The two principal documents examined in this report are “Pricing Scarcity: Easing the Short-Term Surge in Data Center Energy Demand,” written by Domenico Ferraro, Ph.D., Associate Professor at Arizona State University, and published by the Goldwater Institute on May 14, 2026, together with its accompanying Goldwater Institute article, “Data Centers Aren’t Breaking the Grid—Bad Pricing Is.” The official policy paper is available at https://www.goldwaterinstitute.org/policy-report/pricing-scarcity/, and the accompanying article is available at https://www.goldwaterinstitute.org/goldwater-report-data-centers-arent-breaking-the-grid-bad-pricing-is/. The second policy paper is “Powering the AI Boom,” also written by Domenico Ferraro, Ph.D., and published by the Goldwater Institute on June 16, 2026, accompanied by the article “America Can Power the AI Boom, But It Needs to Build Fast Enough.” The policy paper is available at https://www.goldwaterinstitute.org/policy-report/powering-the-ai-boom/, and the accompanying article is available at https://www.goldwaterinstitute.org/new-report-america-can-power-the-ai-boom-but-it-needs-to-build-fast-enough/. Together these four documents provide one of the most comprehensive recent policy analyses of the relationship between artificial intelligence, data-center growth, electric-grid infrastructure, electricity pricing, and utility regulation. (Goldwater Institute⁠)

These papers are important because they address one of the most controversial public policy questions facing the United States: whether rapidly expanding AI and cloud-computing data centers are directly responsible for increasing residential electricity costs or whether broader structural weaknesses within the electric-power system are the primary cause. Rather than approaching the issue from an environmental or political perspective, the papers analyze the economics of electricity markets, the engineering characteristics of power systems, utility regulation, and infrastructure investment. Their central argument is that rapidly growing electricity demand from data centers is real and significant, but that demand is exposing longstanding weaknesses in electricity markets rather than creating entirely new ones. (Goldwater Institute⁠)

The first policy paper, Pricing Scarcity, begins by examining projections showing that U.S. data-center electricity consumption is expected to increase dramatically during the remainder of the decade. Drawing upon the Lawrence Berkeley National Laboratory report prepared for the U.S. Department of Energy, Ferraro notes that data centers accounted for approximately 4.4 percent of total U.S. electricity consumption in 2023 and could grow to between 6.7 percent and 12 percent by 2028 as artificial intelligence computing expands. Rather than disputing these projections, the paper accepts them and argues that policymakers should focus on how to accommodate this growth rather than attempting to suppress it. (Goldwater Institute⁠)

The report explains that electricity differs fundamentally from most other commodities because production and consumption must remain balanced continuously. Electricity cannot be economically stored in large quantities across an interconnected national grid, making instantaneous balancing of supply and demand essential. When demand exceeds available generation or transmission capability, prices increase sharply, congestion develops, reserve margins shrink, and reliability can deteriorate. Consequently, even relatively small changes in peak demand can produce disproportionately large increases in operating costs if additional generation and transmission infrastructure are not already available. (Goldwater Institute⁠)

One of the paper’s most significant technical contributions is its distinction between wholesale and retail electricity markets. Wholesale electricity prices fluctuate continuously according to actual operating conditions, while retail customers generally pay regulated rates that change only after utility commission proceedings. Ferraro argues that this disconnect hides scarcity from consumers and prevents demand from responding efficiently during periods of system stress. In effect, wholesale markets recognize shortages almost immediately, while retail pricing often conceals those shortages until utilities later recover their costs through higher rates. According to the paper, the inability of retail markets to communicate real-time scarcity contributes significantly to congestion, inefficient investment, and ultimately higher electricity costs. (Goldwater Institute⁠)

The report therefore recommends expanding the use of dynamic pricing, demand-response programs, interruptible service contracts, and more flexible electricity tariffs, particularly for sophisticated industrial customers such as hyperscale data centers. Because many computing workloads can be shifted geographically or delayed for several hours without affecting customers, Ferraro argues that data centers possess a unique ability to reduce demand during critical peak periods. Instead of viewing these facilities solely as consumers of electricity, he suggests they should increasingly become active participants in maintaining grid stability by reducing consumption when requested by utilities or regional transmission organizations. (Goldwater Institute⁠)

The accompanying Goldwater Institute article translates these technical concepts into language more accessible to policymakers and the general public. It emphasizes that political debate has largely focused on blaming data centers for increasing electricity demand while overlooking the underlying market design. According to the article, the real problem is not the existence of additional demand but an electricity system that masks scarcity until it appears as congestion, reliability concerns, or higher customer bills. The article reinforces Ferraro’s argument that improved pricing mechanisms and market incentives would encourage both consumers and large industrial users to respond more efficiently during periods of constrained supply. (Goldwater Institute⁠)

The second policy paper, Powering the AI Boom, shifts attention from short-term operational improvements to the long-term physical expansion of the nation’s electrical infrastructure. Whereas the first paper concentrates on improving market efficiency, the second recognizes that pricing reforms alone cannot satisfy the projected increase in electricity demand. The United States must significantly accelerate construction of generation facilities, transmission lines, substations, transformers, and local distribution infrastructure if it intends to remain globally competitive in artificial intelligence and cloud computing. (Goldwater Institute⁠)

Ferraro explains that one of the most serious challenges facing utilities is the mismatch between the construction schedule for data centers and the construction schedule for electric infrastructure. Large AI facilities can often be completed within two to three years, while major transmission projects, natural-gas generating stations, nuclear facilities, or high-voltage substations frequently require much longer because of permitting requirements, environmental reviews, financing delays, equipment shortages, and regulatory approvals. This timing mismatch creates temporary shortages that can increase electricity costs even when adequate long-term generation eventually becomes available. (Goldwater Institute⁠)

The paper also discusses cost allocation. It argues that existing residential and commercial customers should not automatically bear the financial burden of infrastructure constructed primarily to serve new industrial customers. Instead, Ferraro recommends long-term contractual commitments, phased infrastructure investments, and customer-specific cost recovery mechanisms that require large data-center operators to assume an appropriate share of the costs associated with their own growth. Such arrangements reduce financial risk for utilities while protecting existing ratepayers from subsidizing speculative developments. (Goldwater Institute⁠)

Another important aspect of the report concerns transmission congestion. Electricity prices frequently increase not because generation is unavailable somewhere within an interconnected region but because transmission capacity is insufficient to move lower-cost electricity to where it is needed. As data centers become concentrated within particular metropolitan regions, congestion costs can rise rapidly unless transmission capacity expands simultaneously. Ferraro therefore argues that transmission planning should become a strategic national priority alongside new generation development. (Goldwater Institute⁠)

The report further emphasizes that dependable generation remains essential. Although renewable resources, battery storage, and energy efficiency all contribute to meeting future demand, the analysis concludes that dispatchable generation capable of operating whenever required will continue to play a critical role in supporting artificial intelligence infrastructure. The report therefore advocates policies that accelerate construction of reliable generation while preserving market incentives for innovation and investment. (Goldwater Institute⁠)

The accompanying Goldwater article explains that America possesses sufficient technical capability and financial resources to support the AI revolution, provided infrastructure development keeps pace with computing demand. The article stresses that unnecessary permitting delays, fragmented regulation, and uncertainty regarding cost recovery represent more significant obstacles than engineering limitations. Faster approval of energy infrastructure, combined with appropriate market pricing, would allow both utilities and private investors to respond more effectively to growing electricity demand. (Goldwater Institute⁠)

Taken together, the two policy papers present a consistent analytical framework. They reject the simplistic argument that data centers alone are responsible for rising electricity prices while also rejecting the equally simplistic claim that data centers have no impact on electricity markets. Instead, they demonstrate that electricity prices are influenced by the interaction of multiple factors, including rapidly increasing demand, transmission congestion, insufficient generating capacity, regulatory delays, retail pricing structures, infrastructure financing, and utility cost-allocation policies. Data centers represent one important source of new demand, but whether that demand ultimately raises customer rates depends largely upon how electricity markets, utilities, and regulators respond. (Goldwater Institute⁠).

The policy recommendations emerging from these papers provide a practical framework for state legislatures, public service commissions, governors, utilities, and regional transmission organizations. They recommend modernizing retail electricity pricing to better reflect actual system conditions, expanding demand-response participation by large industrial customers, accelerating permitting for new generation and transmission facilities, encouraging long-term contractual commitments from major electricity users, ensuring that infrastructure costs are allocated fairly to those creating new demand, strengthening transmission planning, and expanding dependable generation capacity alongside renewable resources and storage technologies. Rather than treating artificial intelligence as a threat to the electric grid, the papers argue that AI growth should become the catalyst for modernizing America’s electric infrastructure while protecting existing ratepayers from inappropriate cost shifting. In that sense, the studies present an optimistic but disciplined vision: the United States can simultaneously support rapid technological innovation, maintain electric reliability, and protect consumers if policymakers implement market-based pricing, accelerate infrastructure development, and establish transparent rules governing how the costs of serving new large electrical loads are assigned. (Goldwater Institute⁠)