DC Edit | Regulate AI Use, Make Sure It’s All Productive
Currently, the United States leads this technology. It dominates the production of specialised chips, it is the leading developer of large language models (LLMs), and hosts the largest share of data centres — 33 per cent. While China is closely catching up, India, though a late entrant in the race, has become a serious player
Artificial intelligence (AI) is the most talked about topic in the world. Every country, every company and every forward-looking person wants to harness the power of AI for strategic and economic advantage. As a result, ChatGPT usage has quadrupled in just over one year from one billion queries a day in December 2024 to four billion queries a day in January 2026.
Currently, the United States leads this technology. It dominates the production of specialised chips, it is the leading developer of large language models (LLMs), and hosts the largest share of data centres — 33 per cent. While China is closely catching up, India, though a late entrant in the race, has become a serious player.
A viable AI ecosystem rests on three pillars — semiconductors as hardware, language models as software and abundant electricity to power energy-intensive data centres. Leadership will accrue only to those who control all three, compelling nations to compete aggressively for energy resources and capacity.
Among these, energy is likely to become the defining constraint. A simple AI query can consume about 0.33 watt-hours (Wh) of electricity. If the query is complex, the AI agent may need 33 watt-hour (Wh) power. Currently, the world needs 1.8 gigawatts of power per day to run AI. By the end of 2026, AI data centres would need 40 GW of power per day, which is 20 per cent of India’s daily average power demand. Apart from the shortage of power, the AI data centres generate higher amounts of heat, adding to the global warming burden.
The geopolitical dimension is already visible. As the United States disproportionately powers the global AI demand, US President Donald Trump’s aide Peter Navarro raised objections to American technology companies using America’s energy resources to meet global demand. However, power requirements will only intensify, as AI adoption grows further. This has prompted speculative ideas, including SpaceX chief executive officer Elon Musk’s suggestion of space-based data centres to tap large-scale solar energy. While space-borne data centres could solve the issue of power shortage, they will be the first target of enemy attacks during hostilities, making them high-risk assets.
Countries will, therefore, continue to invest in GPUs, LLMs and power production, regardless of environmental costs. If any country chooses to quit the AI race for ecology, it will be damned to remain an underdeveloped and vulnerable country for decades. However, countries cannot ignore the planet’s ecological limits and must strike a balance between technological development and ecology.
A sustained balance could be achieved by forcing companies to price AI services at their cost without subsidising them. Higher costs will curb the intemperate and unproductive use of the AI agents for role-playing, emotional advice, routine web searches, among others, and reserve computational resources for productive use.
Artificial intelligence is a new concept and the world is still unaware of issues that emerge out of it. As it matures, the world needs to regulate it to ensure that it is used for human development and not for frivolous purposes.