India's energy jobs boom exposes the myth of AI-led job destruction

AI has been framed as a job-stealing force. The electricity sector tells a different story

Peter Drucker once warned that the real danger in periods of upheaval is not change itself, but the habit of thinking with yesterday’s logic. That warning fits today’s anxiety around artificial intelligence (AI) almost perfectly. AI is routinely described as a job-destroying force. Yet when one looks closely at where the global economy is actually adding jobs, a very different picture emerges — especially in energy. 

India sits at the centre of this contradiction. Public debate is increasingly dominated by fears of AI-driven unemployment, even as India’s energy sector is quietly generating work at one of the fastest rates in the world. The World Energy Employment 2025 report from the International Energy Agency (IEA) offers a useful reality check. In 2024, the global energy workforce reached 76 million people, growing 2.2 per cent — nearly twice the pace of overall global employment. More than five million energy jobs have been created since 2019, precisely during the years when AI adoption accelerated across industries. 

Electricity has now become the single largest employer in the global energy system, overtaking oil, gas and coal. India is among the countries where this growth is steepest. That fact alone should give pause to the popular claim that AI is steadily eroding human work. 

The fear persists because it tells a simple story: As machines become smarter, humans become redundant. But the electricity sector — the backbone of modern economic life — does not behave this way. Since 2019, electricity has been the fastest-growing source of energy employment worldwide. Solar power today employs more people than any other energy technology in history. Grid expansion, transmission lines, battery storage, nuclear maintenance, electric vehicle (EV) charging networks and renewable manufacturing have together created millions of jobs. 

In India, this shift is happening at exceptional speed. Energy employment grew by nearly 6 per cent in 2024, among the highest rates in major economies. Large solar parks in Rajasthan and Gujarat, new transmission corridors, battery gigafactories and grid modernisation projects are not abstract digital ventures. They are physical, site-specific, and labour-intensive. 

This distinction matters. Automating a customer support desk is relatively easy. Automating the installation of a 400-kilovolt transmission line, or the maintenance of a nuclear reactor, is not. Electricity systems are physical, distributed and unforgiving of error. They require judgement under uncertainty, safety-conscious decision-making and hands-on skill. AI can assist these processes, but it cannot replace them. 

What AI is doing instead is making human work more productive. Sensors embedded in power plants, turbines and substations generate constant streams of data. AI models analyse this information to predict failures and optimise performance — but prediction is not repair. When something breaks, people still show up, diagnose the problem and fix it. By reducing downtime and improving reliability, AI enables systems to scale faster — and scaling infrastructure means hiring more people, not fewer. 

The same logic applies to grid operations. AI tools help forecast demand spikes, detect faults, and improve load management. Yet human operators still rebalance flows, dispatch field crews and upgrade equipment. Even permitting and compliance are becoming faster with AI assistance, allowing projects to move from approval to construction sooner. The result is not job loss, but more work compressed into shorter timelines. 

Training is evolving alongside this expansion. Virtual reality and simulation tools now prepare linemen, nuclear technicians and maintenance crews for hazardous environments. Learning becomes safer and faster, but real-world construction and repair remain human tasks. 

The economic principle underneath all this is simple. When efficiency improves, sectors expand. When sectors expand, labour demand grows. The IEA data shows no evidence that AI has reduced employment in the electricity sector. What it shows instead is a growing shortage of workers — electricians, solar installers, grid technicians, battery specialists and engineers who understand both electrical systems and digital tools. 

This is where India’s real challenge lies. With over 500 gigawatts (Gw) of installed power capacity—more than half from non-fossil sources—India’s energy transition is among the most ambitious globally. Every additional Gw requires designers, factory workers, construction crews, inspectors and operators. The bottleneck is not demand for labour, it is the supply of trained workers. 

Coal-dependent regions illustrate both the risk and the opportunity. As India shifts towards cleaner energy, many technicians and operators can move into grid operations, storage and manufacturing — but only if reskilling pathways exist. That requires curriculum reform, regional training ecosystems and sustained investment in vocational education. AI will not replace India’s workers. But AI will replace India’s untrained workers. The future belongs to technicians who can read wiring diagrams and data dashboards, engineers fluent in turbines and algorithms, and operators who manage electrical parameters alongside predictive models. To capture this future, India must modernise Industrial Training Institutes and polytechnics, integrate AI tools into vocational training and build energy-skills hubs in transition districts. 

AI has been framed as a job-stealing force. The electricity sector tells a quieter, more grounded story. AI is not shrinking work, it is expanding what humans can build and maintain. Energy transitions are labour-intensive. Infrastructure-led growth is labour-intensive. India’s development story remains labour-intensive. The real scarcity is not work. It is a skill. 

The author is a theoretical physicist at the University of North Carolina at Chapel Hill. His forthcoming book is called Last Equation Before Silence