A New Load Required a New Infrastructure Question
A global cloud provider and a regulated utility were assessing how to support prospective large-scale digital load through a more integrated power-and-infrastructure partnership. The cloud provider needed reliable, long-horizon capacity for a rapidly expanding compute footprint. The utility had to plan generation and network investment in line with system reliability, regulatory oversight and its obligations to existing customers.
The potential demand was large enough to shape long-lived decisions on firm power, transmission and grid development, so neither party could treat the question as a conventional supply contract. The partnership had to tie prospective load to infrastructure that could be developed, approved, financed and operated within the wider electricity system.
The Decision Was About More Than Power Supply
The partners were considering whether to co-fund generation, transmission, substations, grid upgrades or related firm-capacity infrastructure. They also had to decide how capital, capacity rights, development risk, operating obligations, governance and customer commitments should be split between them.
Each pathway struck a different balance between certainty and exposure. A deeper commitment from the cloud provider could give the utility more confidence to build capacity, but would increase the provider’s exposure to demand forecasts, location choices and long-lived infrastructure. A more conventional procurement structure could preserve flexibility, but offer too little certainty to support investment at the required pace.
The question was therefore how to build a credible sequence of commitments through which the cloud provider could secure strategic access to capacity without leaving the utility or its broader customer base with disproportionate risk.
Private Demand Met a Regulated Public System
A conventional corporate joint-venture analysis would compare contributions, expected returns, governance rights and risk allocation. A standard power-procurement exercise would assess capacity, price, contract duration and carbon attributes. Both were necessary, but neither could show whether the proposed structure was feasible within a regulated electricity system.
The utility’s investment decisions were shaped by rate-base treatment, prudence requirements, cost causation, service reliability and the need to maintain fair treatment across customer classes. Infrastructure built for a large new load could create wider network benefits, but could also impose costs or operating constraints that could not reasonably be passed on to other customers. The distinction that mattered was therefore between investment serving the wider system and investment attributable to the prospective customer, rather than between private and public capital.
The cloud provider faced different constraints. AI-related demand could become material at speed, yet its exact location, utilisation profile, deployment timing and duration remained uncertain. A structure built around a single load forecast risked overbuilding capacity or creating obligations that drifted out of line with the cloud provider’s evolving compute strategy.
Capacity Had to Be Delivered, Not Merely Contracted
The partnership’s viability depended on the full chain needed to turn prospective demand into usable power. Access to generation did not in itself deliver capacity. Development required compatible progress across generation, licensing, fuel where relevant, interconnection, transmission, substations, distribution infrastructure, grid operations and data-centre commissioning.
Each dependency affected the others. Delayed transmission or substation development could constrain capacity from otherwise available generation. A firm-power project could support long-term reliability and still be exposed to permitting, construction, workforce, supply-chain or operating-readiness risk. Data-centre deployment could also shift the demand profile for which infrastructure had been designed, affecting utilisation, cost recovery and the timing of further investment.
Regulatory and public scrutiny sat within the same delivery chain. Capacity arrangements had to withstand questions on reliability, customer fairness, cost allocation and the treatment of infrastructure that might serve both the cloud provider and the wider grid. The aim was a structure that coordinated demand commitments with infrastructure development and stayed viable within normal network constraints, without seeking an exception to them.
Testing Commitments, Rights, and Development Paths
Bruqe framed the engagement around the partners’ capacity requirements, planning horizons, investment limits, regulatory boundaries, reliability thresholds and acceptable exposures. The work examined the decision as a connected system of prospective AI demand, generation and network development, customer commitments, utility obligations, regulatory treatment and delivery risk.
The analysis tested alternative development pathways, including regulated-utility investment, corporate procurement, direct customer contributions, hybrid commercial structures, staged capital and differentiated capacity-rights arrangements. It assessed how each performed under different futures for AI-load growth, utilisation, infrastructure timing, regulatory treatment, construction and licensing, energy-market conditions and system-reliability requirements.
Bruqe’s proprietary platform’s intelligence graph linked prospective load, infrastructure dependencies, regulatory constraints and commercial commitments (each with its own timing constraints, level of confidence and sequencing challenges). Decision-specific models represented uncertain demand and differing development timelines, allowing the team to test how sensitivities in load or network readiness could affect usable capacity, investment exposure and cost allocation. This showed what timing mismatches could cost: a capacity commitment could stay commercially attractive while drifting out of line with infrastructure delivery or customer demand. The modelling separated commitments that held up across plausible conditions from those needing tighter milestones, conditional rights or further demand validation.
The work also assessed which commitments could create useful early options without forcing either party to commit prematurely to final scale. These included conditional demand agreements, phased infrastructure preparation, structured capacity reservations, decision rights tied to development milestones, and governance arrangements that separated customer-specific obligations from broader system investment.
Rather than favour a single commercial model, the work established which structures stayed feasible across plausible conditions, where risk needed to be allocated explicitly, and which signposts should trigger acceleration, redesign, further investment or deferral.
Building a Partnership the Grid Could Carry
The analysis showed that a durable partnership depended on a commercial and governance structure that reflected the physical and public-system conditions under which capacity would be delivered.
The resulting approach separated the value of firm capacity from the rights needed to access it, and customer-specific infrastructure from investment supporting wider reliability. It identified where binding demand commitments could justify earlier development, where capacity rights should stay conditional, and where direct customer contributions or tailored commercial arrangements could better align incentives.
Sequencing was equally important. Early actions could preserve options through joint planning, regulatory engagement, site and network readiness, capacity-reservation principles and demand-validation milestones. Larger capital commitments could then follow evidence that load, approvals, delivery conditions and system needs were progressing in step.
Securing Access While Retaining Legitimacy
The resulting decision architecture linked the cloud provider’s strategic power access to the utility’s wider reliability and public-service obligations. It let the partners assess how infrastructure could be developed around credible demand while keeping the ability to adjust as technology, load and regulatory conditions evolved.
The broader implication is that large-scale AI power access is not secured through a bilateral agreement alone. It depends on customer commitments, capacity rights, grid investment and public-system legitimacy staying aligned over time.
