AI, while often cited as a major driver of today’s energy demands, could also be seen as an answer to this challenge if used correctly.
That was the message from speakers at Schneider Electric’s Climate Action Week in London last week, where executives from Schneider Electric, Dell Technologies, Stack Infrastructure and Pure Data Centre Group discussed how AI is reshaping energy systems.
“We’re no longer debating targets or long-term intent,” David Hall, U.K. and Ireland zone president at Schneider Electric, which supplies infrastructure to support AI data centers, said in a keynote. “We’re in the phase where success is judged on delivery, on progress we can prove at scale.”
“The question has shifted from where do we want to go to how do we deliver it consistently [while] under pressure,” he added. “The challenge is to improve efficiency while demand continues to rise, to modernize assets that cannot afford downtime, and to do so visibly and transparently.”
AI as a ‘Climate Force’
Matthew Baines, vice president of secure power at Schneider Electric U.K. and Ireland, described AI as one of the most “climate-relevant forces of today.”
“We need energy for AI, but we also need AI for energy. They are two sides of the same coin,” he said.
The stakes are significant. U.K. data centers account for an estimated 2 to 3% of national electricity consumption, a figure that could reach 10% by 2030. Globally, operational data center capacity stands at roughly 100 gigawatts and is expected to double.
“The question now isn’t whether it will shape our energy system. It’s whether we shape it in the right way,” Baines added. “Will AI accelerate the energy transition or put it under more strain?”
Responding to that question, Arash Ghazanfari, former CTO of Dell Technologies UK, argued that AI’s impact will depend heavily on how organizations govern and deploy it.
“AI, by default, is not going to be a net positive contributor,” he said.
Instead, he said organizations need to focus on “right-sizing” AI deployments, ensuring they use the appropriate models and infrastructure for specific tasks rather than defaulting to the largest and most resource-intensive systems.
Ghazanfari also pointed to the rapid growth of data as a key challenge. As generative AI systems create increasing amounts of synthetic data, organizations will need stronger governance for data storage and duplication to avoid unnecessary increases in compute demand.
“We encourage experimentation and innovation,” he said. “But innovation without governance is a liability.”
Governance as Motivator or Hindrance
The panel also explored the tension between the need for governance and the fear that too much red tape will hinder development.
“Europe has an opportunity right now to catch up with the U.S. in terms of AI deployment,” Amy Daniell, senior vice president of strategy and development at Stack Infrastructure EMEA, said. “But by putting more governance into that process, it’s going to slow the whole thing down and we’re going to lose that chance.”
Ghazanfari argued that the debate should focus on “right-sized governance” rather than regulation versus innovation. Effective oversight, he said, should provide clarity and consistency without creating unnecessary barriers to progress.
Using AI to Optimize Energy Consumption
Despite its challenges, the panelists agreed that AI could play an important role in reducing energy consumption across data centers and electricity networks.
Ghazanfari described demand-side flexibility as one of the clearest opportunities for AI to contribute positively to climate goals. He pointed to Dell’s work using telemetry, digital twins and agentic AI systems to determine where workloads should run at any given time based on energy availability and operational efficiency.
“The only way for AI to start adding value to climate initiatives is through the use of AI and data to more intelligently place workloads,” he said.
Rather than running workloads in fixed locations, future systems could shift computing tasks among facilities depending on factors such as renewable energy availability, electricity demand and grid constraints.
Daniel added that AI could help infrastructure developers better understand grid vulnerabilities and identify opportunities to reduce pressure on electricity networks.
“If we can understand where infrastructure potential failure points are through a collaboration with grid operators, we’re able to work a lot more closely with them to alleviate that strain,” she said.
Emerging approaches, including microgrids, modular infrastructure and small modular nuclear reactors, were also highlighted as potential ways to support future AI demand without placing additional strain on public electricity systems.
Success with any of these approaches, however, will depend on cross-industry collaboration. While data center operators have historically developed projects independently, cooperation among operators, utilities, renewable energy developers and regulators will be needed to ensure AI growth does not outpace available energy capacity.
