As executives from around the world gathered in Geneva this week to examine how to best confront the strain AI is putting on the electricity sector with tech and policy, they were greeted with news that flipped the script. During the opening plenary, speaker Varun Sivaram, the CEO of Silicon Valley-based Emerald AI, announced a new industry coalition with Nvidia and Google — the AI Energy Management Alliance — that aims to prove that AI data centers can dynamically manage their electricity use in response to grid conditions instead of drawing constant, unyielding power 24/7 and taxing the global grid.
The coalition, which brings together 20 companies and organizations spanning the AI and energy sectors, maintains that the technology can be a help rather than a hindrance: flexible AI data centers could unlock an additional 100 GW from the existing grid while reducing the interconnection bottleneck and that every 10% improvement in grid utilization could reduce utility rates by 3.4%.
“This could be a win-win and reoptimize energy systems,” says Espen Mehlum, head of energy at the World Economic Forum, which organized the September 16 energy conference in Geneva this week.
Analysts caution that the flexibility being promised needs to be measurable, predictable, and enforceable before grid operators can rely on it, but the technology developed by Emerald AI appears promising, The new Alliance is technology-neutral – it will embrace not just that company’s but any technology that can make AI data centers flexible power users, and champion policies that reward any data center that can measurably provide relief to the grid.
Other technology developments that could unlock additional capacity on the electric grid or make it more efficient were also showcased at the conference. UpLink, the World Economic Forum’s early-stage innovation engine, announced the 10 finalists of its Powering Tomorrow: Electricity Systems Challenge, a global startup competition. The winners, which were chosen with the help of members of the Forum’s Energy Technology Frontiers Future Council, include:
- * Splight, a U.S.-based grid-technology startup that uses AI and real-time data to double the usable transmission capacity of existing electrical grids without building new power lines.
- AssetCool, a UK deep-tech startup that makes metaphotonic coatings that are applied to existing overhead power lines by semi-autonomous robots. Those coatings cool the power lines and let them safely carry more current. The company says its technology offers grid operators up to 30% more line capacity, or up to 10% less energy loss, and can be delivered in weeks, rather than the years typically required for conventional infrastructure upgrades.” (see The Innovator’s Startup of the Week story)
- PlexiGrid, a Swedish and Spanish deep tech scale-up with roots in the University of Oviedo, which has developed AI software that boosts grid efficiency by utilizing integrated analytics to run the grid closer to its operational limits.
“We need to look at the $20 trillion in energy and industry assets deployed today, and leverage technology to use them in a smarter, more sustainable way,” says conference speaker Thomas Leurent, CEO of Akselos, a World Economic Forum Technology Pioneer that used physics-based AI digital twin technology to improve the structural performance of assets.
Getting The Public and Government On Board
But technology alone isn’t enough to bring about the energy transition.
During a workshop breakout session moderated by The Innovator’s Editor-in-Chief, the group consensus was that community support and government buy-in at the highest levels are crucial to overcoming global constraints shaping AI energy readiness. (Constraints include physical infrastructure and access to power; demand efficiency and operational flexibility; governance, permitting, and coordination supporting AI development; and finance, market, and investment conditions.”
Adrian Marcellus, head of the Forum’s Center for the Fourth Industrial Revolution Malaysia, who participated in the breakout session, explained how his country is mobilizing government and public support. Malaysia has built 60 data centers and is planning 70 more in the next two to three years. Today data centers use about 7% of the country’s energy supply, but that number is expected to jump to 31% in the next five to seven years. To ease the public’s concerns, the government first developed a clear-cut plan along with a strategy for communicating to the public why it is necessary for the country’s progress. A government task force was set up and a streamlined process adopted to ensure its work can advance quickly.
AI As Both Load and Lever
While other nations can learn from Malaysia’s example, each country will have to carefully calibrate its own path. AI infrastructure is creating new electricity demand at a pace that can exceed conventional energy and infrastructure planning cycles, says Ginelle Greene-Dewasmes, PhD, the Forum’s lead, artificial intelligence and energy. Readiness depends on more than aggregate power supply, she says. Where and when demand emerges, how AI infrastructure operates, and whether grids, investment, regulation, and communities are prepared can determine whether AI growth strengthens or strains wider energy systems.
“This creates a dual challenge of treating AI as both load and lever,” she says. “As a load, new demand must be balanced with affordability, reliability, energy security, and wider development needs. As a lever, AI growth can create opportunities to optimize existing capacity, accelerate grid modernization, unlock new energy investment, and strengthen the systems on which it depends.”
At the same time, the range of possible responses is evolving rapidly. Clean firm power, grid-enhancing technologies, storage, flexible AI infrastructure, and compute efficiency can reshape constraints and create new strategic choices, says Greene-Dewasmes. The challenge is not simply identifying what constrains AI growth, but determining which combination of innovations, infrastructure choices, and coordinated actions is most appropriate for a given context.
An Innovation Playbook And An Energy Readiness Compass
Earlier this year The Forum published an Innovation Playbook for Future Power Systems, which is meant to serve as a practical resource to show how innovation can deliver system-level progress. It combines real-world examples and insights that support planning, operations, and financing, highlighting best practices for large-scale renewable integration, greater flexibility, higher efficiency, and investment in resilient, digital infrastructure.
The report “lays out a framework and approach for building effective ecosystems for energy technology across digitization, AI, computer hardware technology, software tech, and platforms with the view that energy technologies need to continue to evolve to meet growing demand, security, reliability, affordability, and sustainability,” says Doug Arent, co-chair of the Forum’s Energy Technology Frontiers Future Council and a speaker at the Sept. 16 Energy Forum in Geneva.
“The Innovation Playbook was derived in part from discussions within the Global Future Council on Energy Technology Frontiers of successful approaches, such as the ecosystems built in the U.S. and in China, which effectively combine the public and private sectors,” says Arent, who is also a Global Fellow at the Center for Global Energy Policy at Columbia University and at the World Resources Institute.
The Forum’s Centre for AI Excellence is currently working on an AI Energy Readiness Compass as an additional way of helping ecosystem actors, ranging from governments and companies to civil society, to translate AI ambition to an energy-ready action plan, says Greene-Dewasmes. The compass is designed to support stakeholders in understanding their interdependencies, aligning priorities, and coordinating practical actions to actively steer toward competitive, resilient, and net-positive AI ecosystems.
A Call For New Types Of Collaboration
“There are no simple answers to the challenges the sector is facing,” Mehlum, the Forum’s head of energy, said in an interview (For the full interview, read the Q&A). “That is why we invited 170 people from 34 countries, including a big delegation from China, to grapple with similar challenges from different vantage points. We need new forms of cooperation — multilateral but also between the public and private sector — to solve security and sustainability challenges. The Energy Forum sparked ideas, and we will incorporate these insights into our upcoming initiatives and events
This article is content that would normally only be available to subscribers. Become a subscriber to see what you have been missing
