Top News

New Delhi: Energy-intensive artificial intelligence (AI) data centers are expected to drive a sharp rise in power consumption, with data center power usage projected to increase by 160 per cent by 2030, according to a report by Goldman Sachs. The report noted that after nearly a decade of flat power demand growth, the emergence of AI-driven data centers is set to transform the global energy landscape.

It stated, “Data center power usage is expected to increase +160 per cent by 2030, driven by energy-intensive AI data centers.” The report highlighted that power generation is only one part of the issue, as transmission remains a significant bottleneck in bringing new power plants online.

Citing the example of the United States, the report stated that most data centers are powered by natural gas due to its abundant supply.

However, challenges in permitting, transmission, and critical supply chains–such as those to gas turbines, have led to long timelines of around 5-7 years for getting new natural gas plants operational and connected to the grid.

The report also estimated that about 60 per cent of the data center demand growth will need to be met with new capacity.

This new capacity is likely to be powered by 30 per cent natural gas combined cycle gas turbine (CCGT), 30 per cent natural gas peaks, 27.5 per cent solar, and 12.5 per cent wind.

The report further stated that while natural gas remains a key source of power, renewables are expected to play a growing role as they offer faster and more efficient ways to secure incremental power supply amid the long development timelines for gas-based plants.

Hyperscale companies are adopting a mixed approach by combining different power sources to meet short-term energy needs. At the same time, they are investing in long-term solutions such as nuclear energy, though cautiously.

Tech giants have been willing to invest in power infrastructure but have avoided direct development risk or asset ownership. Instead, they are turning to strategies such as forward-start power purchase agreements (PPAs) to accelerate progress.

The report cited Alphabet’s recent agreement with Elementl Power to pre-position three sites for advanced nuclear energy as an example of this growing trend among major technology companies to secure sustainable and reliable energy for future AI operations.