Utilities and Regulators Accelerate Pilot Innovation to Shorten Technology Deployment Cycles

A recent study by the U.S. Department of Energy shows that through smarter pilot program design, utilities can more efficiently apply innovative technologies when addressing challenges such as load growth and affordability.

A June report from Lawrence Berkeley National Laboratory (LBNL) on pilot program design notes that current pilots are often duplicative, yield unclear conclusions, and lack a clear path to scale.

The report states: "Safety is a utility's top priority, so they are cautious about adopting new technologies, but quickly validating good ideas is critical." Facing growing demand, especially from data center loads, some utilities have begun accelerating their efforts to address the risk that new generation and storage may not come online in time.

A data center load flexibility demonstration conducted by Salt River Project on May 3 using Emerald AI software has led to an announcement of scaled deployment within the PJM Interconnection system. Many utilities are seeking ways to achieve this "speed of innovation."

"Because customer adoption, expectations, and technology are evolving at an exponential rate, it is more critical than ever to move innovative programs and pilots to scale quickly," said Chanel Parson, director of clean energy and demand response at Southern California Edison. She added that the faster utilities can scale solutions, "the better they can keep pace with change."

The LBNL study found that utilities are working with regulators to find pilot designs that can accelerate innovation. To address rapidly growing electric vehicle penetration, Pacific Gas and Electric launched a controlled charging program for 1,000 customers in January, and it is nearing its next phase. "This demonstrates the pace at which utilities want to move," said Marina Donovan, vice president of global marketing at smart meter supplier Itron.

The LBNL report notes that streamlined pilot frameworks, often called "regulatory sandboxes," can support the speed of innovation and have validated new value propositions as well as approaches to implementation, operations, planning, and investment.

Barriers include financial and regulatory disincentives

Pilots are key to innovation, but according to another LBNL analysis from 2022, utilities tend to underinvest in new technologies and methods compared with other, more innovative industries.

Some utilities have achieved success by gradually scaling pilots. Kristin Carlson, spokesperson for Green Mountain Power, a Vermont investor-owned utility, said its pilots "laid the foundation for current developments." She added that during a peak event in June 2025, the company's distributed storage virtual power plant (VPP) program saved customers "approximately $3 million" in one hour.

But like traditional pilots, the company's VPP program grew slowly: starting with 20 customers in 2017, it now has 5,035 participants providing more than 75 megawatts of dispatchable peak resources. Jeff Monford, spokesperson for Southern California Edison, said the company also gradually cultivated its VPP innovation. Its 5-megawatt Power Flex solar-plus-storage VPP pilot launched with Sunrun in 2020 now has multiple vendors, 25 megawatts of capacity, and plans to continue expanding after 2027.

The LBNL analysis found that financial disincentives for utilities are a major barrier to accelerating innovative pilot programs. Analysts noted that to obtain approval for cost recovery, utilities must demonstrate to regulators that proposed technologies will benefit customers, which discourages investment in pilots of unproven tools and methods in favor of "the status quo that has historically provided a stable business environment." Additionally, traditional pilot frameworks often fail to clearly define pilot parameters, leading to misleading metrics, biased or uncertain results, duplication, participant disputes, delays in approval and implementation, and a lack of clear paths to scale.

Anne Hoskins, an executive at Generac Power Systems and former commissioner on the Maryland Public Service Commission, said these challenges can be overcome with better policies. For example, pilot frameworks could include cost limits, timelines, requirements for sharing lessons learned, and clear next steps for completed projects. "Regulators can create 'sandbox' frameworks to quickly approve innovative pilot projects that meet guidelines," she said, adding that this could be a low-cost way to test new technologies, tools, or approaches.

Piloting large load flexibility

A key innovation emerging in recent pilots is data center flexibility. Although the concept is still new, it has been embraced by Energy Secretary Chris Wright, who has called on federal regulators to streamline interconnection approvals for flexible data centers.

In May, Salt River Project publicly demonstrated the potential of data center flexibility in collaboration with participants in the Electric Power Research Institute (EPRI) DCFlex initiative. Anuja Ratnayake, EPRI's emerging technologies lead, said that as the first demonstration in the initiative, it showed that flexibility can be achieved without compromising the value proposition of participating data centers. She noted that unlike the rigorous, repeatable pilot process described by LBNL, the Arizona demonstration was "a specific controlled test" with predetermined conditions, with plans to test the software in live operations in four steps by mid-2026.

Emerald AI's fourth demonstration is scheduled for mid-2026 at the 96-megawatt Aurora AI facility on the PJM system in Virginia, and will be conducted at full scale, in real time, under real conditions, to show how much flexibility the software can provide at scale. Varun Sivaram, CEO of Emerald AI, said AI companies lack patience because data center interconnection speed will determine when artificial general intelligence is developed. However, according to a November PJM market monitoring report, system operators must first confirm the certainty of data center flexibility before adopting it as a solution for faster interconnection of large loads. The report questioned the viability of flexibility as a workable solution for the largest U.S. grid operator. Sivaram said Emerald AI's streamlined four-step demonstration process is designed to give utilities confidence to connect flexible data centers faster. He believes that once flexibility is successfully demonstrated at scale, "this cautious industry will open the floodgates and move faster."

"I want to see utilities use our sandbox"

State regulators and regulated utilities are also working to achieve speed of innovation through pilots, with mixed results. Researchers and stakeholders say innovation can happen faster if policymakers set clear goals and remove regulatory barriers. The LBNL report notes that policy directives and vision statements can identify "goals that stakeholders can jointly support and work toward." It cites examples such as New Jersey's Energy Master Plan, Hawaii's 2014 "Commission Inclination on the Future of Hawaii's Electric Utilities" and its 2024 follow-up, and the District of Columbia's PowerPath DC, all of which have advanced "ambitious energy goals."

To further advance the pace of innovation, the Hawaii Commission's 2020 innovation pilot framework streamlined pilot implementation, limiting regulatory review to 45 days and reducing uncertainty around cost recovery. In 2019, as Green Mountain Power's distributed storage program grew, the Vermont Public Utility Commission ordered a multi-year regulatory plan that included a framework for innovative pilot projects, allowing the utility to scale multiple pilots. VPUC Commissioner Riley Allen said two important features of the plan were requirements for advance notice to stakeholders and a cap on cost increases at 2%. "Utilities are reluctant to deviate from known frameworks, but the industry needs to address emerging factors like growing electricity demand," he said. The sandbox concept "creates space for doing good within the regulatory process."

Michigan's 2019 Power Grid plan includes a 90-day pilot approval process and incentives for utilities to collaborate with stakeholders. But LBNL notes that according to Michigan's pilot database, pilots "have not yet advanced to full programs." Michigan Public Service Commissioner Katherine Peretick said progress is still being made through slower general rate cases. "I want to see utilities use our sandbox," she said, adding that "growing electricity demand may increase utility interest in using its fast-track pilot framework."

LBNL says two other models—New York's "Reforming the Energy Vision" and Oregon's "Smart Grid Testbed"—are also important sandbox framework efforts. Randomized controlled trials (RCTs), advocated by customer-owned device aggregator Renew Home, represent an approach to innovation speed outside the sandbox. Ben Brown, CEO of Renew Home, said RCT data can approximate the value of customers' smart devices, bypassing the pilot process entirely.

The Connecticut model

Connecticut's "Innovation Energy Solutions" (IES) sandbox framework is gaining attention as a potential model, as the state prepares to evaluate its first completed round of pilots. In 2023, the Connecticut Public Utilities Regulatory Authority (PURA) launched its first round of pilots, with a new round each year since. LBNL says the framework requires four phases: proposal, selection, deployment, and evaluation. To reduce deployment barriers and foster collaboration, IES allows pilots to be led by vendors, utilities, or utility-vendor partnerships.

Former PURA Chair Marissa Paslick Gillett said IES is designed to overcome regulatory hurdles and provide regulatory certainty that the pilot process will "produce results." She also said it provides an "innovative regulatory environment." "There's a 'fail fast' mindset built into the process, and milestones must be met to receive continued funding," Gillett said. She added that scalability is not yet clearly defined, but Connecticut law allows PURA to mandate that utilities scale up pilot projects that have demonstrated viability.

"Regulatory sandboxes accelerate innovation within existing regulatory frameworks, and the IES program is one of the best examples," said Matt McDonnell, co-founding partner at regulatory consulting firm Current Energy Group and independent administrator for the Connecticut program. McDonnell said IES shortens the typical pilot timeline from at least four years to two years, and all three participation pathways include ratepayer protections.

Piclo's grid flexibility marketplace recently completed its first IES cycle with Eversource Energy and United Illuminating. In the final report, both utilities and Piclo said the market launched faster than expected and attracted more flexibility providers than anticipated. "But the most important part of building a market is continuity," said James Johnston, CEO of Piclo. "If there's a second year and the prospect of a five-year revenue opportunity, participation will grow." However, Connecticut's utilities were not convinced. Eversource reported that the Piclo pilot fell short of expectations in provider engagement, locational specificity, cost-effectiveness, and operations. Due to a lack of system infrastructure to visualize and control customer-owned resources, both Eversource and United Illuminating considered Piclo's market premature for Connecticut. United Illuminating said Piclo's solution could become viable "as market conditions and grid needs evolve."

Former PURA Chair Gillett said load growth, reliability, and affordability concerns make it important to fully utilize the grid. "The IES framework gives PURA commissioners final decision-making authority over pilots in the sandbox within two years," McDonnell said. He added that its rulings on Piclo and other projects in each round "will provide guidance on how to resolve disagreements between utilities and vendor participants." The IES administrator's final report is due December 15. The commission has said it is committed to issuing final rulings as soon as possible.