NO BLACKOUTS. A grid drawing closely on renewable energy can use AI to quickly fill demand-supply gaps
| Photo Credit:
Ravi Muchhapothula
There is {a photograph} that energy firms love to point out at conferences. A gleaming management room, banks of screens, operators watching dashboards fed by 1000’s of sensors throughout pipelines, wells and grids. Everything seen, all the things linked, all the things clever.
The KPMG Global Tech Report 2026, which surveyed 258 expertise leaders throughout the oil and fuel, mining, chemical substances, energy and utilities, and renewables sectors in 22 international locations, tells a extra difficult story. The screens are actual. The dashboards exist. But the information feeding them is usually incomplete, poorly ruled or trapped in legacy programs.
One in 5 energy firms is reaching a return of greater than 200 per cent on its expertise investments. The majority, 57 per cent, are at break-even. The cross-sector common for expertise ROI sits at 200 per cent. Energy lags, and has been lagging for a decade of heavy funding.
The pilot entice
Almost each massive energy firm has run AI pilots. Many have run dozens. Predictive upkeep. Production optimisation. Automated doc processing. These pilots usually work. They display worth. And then they cease.
Twenty-nine per cent of energy firms are nonetheless in the piloting section, operating AI tasks with out clear returns. Executives count on that to fall to 2 per cent inside a 12 months, a declare so assured it features as a check of whether or not the sector’s dedication to scale is actual or rhetorical.
The impediment isn’t AI. Around 60 per cent of energy executives say legacy programs are blocking full returns on their expertise investments. An oil refinery in-built the Eighties runs on management programs designed again then, sturdy and long-lived and wholly unequipped to share information with cloud platforms or machine studying fashions. Connecting them is sluggish, costly and operationally dangerous. Upgrades are deferred, and the hole between the trendy programs constructed on high and the ageing ones beneath — what one KPMG accomplice calls “digital debt” — retains widening.
India makes this seen at scale. The National Thermal Power Corporation and the Oil and Natural Gas Corporation run subtle digital programmes. The State distribution firms that really ship electrical energy to properties are nonetheless preventing unhealthy meter information, billing failures and grid losses, which in the worst-performing States exceed 1 / 4 of the energy generated. For them, the dialog about agentic AI just isn’t yet related. It is about whether or not the underlying information is dependable sufficient to construct on.
What the numbers disguise
The monetary returns that do exist come largely from the again workplace, finance, procurement, HR and compliance, the place generative AI accelerates document-heavy work. Over 50 per cent of energy organisations report that AI contributes 31–40 per cent of their whole monetary advantages. That is substantial, and a great distance from the story of AI altering how energy is produced and delivered, which stays, for many organisations, forward of them.
At the operational front-end, AI’s most categorical worth is in grid administration. A coal or fuel grid is predictable: Burn extra gas, get extra energy. A grid drawing closely on photo voltaic and wind should steadiness variable provide in opposition to variable demand throughout 1000’s of connection factors in actual time, at a pace past human operators. India’s goal of 500 GW of non-fossil gas capability by 2030, formalised in its Nationally Determined Contribution to the UNFCCC, will make this drawback acute. The technical case for AI-driven grid administration in India is as robust as elsewhere in the world. The implementation problem, given the grid infrastructure and information high quality throughout a lot of the nation, is correspondingly more durable.
Cutting corners
Nearly three-quarters of energy executives say that prioritising pace and cost-efficiency results in trade-offs in safety, scalability and information standardisation.
The penalties in energy are bodily. A cybersecurity hole in a pipeline invitations infrastructure disruption. A poorly ruled AI mannequin in grid administration can contribute to blackouts. Improved cybersecurity administration is the single most anticipated profit from expertise funding, ranked above income development and operational effectivity.
AI sharpens the risk on each side. It permits sooner, extra correct risk detection. It additionally places subtle assault instruments inside the attain of unhealthy actors. Intrusions into Indian grid infrastructure have been documented by cybersecurity researchers, although attribution in particular instances stays contested. As India’s grid turns into extra linked and AI-dependent, the assault floor expands. The hole between nationwide cybersecurity coverage and the operational actuality at smaller State utilities is consequential in the nation’s energy transition.
The lacking ability
Ninety-six per cent of energy leaders imagine that managing AI brokers can be a key workforce ability inside 5 years. The present era of AI instruments suggest; the subsequent will act. An agentic system is not going to recommend adjusting a valve — it can modify it. Managing that shift requires individuals who perceive what these programs are doing, when to override them, and the place human judgment stays indispensable.
India has engineering depth and a rising information science neighborhood. What it lacks, at scale, is the overlap: People who perceive each the physics of energy programs and the structure of AI. Shreyansh Upadhyay, KPMG India’s AI for Energy chair, places it plainly. The greatest problem is in constructing fashions which can be context-aware and grounded in the bodily behaviour of precise equipment, not simply historic information patterns. That mixture is tough to rent and more durable to construct shortly.
Published on July 13, 2026