The artificial intelligence boom is driving a new wave of innovation in the energy sector, as governments and corporations race to meet the enormous increase in electricity demand expected from the rapid expansion of data centers.

Experts estimate that power demand from US data centers alone could surge by about 360% by 2030, reaching 110 gigawatts. Meeting this demand while maintaining affordable and sustainable energy is widely viewed as a major challenge that will require innovative solutions and significant technological advances.

There is no way to get there without a technological breakthrough, said Sam Altman, founder of OpenAI, during the 2024 World Economic Forum in Davos.

For Altman, the scale of the challenge is a direct argument for increasing investment in nuclear fusion research, which supporters believe could eventually provide an unlimited source of clean energy.

Alongside figures such as Sam Altman and Bill Gates, many Silicon Valley investors have spent years backing fusion technology. Those efforts are now beginning to show results as new startups enter the sector and technological breakthroughs attract growing interest from Wall Street, reviving research that had seen limited progress for decades.

Big Techs interest extends beyond nuclear fusion to other advanced energy technologies, including enhanced geothermal power and space-based solar energy.

However, the most intriguing innovation emerging at the intersection of artificial intelligence and energy is not about generating new electricity at all. Instead, it focuses on improving the distribution of existing power and making energy consumption more flexible.

Virtual power plants instead of building new power stations

This week, [Google](https://www.google.com?utm_source=chatgpt.com) signed an unprecedented agreement with [Voltus](https://www.voltus.co?utm_source=chatgpt.com) to create a virtual power plant.

Under the agreement, Google will fund a program within the Mid-Atlantic power grid that pays households and businesses to reduce electricity consumption during specific periods.

The companies say the arrangement will provide Google with 100 megawatts of power capacity without requiring the construction of any additional infrastructure.

Google thus becomes the first customer of Voltus Bring Your Own Capacity program, which allows energy-hungry companies to finance electricity demand flexibility among communities surrounding their data centers.

Voltus connects a wide range of devices into a single virtual network, including electric vehicles, smart thermostats, and other energy-connected equipment. Participating households and businesses receive compensation, while the company can manage energy flows and utilize stored power when needed.

According to Latitude Media, technology companies participating in the program effectively finance the creation of a virtual power plant in regions where they need to operate data centers, while Voltus delivers that capacity directly to utility providers.

The model is designed to help data centers bridge the anticipated energy gap through the early 2030s.

The current pilot project is the largest and first of its kind, and it is expected to provide valuable insight into whether energy flexibility can help address the rapidly growing electricity demands of data centers.

Making data centers themselves more flexible in their energy use will also be an important part of the solution. A study from Duke University last year found that reducing data center electricity consumption during peak-demand periods could enable the addition of roughly 100 gigawatts of new data center capacity without building new power plants or transmission lines.

However, that approach remains unpopular among AI companies because it could reduce revenue during periods when energy consumption is curtailed.

As a result, virtual power plants have become one of the most popular solutions currently under consideration. Instead of reducing their own electricity use, major technology companies can pay others to consume less power, allowing them to secure the energy capacity they need while avoiding disruptions to their operations.