SUSTAINABILITY

Powering AI: Turning grid constraints into advantage

By: Celine Le Goazigo, Energy Lead at World Business Council for Sustainable Development

Editor’s note: As part of a special Climate Week NYC 2026 guest-contributor series on USA Stories, featuring voices from organizations working alongside Siemens to advance sustainability through collaboration and collective action, we asked Celine Le Goazigo, Energy Lead at World Business Council for Sustainable Development (WBCSD), to address key topics about powering the AI economy. (Siemens is a member of WBCSD.)

Electricity demand is entering a new era. After two decades of largely flat demand in many mature markets, the rapid expansion of artificial intelligence (AI) and data centers is creating sustained load growth. This shift is changing assumptions about grid capacity, energy infrastructure, and the pace of clean-power deployment.

AI growth is compelling utilities, policymakers, regulators, data-center operators, technology companies, and communities to reconsider how they plan, finance, connect, and operate energy infrastructure. The challenge extends beyond the amount of electricity AI infrastructure may require. It also involves the speed, geographic concentration, and investment intensity of that demand, along with the ability of grids and markets to respond quickly enough to support economic growth. The choices made now about where and how this infrastructure is built will shape not only its long-term carbon footprint, but also the resilience, affordability and competitiveness of the wider energy system.

Grid constraints do not have to become barriers to the AI economy. Addressed early and collaboratively, they can create an incentive for grid modernization, clean energy investment, digital deployment, and more flexible electricity use. Regions and companies that connect AI infrastructure development with energy-system planning can strengthen local capacity, reduce long-term energy risk, and establish a competitive advantage in attracting investment.

The key is to begin infrastructure planning alongside investment decisions. To realize that opportunity, stakeholders must treat the grid as a design partner and design AI data centers as active, flexible participants in the power system rather than passive points of consumption.

Despite hurdles including higher energy costs and grid congestion, electrification remains one of the most important levers for decarbonizing industry. Technical, economic, and regulatory barriers will continue to slow the deployment of electrified solutions unless key actors join forces to address them. Accelerating the energy transition therefore requires businesses, governments, utilities, technology companies, and grid operators to turn rising AI-driven demand into a shared opportunity for faster electrification, clean energy investment, and a more resilient and competitive energy system.

How AI data centers are reshaping energy demand and grid capacity

The return of sustained load growth is upending decades-old assumptions about grid capacity, pricing, and energy risk, with implications across sustainability, infrastructure, and economic development.

On sustainability, AI infrastructure moves the choice point earlier rather than later. Data centers built today will run for decades, and the sourcing and energy-flexibility decisions made now will influence a facility’s carbon profile for the lifetime of the asset. Integrating sustainability at the design stage, therefore, costs far less than retrofitting it later. It can also provide greater certainty around future energy costs, climate commitments, and access to clean power.

On energy infrastructure, large, creditworthy anchor loads can underwrite grid modernization that benefits the whole system. A substation upgrade or transmission reinforcement built for one operator can unlock capacity for the surrounding industrial base and for residential connections. Communities near new data-center development want confidence that added demand will strengthen local infrastructure while preserving their environment. Leaders who can connect data-center investment to a more resilient and affordable local grid can build support faster than those who address community impact separately from the infrastructure story.

On economic development, the scale of capital in motion proves how fiercely regions and countries are competing for investment. Capital expenditure among the 14 largest data-center operators globally is set to reach close to $750 billion this year, up from a little less than $450 billion last year. More than 23 gigawatts of data-center capacity was under construction worldwide as of late 2025, with about three quarters of it in the United States. Success will go to the regions that can connect projects quickly without compromising sustainability, jobs, affordability, or shared infrastructure benefits.

How leading companies combine AI growth with climate commitments and clean-energy investments

The pattern seen across leading companies confirms that electrification is not stalling because of technology, but because of gaps in value-chain coordination. We reached that conclusion after connecting companies across the data-center value chain through targeted dialogues and at WBCSD Annual Meeting 2026 in April 2026.

The organizations successfully balancing growing energy demand with climate commitments share several practices:

They begin substantive conversations with grid operators before a site decision is final, turning a potentially multi-year interconnection queue into a more manageable planning exercise.
They design for energy flexibility from day one, treating demand response and load-shifting capability as core facility specifications rather than features added later.
They pair long-term clean-power procurement with explicit grid-investment commitments, so the resulting agreement does more than move electrons on paper.
They look beyond national grid averages to locational data, because the carbon intensity and available capacity at a specific substation can differ substantially from the national picture used in much of today’s reporting.

The common thread is that these companies treat the grid as a design partner and the facility itself as an asset within the power system, not simply as a large point of consumption awaiting connection. This approach allows grid constraints to inform better siting, design, and operating decisions, turning energy planning into a source of business resilience and competitive differentiation.

To help accelerate progress, WBCSD is developing a set of practical resources: a brief for policymakers, drawing on insights from WBCSD’s Business Breakthrough Barometer 2026, and an electrification navigator, identifying priority actions for business and regulators to unlock barriers by 2030. Both are set for release in the second half of 2026.

These resources explore how businesses, policymakers, and regulators can create the conditions for a faster and more effective transition. By identifying shared priorities, practical actions, and opportunities for collaboration, they provide a roadmap for accelerating implementation, reducing barriers, and unlocking long-term economic, environmental, and societal value.

Building energy flexibility through grid and data-center cooperation among stakeholders

The most common gap we see in energy development is between infrastructure planners and infrastructure users. Utilities and grid operators plan on decade-long horizons, while AI data-center operators may make investment decisions on horizons measured in months. Closing this gap requires structured dialogue about how data centers can run as active grid assets.

There are multiple opportunities for alignment:

Shifting load in response to grid conditions:
Load shifting only works if utilities and grid operators can send data centers a signal about grid constraint that the facility's systems can act on automatically. This type of energy flexibility requires a partnership that should become a standard approach to reducing emissions and managing peak demand.
Rewarding flexible electricity use:
Policymakers and regulators need to set up incentives that make flexibility economically valuable. Utilities and operators must then co-design mechanisms that reflect how facilities operate and consume electricity.
Turning idle capacity into a grid asset:
Facilities, backup batteries, and other forms of spare capacity can potentially support the power system when they are not being used for their primary purpose. But to capture that value, regulators need to write rules that let companies deploy this capacity and get paid for it. This is one of the easier wins available in the near term.
Shifting computing workloads across regions:
Grid operators need to show clearly where capacity is tight and where it's available, using comparable capacity information across regions and markets. This approach requires cooperation among grid operators as well as collaboration with data-center companies.

Across all these opportunities, the clearest need is for policy that lowers the risk of making long-term commitments. Much of the hesitation comes down to who bears the cost if demand, market conditions, or infrastructure outcomes do not develop as planned. Clearer rules and incentives can allow businesses to invest in energy flexibility while giving utilities greater confidence in future demand.

As a first step, WBCSD is partnering with Climate TRACE on an online, interactive tool expected to launch by the end of 2026. The tool aims to help data-center owners and operators identify emissions-reduction opportunities across siting decisions, clean-power procurement, and flexible compute loads. This collaboration is part of WBCSD’s Emissions Reduction Accelerator, which aims to identify the highest-impact points across the value chain for collective action.

Turning AI demand into grid modernization and clean-energy investment

Three factors indicate that AI-driven electricity demand can support grid modernization while strengthening economic competitiveness.

The first is the scale of capital involved. Anchor loads of this size can justify grid investment and digital deployment that might otherwise take years to build a business case for. When planned effectively, this investment can support everyone connected to the same network. It can also make regions more attractive to manufacturers and other electricity users seeking reliable infrastructure, cleaner power, and room to expand.

The second factor is the demonstration effect. AI data centers that operate flexibly, shifting or shedding load in response to grid conditions, can prove at meaningful scale that flexible demand works both operationally and commercially. That evidence is important for the business case facing other sectors, from electric vehicles (EV)-charging networks to heavy industry, as they consider their own energy-flexibility investments.

Third is the urgency created by competition for AI investment. Regulators and industry are re-examining interconnection processes that move at a traditional utility pace because the economic cost of delay has become visible and immediate. They are now testing grid-enhancing technologies and faster permitting pathways that remained on policy wish lists for years because the pressure to connect new infrastructure has become concrete rather than theoretical. None of this will happen automatically. Progress depends on closing the coordination gaps described earlier. But the raw ingredients for acceleration, including capital, commercial proof points, technology, and policy attention, are present in a way they were not five years ago. The companies and regions that bring those ingredients together most effectively can turn infrastructure constraints into a source of resilience, investment attraction, and long-term economic advantage.

A new model for powering the AI economy

The rapid growth of the AI economy is testing assumptions that have shaped energy planning for decades. It is also creating a rare opportunity to advance changes the power system already needed. Capital investment, large anchor loads, growing policy attention, and real-world demonstrations of energy flexibility can help accelerate grid modernization, clean-power deployment, and infrastructure resilience.

The critical question is not whether AI growth will place new demands on the energy system. It is whether stakeholders will coordinate early enough to convert those demands into durable benefits for businesses, communities, and the grid.

Doing so will require a different model of planning. Data-center owners and operators must bring utilities and grid operators into site-selection discussions before decisions are final, design facilities for flexible operation, and integrate sustainability from the outset. Utilities, regulators, and policymakers must provide clearer signals about local capacity, reward energy flexibility, enable spare capacity to support the grid, and reduce the risks associated with long-term infrastructure commitments.

These actions can help ensure that faster connections do not come at the expense of cleaner power, affordability, or system resilience. More importantly, they can turn energy availability from a constraint managed late in the development process into a strategic consideration that guides investment, strengthens operating resilience, includes communities, and creates competitive advantage.

The frameworks established now will determine how much of this decade’s AI investment contributes to a stronger and cleaner energy system. Business leaders, policymakers, regulators, utilities, grid operators, technology companies, and AI data-center owners must act together to align investment, infrastructure planning, and climate ambition. By doing so, they can turn rising demand and grid constraints into catalysts for clean energy investment, economic growth, and a more resilient energy future.

Business calls to action for AI data centers, clean energy, and grid growth over the next decade

As business leaders head to Climate Week New York City 2026, our key message is that AI growth and the energy transition are not on opposing tracks. Sustainability does not have to be traded away when demand rises quickly.

AI growth and the energy transition only appear to be competing priorities when infrastructure planning happens after an investment decision rather than alongside it. Treated as one planning problem from the outset, AI-driven demand can become one of the strongest forces available for grid modernization and clean energy investment, both of which the energy transition already requires.

For businesses advancing energy development today, three calls to action stand out:

First, bring companies, utilities, and grid operators into site-selection conversations early, before a location is locked in.

Next, build energy flexibility and sustainability into facility design as core operational and commercial requirements.

Lastly, engage in the platforms shaping the enabling policy and regulatory environment because the frameworks regulators are establishing now around connection speed, locational transparency, infrastructure investment, and flexibility markets will determine how much of this decade’s AI investment translates into stronger, cleaner grids. COP31 in Antalya is an important milestone for this policy and advocacy work. Companies involved in these discussions can help shape rules that support faster connections while preserving clean-power objectives, system reliability, and shared economic value.

Published: September 9, 2026