Artificial intelligence has become one of the fastest-growing users of electricity on the planet. It is no longer just software running on machines. It is a global industrial system that must be powered every second. The future of AI and the future of energy are now directly linked. You cannot plan one without planning the other.

In 2024, data centres around the world consumed about 415 terawatt-hours of electricity — around 1.5 per cent of global power use. By 2030, demand could rise to 620–1,050 TWh, which would place the electricity used by AI and cloud systems close to the annual consumption of countries like Japan or Germany. This surge is driven mainly by AI.

Training the largest AI models requires 5 — 10 gigawatt-hours of electricity — the same energy needed to power 1,000 – 2,000 homes for an entire year. But the bigger long-term demand will come from inference —the billions of small interactions that happen every day when people use AI. Every search, every medical AI query, every factory prediction, every financial decision consumes electricity. By 2030, nearly 70 per cent of AI’s global energy demand will come from this

continuous usage. AI is now behaving like a heavy industry. It demands energy the way steel or cement does. Countries that treat AI only as a computing challenge—buy more chips, build larger servers — will fall behind. The next phase of AI progress will happen in countries with cheap, stable, scalable electricity built around AI’s needs.

This is why solar power is moving to the centre of economic strategy. It is not just a green ideal. It is becoming the cheapest and most reliable way to power intelligence. Solar has undergone one of the most dramatic cost drops in industrial history.

In the last decade, the cost of solar power has fallen by almost 90 per cent and battery storage by more than 80 per cent. Solar has a special property: every time global solar capacity doubles, costs drop by about 20 per cent. Fossil fuels and nuclear power do not follow this “learning curve,” but solar does because it is manufactured, not mined. The world currently has about 1.6 terawatts of installed solar capacity and this is expected to reach 3 TW by 2027. In sunny regions, new solar projects produce electricity at 2–3 cents per kilowatt-hour, which is cheaper than operating many coal plants. For data-centre operators — who live or die by energy costs — this shift is game-changing. Solar paired with modern batteries can cut the total energy costs of AI clusters by 30–50 per cent.

In other words, the economics of intelligence are becoming the economics of electricity. And the cheapest electricity increasingly comes from sunlight. This is where India enters the global picture with remarkable strength.

India’s Opportunity to Build the World’s First Solar-Powered Intelligence Economy

 India enjoys some of the highest solar exposure in the world, with 5.5–6.5 kWh/m²/day of sunlight across large regions.

This natural advantage is matched by rapid growth in solar capacity. India has already built 82 gigawatts of solar power and is on track to cross 110 GW by 2026. Even more crucially, India’s solar tariffs — Rs 2.2–2.5 per kWh —are among the lowest anywhere. Coal costs Rs 3.5–4.5 per kWh, and its price depends on imports and global volatility. This energy advantage is emerging just as India’s digital infrastructure is expanding at a historic pace. Data-centre electricity demand in India is expected to rise from 1.5 GW today to more than 5 GW by 2028. AI-specific facilities may require 1.2–1.5 GW, equal to the electricity used by a mid-sized city. Several states are especially well-positi-

oned. Rajasthan, Gujarat, Maharashtra, Karnataka and Telangana already have large solar parks, stable grids and pro-industry policies. These states can develop solar – AI corridors — zones where solar farms, advanced data centres, AI training clusters, semiconductor plants and high-performance computing facilities operate as a single ecosystem powered by sunlight.

Rajasthan alone has more than 140 GW of solar potential. Gujarat has some of the strongest solar radiation levels in Asia. Karnataka and Telangana host major cloud infrastructure. Maharashtra combines financial

and technical strength. Together, these regions can support the world’s first solar-backed cognitive economy — an industrial system where computation, not fossil fuel combustion, is the engine of growth. This shift gives India three major strategic advantages. The first is cost leadership. AI development will naturally move to locations where electricity is cheapest. India’s low-cost solar gives it a competitive edge over the US, China and Europe. No subsidy can match the power of structural, naturally low energy costs.

The second is energy resilience. Solar-powered data centres reduce India’s dependence on imported fuels and protect the economy from global price shocks. In a world where energy security shapes national power, this stability is invaluable.

The third is technological independence. AI is becoming as fundamental to national strength as telecom or defence. Finance, medicine, agriculture, logistics, governance — all will depend on AI. Controlling the energy that powers AI becomes a matter of sovereignty. India can build its AI future on domestic sunlight rather than

expensive imported fuels. But the shift goes deeper than economics or geopolitics. It represents a civilisational change in how human societies generate and use energy for intelligence.

For two centuries, industrial growth has been powered by fossil fuels — essentially ancient sunlight stored underground. AI represents the moment when we begin powering intelligence with today’s sunlight, captured in real time. The most advanced forms of machine cognition will be powered not by coal or oil, but by the star above us. The Sun sends Earth more than 10,000 times the energy humanity uses each year. The challenge is not sunlight itself, but building the infrastructure to convert it into dependable, round-the-clock electricity for an AI-driven economy.

Fortunately, AI’s rhythms align naturally with solar cycles. Large tasks like training checkpoints, simulations and batch processing can be scheduled during peak sunlight hours, when energy is most abundant

and cheapest. For Indian policymakers and industry leaders, the message is clear and urgent. AI strategy, energy strategy, semiconductor strategy and industrial strategy must now be treated as one integrated national plan. They cannot exist in separate silos.  India needs to accelerate solar-powered data-centre parks, modern grid systems, long-duration battery storage, renewable-based semiconductor fabrication units and university solar computing farms. It must also build secure, solar-driven AI networks for defence and national intelligence.

Ultimately, India’s AI ambitions will be limited not by talent or innovation, but by electricity. And electricity itself will depend on how rapidly India can expand low-cost solar capacity.

The defining idea of this coming era can be stated simply: The cost of intelligence is becoming the cost of electricity. And the cost of electricity is becoming the cost of sunlight.  If India acts decisively, it can lead this new global landscape — not only as an adopter of AI, but as a superpower in the energy systems that will run the world’s most advanced intelligence. India has the sunlight. India has the talent. And this is India’s moment to lead.

Author is a theoretical physicist at the University of North Carolina, United States; views are personal