AI power demand underestimated in current forecasts, 81% of energy executives say

More than 80 percent of energy industry decision-makers expect the rise of artificial intelligence to push electricity consumption beyond current projections, raising concerns about whether global infrastructure can keep pace with the rapid expansion of data centres.

A global survey commissioned by Adipec, the annual energy conference held in Abu Dhabi, found that 81 percent of energy insiders believe AI-driven electricity use will exceed existing forecasts. Of those, 32 percent anticipate demand will be significantly higher. Only 3 percent of respondents think consumption will fall short of current estimates, according to the Adipec Global Energy Report 2026, released on Thursday.

The findings come as technology companies accelerate investment in energy-intensive data centres, creating opportunities for power producers while adding strain to electricity networks.

The International Energy Agency expects data centres to consume about 950 terawatt-hours of electricity by 2030, roughly double the level seen in 2025, according to figures cited in the report.

Growing influence of technology firms

The expansion of AI is also expected to shift the balance of influence within the energy sector. About 43 percent of respondents believe technology and AI companies will gain the most influence over the industry, compared with 15 percent who selected governments and regulators. Financial institutions were also chosen by 15 percent, followed by international oil companies at 14 percent and national oil companies at 9 percent.

The findings reflect the increasingly close relationship between AI and energy, as technology companies become major drivers of investment in electricity generation and infrastructure.

For oil and gas companies, the shift is creating new commercial opportunities. While gas and liquefied natural gas growth currently ranks as the biggest commercial opportunity among oil and gas insiders, AI-driven efficiency is expected to take the top position over the next decade. About 27 percent identify AI-driven efficiency as a leading opportunity in 10 years, while 25 percent point to supplying power for AI and data centres.

The report also found that 81 percent of oil and gas insiders believe adopting AI will become essential to remaining profitable over the next decade.

UAE context and infrastructure challenges

The findings are particularly relevant to the UAE, where Abu Dhabi is developing a 5-gigawatt AI campus that will house the Stargate UAE computing cluster, expected to cost more than $30 billion.

However, a quarter of energy insiders believe the industry is unprepared for AI-driven electricity demand, with grid delays among the biggest obstacles. At the previous Adipec conference, Dr Sultan Al Jaber, Minister of Industry and Advanced Technology and managing director and group chief executive of Adnoc, called for massive energy investment to meet growing requirements.

During Abu Dhabi Sustainability Week in January, Dr Al Jaber said global electricity requirements from data centres could increase fivefold by 2040.

The report’s findings build on the “power is knowledge” message that was the theme of discussions at last year’s Adipec. The relationship between AI and electricity remains a central focus of this year’s event, which will take place at Adnec Centre Abu Dhabi from 2 to 5 November 2026.

Survey methodology

The Adipec report, conducted by global strategic communications consultancy Kekst CNC between August and October 2026, surveyed 1,832 energy industry decision-makers across 78 countries and included 17 executive interviews. Its findings reflect industry expectations rather than independently verified demand forecasts.