The Electric Power Research Institute (EPRI) has officially introduced Flex MOSAIC, a comprehensive and uniform flexibility classification framework designed to categorize and manage large electric loads, primarily targeting the rapidly expanding data center sector. Developed under the umbrella of EPRI’s DCFlex initiative, the framework is the result of a massive collaborative effort involving more than 65 utilities, system operators, state and federal regulators, hyperscale data center operators, and technology providers. This voluntary system aims to solve one of the most pressing challenges in the modern energy landscape: the "time to power" bottleneck that currently delays the deployment of critical digital infrastructure across the United States and globally.
As the demand for artificial intelligence (AI) and cloud computing services reaches unprecedented levels, the traditional methods of grid planning and interconnection are proving insufficient. Flex MOSAIC provides a standardized language and technical foundation to define how large loads can adjust their power consumption in response to grid conditions. By establishing clear metrics for the magnitude, timing, duration, and frequency of a load’s response, the framework seeks to build the necessary confidence among grid operators to integrate massive new demands without jeopardizing the stability or affordability of the electric system.
The Critical Challenge of Data Center Interconnection
The surge in AI development has fundamentally altered the trajectory of electric load growth. In many regions of the United States, utilities are reporting demand forecasts that have doubled or tripled compared to projections made only a few years ago. Data centers, which serve as the backbone of the digital economy, require vast amounts of reliable, 24/7 power. However, the timeline for building a data center—often 18 to 24 months—is frequently at odds with the timeline for upgrading transmission lines or commissioning new power plants, which can take five to ten years or longer.
This temporal mismatch has created a significant backlog in interconnection queues. According to recent industry data, hundreds of gigawatts of capacity are currently waiting for approval to connect to the grid. EPRI President and CEO Arshad Mansoor noted that flexibility must now be viewed as a primary solution to this crisis. "As demand from AI and data centers grows at unprecedented speed, flexibility is becoming the third leg of the speed-to-power stool, alongside generation and transmission," Mansoor stated. He emphasized that the Flex MOSAIC framework allows for a "shared, credible way" to define and trust the flexibility that large loads can offer, which is essential for accelerating access to power.
Understanding the Flex MOSAIC Framework
The Flex MOSAIC framework is built upon practical performance characteristics that allow for the granular classification of how a large load behaves under different grid scenarios. Unlike previous, more nebulous definitions of "demand response," Flex MOSAIC provides a technical taxonomy that categorizes loads based on four key pillars:
- Magnitude: The total amount of power (measured in megawatts) that a facility can reduce or shift from its peak demand.
- Timing: How quickly a load can respond to a signal from a grid operator (latency) and the specific times of day or seasons when that flexibility is available.
- Duration: The length of time a load can maintain a reduced or shifted state before operational constraints require it to return to normal levels.
- Frequency: How often a facility can be called upon to provide flexibility over a given period (e.g., daily, weekly, or monthly) without impacting its primary business functions.
By organizing these characteristics into uniform flexibility classes, the framework enables utilities and system operators to treat large loads as dynamic assets rather than static liabilities. For instance, a data center capable of shifting its non-essential computational tasks (such as background AI model training) to overnight hours can be classified differently than a facility that requires constant, unyielding power for real-time financial transactions.
A Collaborative Ecosystem for Grid Modernization
The development of Flex MOSAIC was driven by the DCFlex initiative, a multi-year program launched by EPRI to bridge the gap between the energy and technology sectors. The initiative’s strength lies in its broad coalition of participants, representing every facet of the electricity value chain.
Initial participants include major investor-owned utilities like Southern Company, Entergy, Exelon, and DTE Energy; regional transmission organizations such as MISO, SPP, and the California ISO; and technology giants including Google, Meta, and NVIDIA. The inclusion of hardware and software providers like Honeywell and Siemens ensures that the framework is grounded in the reality of what modern building management systems and power distribution units can actually achieve.
Federal and regional reliability organizations have also signaled their support. Jim Robb, President of the North American Electric Reliability Corporation (NERC), highlighted the framework’s importance for maintaining grid integrity. "As large, flexible loads play a growing role in the power system, having clear, technically grounded definitions of flexibility is critical for reliability," Robb said. He noted that a common framework allows planners to speak the same language, which is vital for maintaining a reliable grid as the resource mix shifts toward more variable renewable energy.
Economic and Regulatory Implications
One of the most significant hurdles to integrating large loads is the concern over cost-shifting. State regulators are tasked with ensuring that the massive infrastructure investments required to serve new data centers do not lead to higher rates for residential and small business customers. The National Association of Regulatory Utility Commissioners (NARUC) has been a key partner in the DCFlex initiative to address these concerns.
NARUC President Ann Rendahl emphasized that state regulators are focused on maintaining affordability while supporting economic growth. "NARUC looks forward to engaging with EPRI and others on how a voluntary, standardized framework like Flex MOSAIC can create a common language and shared understanding of flexibility," Rendahl stated. She added that such a framework provides benefits to regulators when evaluating data center integration, ensuring that flexibility can be used to mitigate the need for expensive grid upgrades that might otherwise be passed on to the general ratepayer.
From a market perspective, Flex MOSAIC could pave the way for new revenue streams for data center operators. By quantifying their flexibility, these companies can more easily participate in wholesale electricity markets, providing services like frequency regulation or spinning reserves, which were traditionally the domain of power plants.
Chronology of the Shift Toward Flexible Load Management
The journey toward the Flex MOSAIC framework reflects a broader evolution in the power industry:
- 2015–2020: The rise of hyperscale cloud computing begins to put localized pressure on grids in hubs like Northern Virginia and Silicon Valley.
- 2021–2022: The emergence of generative AI leads to a spike in demand for high-density power racks. Utilities begin to report that existing "first-come, first-served" interconnection rules are becoming unworkable.
- 2023: EPRI launches the DCFlex initiative to create a dedicated forum for utilities and data center operators to collaborate on technical solutions.
- 2024: Pilot programs are initiated across various regions to test the feasibility of "load shifting" in data centers.
- Early 2025: EPRI officially releases the Flex MOSAIC framework, moving from the conceptual phase to a technical standard ready for jurisdictional adaptation.
Future Outlook: The Road to 2026 and Beyond
The introduction of Flex MOSAIC is expected to be a major topic of discussion at upcoming industry gatherings, most notably the DTECH Data Centers & AI event scheduled for May 12-14, 2026, in Scottsdale, Arizona. This assembly will bring together engineers, technical decision-makers, and utility executives to discuss the practical implementation of these standards. Key topics will include modernizing infrastructure, streamlining interconnection studies, and the integration of onsite generation—such as microgrids and battery storage—into both utility and customer-side systems.
As the industry moves forward, the focus will shift from defining flexibility to implementing it at scale. The success of Flex MOSAIC will depend on its adoption by regional transmission organizations (RTOs) and independent system operators (ISOs). If integrated into grid planning models, the framework could potentially shave years off the time required to bring new data centers online.
Furthermore, the framework’s principles are likely to be applied beyond data centers to other massive industrial loads, such as green hydrogen production facilities and large-scale battery manufacturing plants. By creating a standardized way to measure and verify flexibility, EPRI has provided a blueprint for a more agile and resilient power grid capable of supporting the next generation of technological innovation.
In the long term, the Flex MOSAIC framework represents a pivot in how society thinks about electricity consumption. Rather than the grid simply reacting to demand, the framework envisions a future where demand actively supports the grid. This bidirectional relationship is viewed by many experts as the only viable path forward for achieving a carbon-neutral economy while meeting the insatiable energy appetite of the digital age. Through the collaboration of over 60 organizations, the launch of this framework marks a significant milestone in ensuring that the energy transition and the AI revolution can proceed in tandem without compromising the stability of the world’s most complex machine: the electric grid.