Tim Hughes is chief development officer at Stack Infrastructure, a data center developer.
After nearly two decades of flat load growth, the data center industry is now driving a surge in large load requests. This shift has caught many utility planners, regional transmission organizations and regulators off guard. The response has been to dust off tools from the pre-load-boom era, but those tools are not suited for modern infrastructure planning. Data center tenants, developers and utilities are all seeking one thing: certainty. Tenants need capacity and timing certainty; developers need certainty that their capital investments will become economically viable; and utilities need certainty that loads will materialize — otherwise, they risk being left with stranded assets and unfunded costs.
From Stack Infrastructure’s position as a large data center developer, we understand that both the amount and timing of load can be highly dynamic over a data center’s decades-long lifespan. Rather than focusing solely on the question of how much load will materialize at a data center in the long term, we propose shifting the focus to transparency around load drivers and the urgent need for modernized, adaptive planning frameworks.
Contingency planning is not misbehavior
Responsible loads of all types, including hyperscalers, often embed contingency into their interconnection requests. Utilities know this. Developers know this. However, this practice historically has been a quiet secret and not openly acknowledged by either party. That opacity creates misalignment: More accurate load requests aren’t rewarded, and the same quiet rules that apply to speculators are often applied across the board. As a result, utilities tend to cut load estimates while developers inflate them, creating a loop of mistrust. When this dynamic is compounded by continued speculative behavior, the system becomes overwhelmed and unable to keep pace.
Explaining the contingency planning factors developers commonly address can bring clarity to this opaque process.
- N-1 reliability: To ensure resilience, both utilities and data centers plan for redundancy. Data center loads can shift dynamically through global software-based traffic balancing. What may appear as "excess" load is essential for maintaining service continuity.
- Schedule uncertainty: Developers often request the fastest possible delivery timelines because utilities do not provide contractual guarantees. Including contingency is a practical and necessary hedge against delays.
- Ramp uncertainty: Hyperscalers and developers often have a high degree of confidence in the early phases of load ramp-up, but they have less certainty about the timing and scale of their full capacity needs. With no effective mechanism to support flexible, phased growth, they are incentivized to request full build-out capacity up front — even if they might not need that capacity right away.
- Site selection flexibility: Like utilities, tenants and developers require flexibility until final site decisions are made. This is not an attempt to manipulate the system, but a reflection of responsible and adaptive planning.
- Speculators: Unlike developers engaged in thoughtful contingency planning, speculators are opportunists with neither the intent nor the capacity to build. Fortunately, mechanisms such as development milestones and financial requirements can help filter them out of the system.
The industry must move toward a shared understanding. The need for power certainty — in timing, cost and continuity — is fundamental to the success of digital infrastructure. At the same time, the practice of requesting the maximum amount of power as quickly as possible is a response to outdated systems that can no longer support today’s level of growth and activity. A modern, flexible planning paradigm, one rooted in transparency and collaboration, is critical.
The utility response: The 4Cs
As demand has surged, we have observed utilities increasingly relying on four classic tools to manage ballooning uncertainty in an industry that now operates at unprecedented speed and scale. We refer to these tools as the 4Cs:
- Clawbacks: Reducing allocated capacity if progress stalls.
- Charges: Setting minimum demand or usage fees.
- Collateral: Requiring higher financial guarantees.
- Contract assignments: Putting limits on the transferability of contracts.
As developers, we understand the rationale of the 4Cs. These tools can improve queue integrity and ensure that other customers aren’t bearing the costs of our projects. However, when applied to unprecedented demand surge, they can have a chilling effect on development and delay critical infrastructure. In particular, clawbacks can disrupt continuity and undermine business models.
Instead, we advocate for a modernized toolkit: transparent queues, milestone-based gating, multiple notices to proceed with adequate time to perform development activities, more accurate and dynamic pricing that reflects system costs and locational value, early screening of nonviable applicants, and more dynamic contract structures, such as take-or-pay mechanisms with commitment-requirement milestones appropriate for developing the site. Many of these tools build upon the intent of the 4Cs but apply them in more flexible and constructive ways. We also recognize that these changes require new behaviors and commitments from developers, particularly a stronger emphasis on transparency in all aspects of contingency planning.
A call for a new planning paradigm
Developers operate at the intersection of vision and execution. We accept risk and adapt to change, but our success depends on strong, reliable partnerships. To keep pace with accelerating digital demand, we need utility frameworks that are more transparent, more flexible and more responsive.
As the next wave of infrastructure reshapes the grid, we must ask:
- How can interconnection and contracting evolve to support both flexible growth and long-term stability?
- How can optionality and certainty coexist in the planning process?
- How can transparency among all parties become a foundation of sustainable growth?
- How can hyperscalers, developers and utilities collaborate more creatively to ensure the U.S. remains the global leader in digital infrastructure to support AI?
The answers we choose today will shape the next decade of infrastructure investment and the economy that depends on it.
