Seyed Madaeni is CEO and co-founder of Verse, an energy software company specializing in battery dispatch optimization for data centers.
Even the best-funded project can’t operate without electricity, and the grid often can’t deliver hundreds of new megawatts as quickly as AI developers need them. Access to power has become a competitive advantage, and companies are searching for ways to get it faster.
Today, developers see two obvious options. They can wait years for traditional grid upgrades, or they can build private power plants onsite. The first can move too slowly. The second brings fuel-supply, equipment, permitting and community challenges of its own.
A third path to data center interconnection is taking shape. Across the country, utilities and grid operators are creating clearer ways for large loads that can reduce the risks they bring to the system. Depending on the market, a data center might guarantee a minimum payment, agree to reduce its grid demand when the system is strained, provide better operating data or contract for new generation and storage.
These policies differ, but they share one idea: If a data center wants faster or more certain access to power, it should give the grid greater certainty in return.
That is a better approach than advancing projects mainly because they promise jobs or investment. Projects that make their demand more predictable, controllable and financially secure should have a clearer path to service.
Ohio is asking data centers to pay for what they reserve
Utilities often need to invest in substations, transmission equipment and other infrastructure years before a data center reaches full operation. They are building against forecasts that can change, and their queues may include projects that never get built.
Ohio is addressing this problem through financial commitments for large loads. After AEP Ohio proposed a data-center-specific tariff, Ohio regulators approved a settlement requiring large customers to pay for up to 85% of their subscribed capacity, even when they use less. The agreement also includes longer contract terms and additional collateral requirements. In exchange, the connection moratorium was lifted and developers received a defined framework for service.
Another utility might choose a different percentage or structure. But a company asking the grid to reserve an extraordinary amount of power should be willing to stand behind that request financially. This protects existing customers from paying for infrastructure built around demand that does not materialize. It also helps serious projects distinguish themselves from speculative ones.
Texas and SPP want more predictable operations
Other markets are focused on how data centers behave once they connect.
A large facility can draw hundreds of megawatts, change its consumption quickly or disconnect during a grid disturbance. At that scale, it’s part of how the power system operates, whether it considers itself a grid participant or not.
Texas is responding with clearer technical and operating expectations. A facility that meets the Electric Reliability Council of Texas’ ride-through requirements clears one important gate on the path to energization, alongside required modeling, stability and interconnection reviews. Through controllable-load and bring-your-own-generation pathways in ERCOT's new Batch Zero process, developers can elect dispatchable curtailment under ERCOT’s real-time control in exchange for earlier energization.
Meaning, a data center must be able to reduce what it takes from the grid when the system is strained. It could curtail consumption, switch to onsite resources or coordinate several assets. What matters is that the response is available when promised.
Southwest Power Pool is making a similar trade through its High Impact Large Load framework. Developers that accept real-time forecast reporting, ramp-rate limits and remote-disconnect capability can have their interconnection study completed within 90 days, clearing the way for an agreement soon afterward.
In both markets, developers give the grid better visibility and more control during constrained periods. In return, they receive a faster and more predictable process.
PJM asked large loads to help cover the capacity they need
PJM Interconnection is approaching the challenge through resource adequacy that prioritizes whether the region has enough capacity to meet demand when the system is under the most stress.
Under PJM’s proposed Interim Resource Adequacy Service framework, certain new large loads could connect even when the region doesn’t have enough capacity to serve all forecast demand at its traditional reliability standard. Loads that didn't bring qualifying new capacity or receive an applicable capacity allocation could have been required to reduce their grid consumption during limited emergency conditions. FERC rejected most of the framework in October on jurisdictional grounds, but it encouraged states to decide which retail customers are curtailed when the system is stressed, which keeps the same trade on the table.
A data center could reduce that exposure by bringing qualifying generation, storage or demand-side resources. Otherwise, it would accept service that is less firm than what existing customers receive.
Bringing new capacity does not make transmission and distribution constraints disappear. A project may still need local upgrades, engineering studies and regulatory approvals. But if the system is short on capacity, a new load should either help fill the gap or be prepared to reduce its demand when the shortage becomes critical.
Different rules, one underlying question.
Each market is responding to its own constraints, but the question at the heart of these efforts is the same: What can a data center guarantee in exchange for access to power?
Federal regulators are now pushing that question across the country. In June, FERC directed all six regional grid operators under its jurisdiction to justify their existing large load tariffs or propose reforms. States are moving in parallel through large load tariffs and other requirements intended to ensure that data centers pay their share and manage their impact on the grid.
Today, developers learn the answer after spending time and money evaluating a site. Requirements can differ significantly from one utility or region to the next, making it difficult to compare locations, price the obligations attached to service and design the right energy infrastructure from the beginning.
Utilities and grid operators can make this process clearer. A developer should understand early on how specific commitments affect access to power.
The answer isn’t one national tariff or one required technology. Every grid has different needs, and data centers can provide flexibility through batteries, onsite generation, workload management or a combination of resources.
Utilities and grid operators should define a clear menu of performance-based pathways. Developers should know what they can commit, what the grid will require and what faster or more certain service they can receive in return.
Clearer options will help utilities separate real projects from speculative requests, allow developers to understand the cost of speed before committing to a site and give regulators a stronger basis for determining whether an arrangement protects reliability and existing customers.
The fastest data centers will not necessarily be the ones that find a way around the grid. They will be the ones that give the grid enough confidence to move forward with them.
