delivery and strategic autonomy.
But India’s challenges are serious. It does not yet control the most advanced semiconductor manufacturing stack. Its domestic AI model ecosystem is still developing. Many Indian startups depend on foreign cloud infrastructure. High-quality Indian language datasets remain uneven. Public-sector AI adoption is promising but must avoid surveillance excesses, exclusion errors and opaque decision-making.
India’s AI strategy must therefore answer six hard questions.
Can India build affordable compute capacity at national scale?Can Indian languages become central to AI development instead of afterthoughts?Can public datasets be opened safely without violating privacy?Can domestic firms build models that serve Indian needs rather than simply imitate Western products?Can AI be used in governance without creating algorithmic injustice?Can India protect strategic autonomy while still participating in global AI supply chains?
These questions are not technical alone. They are political, economic and constitutional.
AI and the Global South
The AI race may widen the gap between rich and poor countries. Advanced AI requires capital, compute, specialised talent, electricity, cooling, data infrastructure and regulatory capacity. Many developing countries lack these foundations. If AI becomes the next general-purpose technology, countries that cannot access it affordably may fall further behind.
The International Energy Agency estimates that data centres consumed around 415 TWh of electricity in 2024, about 1.5% of global electricity consumption, and projects that global data-centre electricity consumption could roughly double to around 945 TWh by 2030 in its base case.
This has major implications. AI power will not be distributed evenly because energy, land, water, chips and capital are not distributed evenly. The countries that host large AI data centres may gain new economic leverage. The countries that cannot provide reliable power may remain dependent on foreign compute.
For the Global South, AI can be both opportunity and trap. It can improve healthcare diagnosis, education access, disaster prediction, agricultural productivity and government service delivery. But it can also create dependency on imported platforms, drain local data, automate low-end service jobs and reinforce global hierarchies.
India has a potential leadership role here. It can present an alternative AI model: affordable, multilingual, development-oriented and sovereignty-conscious. Just as India’s digital public infrastructure attracted global attention, India could shape a Global South approach to AI that focuses on inclusion rather than only frontier competition.
But this will require execution, not slogans.
The Counter-View: Is the AI Race Overhyped?
There is a serious counter-view. AI may be powerful, but not every claim about it is realistic. Some productivity gains remain uneven. Many models still hallucinate. Deployment is expensive. Legal liability is unclear. Data quality is messy. Energy requirements are rising. Regulatory uncertainty is high. Many AI startups may fail. Some public-sector use cases may produce more harm than efficiency.
There is also a danger of strategic exaggeration. When governments describe every technology as a national security issue, they can justify protectionism, surveillance and censorship. AI should not become an excuse for techno-nationalism without democratic accountability.
The hype cycle matters. If countries overinvest in flashy AI projects but underinvest in education, healthcare, manufacturing, basic research and institutional reform, the result will be shallow technological nationalism. A country cannot become an AI power by announcing missions alone. It needs universities, research labs, patient capital, procurement reform, regulatory clarity, infrastructure and talent retention.
So yes, AI is sometimes exaggerated. But dismissing AI geopolitics would be a bigger mistake. The question is not whether every AI promise will come true. The question is whether enough of them will come true to reshape power. The answer is almost certainly yes.
What Happens Next
The next phase of AI geopolitics will be shaped by five battles.
The first battle is over chips. Export controls, semiconductor subsidies, Taiwan’s security, supply-chain diversification and domestic fabrication will decide who can access frontier compute.
The second battle is over models. Countries will increasingly ask whether they need sovereign AI models for defence, governance, education and critical infrastructure.
The third battle is over standards. Technical standards, safety benchmarks, audit rules and liability frameworks will shape how AI is deployed globally.
The fourth battle is over data. States will fight over data localisation, cross-border flows, public datasets, privacy and the rights of citizens whose data trains commercial models.
The fifth battle is over trust. Democracies must prove that AI can be used without destroying rights, jobs, privacy and public confidence.
For India, the right strategy is not isolation. India cannot build everything alone. It needs partnerships with the United States, Europe, Japan, Taiwan, South Korea, France, the Gulf and others. But partnership should not become dependency. India must participate in global AI networks while building domestic capacity where it matters most.
The goal should be strategic interdependence, not strategic vulnerability.
Conclusion: The New Test of Sovereignty
Artificial intelligence is becoming the new frontier of global power because it connects everything that matters: economy, defence, governance, information, industry, energy and diplomacy. It is not one sector. It is a force multiplier across sectors.
The countries that understand this early will shape the rules. The countries that delay will consume systems built by others. The difference between the two will not be measured only in technology exports. It will be measured in sovereignty.
India’s AI challenge is therefore not simply to build apps or celebrate startups. It is to build capability. Compute capability. Semiconductor capability. Data capability. Regulatory capability. Research capability. Security capability. Linguistic capability. Institutional capability.
The AI age will reward countries that combine imagination with infrastructure. It will punish countries that confuse digital usage with digital power.
The world is not entering a future where algorithms replace geopolitics. It is entering a future where geopolitics will increasingly be fought through algorithms.
And in that future, artificial intelligence will not merely be a tool of power. It will be one of the foundations on which power itself is built.