Author: Benji Siem, IOSG

I. Introduction

This study begins with a simple observation: power systems are being asked to perform a task they were never designed to do.

With the accelerated penetration of renewable energy, the full-scale advancement of electrification, and the surge in demand for AI-driven data centers, the traditional model of "building more power generation and transmission facilities to meet peak loads" is crumbling. Infrastructure construction cycles are excessively long, grid connection queues are severe, and capital intensity remains high.

In this context, flexibility—the ability to dynamically adjust supply and demand in real time—has risen from an auxiliary function to a core pillar of grid reliability. The flexible supply that previously relied primarily on large industrial loads and peak-shaving power plants is evolving into a complex, multi-tiered market where distributed energy resources (DERs), software platforms, and aggregators coordinate millions of assets to maintain system balance.

We are at a structural inflection point. The winners of this transformation will not be the players who control power generation assets, but rather those who build connectivity and orchestration layers and unlock flexibility on a massive scale. Emerging crypto-native coordination models and token-based incentive mechanisms may further accelerate this shift by enabling global liquidity for decentralized participation, transparent settlement, and flexible services.

As this article will explore in depth, flexibility is no longer just a technological capability; it is becoming an emerging economic infrastructure—creating new value pools and reshaping how energy is traded, managed, and monetized by stacking revenues across capacity markets, ancillary services, demand response, and local markets.

Core Argument

The power flexibility market is at an inflection point. Rising renewable energy penetration, increased demand from data centers, and regulatory initiatives are creating a structural imbalance between supply and demand for flexibility services.

• The demand for power to fuel AI and application development is rapidly exceeding the available supply capacity of the power grid, driven primarily by factors including:
• Global data center power consumption is projected to double to approximately 945 TWh by 2030, slightly higher than Japan's current total power consumption. AI is the most significant driver of this growth, while demand for other digital services is also continuing to rise. It is worth noting that a lack of flexibility could also be a constraint on AI growth.

The electricity market urgently needs operational efficiency and flexibility to mitigate risks. Against the backdrop of lagging infrastructure development, the demand for and necessity of flexibility services have increased significantly.

• Power grids in many regions are already under immense pressure: it is estimated that approximately 20% of planned data center projects may face delays unless capacity risks are addressed.
• Currently, approximately 10,300 power projects with a total capacity of 2,300 GW—equivalent to twice the total installed power generation capacity in the United States—are waiting in line due to grid congestion difficulties faced by grid operators.

The middle layer of aggregation and connectivity infrastructure will be the biggest winner. It builds a crucial bridge between the supply side (users with idle capacity) and the demand side (stretched grid operators).

• Software-centric platforms that aggregate and optimize distributed energy resources (DERs) will capture a disproportionate share of value as the market expands from approximately $98.2 billion in 2025 to approximately $293.6 billion in 2034 (CAGR of 12.94% from 2025 to 2034).

II. Flexibility Market Overview

What is flexibility in the energy market?

In a power system, flexibility equals the system’s ability to quickly adjust generation and/or demand in response to signals (electricity prices, grid congestion, frequency, etc.) to maintain supply and demand balance and avoid power outages.

Historically, flexibility came almost entirely from flexible generating units (gas-fired peaking plants, hydropower). With the scaling up of renewable energy and electrification, system operators are now also sourcing flexibility from the following channels:

• Demand Response: Loads that can be reduced or shifted.
• Energy storage: batteries, electric vehicles, thermal energy storage
• Distributed generation: rooftop solar power, small-scale combined heat and power, etc.

The "flexibility market" is a collection of markets and contracts where flexibility is bought and sold, including wholesale markets, balancing/ancillary service products, capacity markets, and local distribution system operator (DSO) flexibility platforms. Aggregators act as intermediaries, providing platforms that enable grid operators to procure flexibility from end users, forming a critical infrastructure layer (see the "Flexibility Trading and Pricing" section for details). Settlement is handled by transmission system operators (TSOs), who pay aggregators fees, and the aggregators deduct commissions before paying customers.

There are two ways to deliver flexibility:

• Implicit flexibility: This is achieved automatically through static price signals, such as time-of-use pricing. For example, smart EV chargers automatically delay charging during off-peak hours at night. Price signals drive behavior.
• Explicit flexibility involves proactive responses to specific requests from grid operators. These actions are consciously performed and directly compensated through market coordination.

Detailed example

#Step 1: Customer Registration

Aggregators (such as CPower) contract with a manufacturing company to install monitoring equipment (smart meters, controllers) and integrate them into its building management system. The customer agrees to reduce 2 MW of load when invoked.

#Step 2: Register with the power grid operator

The aggregator registered this 2 MW (along with thousands of other sites) as a "demand response resource" with ISO. The aggregator must demonstrate that the resource can indeed be delivered, including baseline calculations, metering protocols, and sometimes test scheduling.

#Step 3: Market Participation

Aggregators bid for aggregation capacity across various markets:

• Capacity Market (Annual/Multi-Year): "I commit to maintaining 500 MW of availability during the summer peak electricity demand period."
• A recent statement in the energy market: "I can cut 200 MW of load tomorrow from 16:00 to 20:00."
• Real-time assistance service: "I can respond to frequency deviations within 10 minutes."

#Step 4: Scheduling

When the power grid requires flexibility, the TSO sends a signal to the aggregator. The aggregator's software platform then executes the following: sending notifications to registered customers (SMS, email, automatic control signals); activating pre-programmed load shedding (such as adjusting temperature control setpoints, dimming lighting, pausing industrial processes); and monitoring execution performance in real time.

Step 5: Settlement

After the event, ISO measured the difference between the actual delivery volume and the promised volume, and the flow of funds was: ISO → Aggregator → Customer (minus aggregator commission).

III. Key Participants

Exchange - Market Platform

These flexible trading venues connect buyers (DSO/TSO) with sellers (aggregators, DER owners). Fast frequency reserve markets also offer another trading platform.

#Representative Projects

EPEX SPOT, Nord Pool, Piclo Flex, NODES, GOPACS, Enera

#Business Model

• Fees for cleared transactions (typically 0.5-2% of the transaction amount or €0.01-0.05/MWh)
• Market access subscription/membership fees (participant annual fees)
• Some platforms operate as regulated public utilities (recovering costs through grid tariffs), while the rest operate commercially.

#Pricing

• The platform does not set prices, but instead facilitates price discovery through auctions (payment based on bids or unified settlement).
• Congestion management for local flexibility platforms (Piclo, NODES) typically costs €50-200/MWh.
• The wholesale balanced market can surge to €1,000+/MWh during scarcity events.
• Prices in classic wholesale markets (such as EPEX) can be negative, which is equivalent to proactively purchasing flexibility in a dedicated flexibility market.

Aggregator/Virtual Power Plant (VPP)

Control a flexible cluster of assets, whose revenue depends on winning contracts and properly scheduling load/storage.

#Representative Companies

Enel X, CPower, Voltus, Next Kraftwerke, Flexitricity, Limejump

#Business Model

• Revenue sharing with asset owners: Aggregators retain 20-50% of market revenue, with the remainder paid to customers.
• Some companies charge asset owners an upfront registration fee or a monthly SaaS fee.
• Performance bonuses may be awarded from utilities for exceeding dispatch targets.

#Pricing

• Capacity-based pricing: $30-150/kW·year (varies depending on market and product)
• Energy-based payment: the transmission of market prices (after deducting aggregator profits).
• Typical customer revenue: $50-200/kW·year for commercial and industrial (C&I) loads, $100-400/year for residential batteries.

Distributed Energy Resource Management System (DERMS) / Optimization Software

The software that enables forecasting, control, bidding, and compliance forms the intelligent layer of the entire system. It can be embedded within the aggregator platform.

#Representative Companies

AutoGrid (Uplight), Enbala (Generac), Opus One, Smarter Grid Solutions, GE GridOS, Siemens EnergyIP

#Business Model

• Enterprise SaaS License: Annual contract based on the number of MW managed or the number of assets controlled.
• Implementation/integration costs: One-time project cost for utility deployment ($500,000 - $5,000,000+).
• Managed services: Performance-based continuous optimization as a service

#Pricing

• Software licenses typically cost $2-10 per kW per year (depending on functionality and scale).
• The total contract value for large-scale utility DERMS deployments can reach $5 million to $20 million+ (5+ years).
• Some suppliers offer revenue sharing models (5-15% of incremental value).

Asset side

Physical suppliers: electric vehicles, batteries, thermostats, heat pumps, industrial loads, etc.

Power grid buyer

Demand side: Utilities and system operators, including DSOs, TSOs, suppliers and municipal utilities, that seek procurement flexibility to manage congestion, balance and peak loads.

#Representative Organization

PJM, CAISO, National Grid ESO, TenneT, UK Power Networks, E.ON, Con Edison

#Business Model

• For regulated entities, costs are recovered from users through grid tariffs or capacity fees.
• Procurement when flexibility is cheaper than infrastructure alternatives (“non-line alternatives”)
• Some vertically integrated utility operations handle internal DR projects, while the rest are outsourced to aggregators.

#Procurement Pricing

• Capacity procurement: $20-330/MW·day (PJM 2026-27 auction reached $329/MW·day)
• Ancillary services: $5-50/MW·hour (frequency response, spin-off standby)
• DSO local flexibility: €50-300/MWh (usually for auctions where payment is based on the bid)
• Rule of thumb: Flexibility must be cheaper than grid hardening (target savings of approximately 30-40%).

#Figure 1: Schematic diagram of the mechanism

• Distribution System Operator (DSO): A company that manages the local power network (distribution lines, substations) and is responsible for delivering electricity from the main transmission lines to homes and businesses.
• Transmission System Operator (TSO): A key entity that manages and maintains high-voltage networks (electric grids and natural gas pipelines) and is responsible for transporting energy over long distances from producers to local distributors or large users.

Revenue estimates for each participant

IV. Current Status of the Industry

The power system faces a structural imbalance between supply and demand in generating capacity and grid infrastructure. This contradiction manifests in two interconnected problems: an unprecedented backlog of grid connection requests and a surge in demand from electrification and data centers.

Grid connection queue backlog

By the end of 2024, more than 2,300 GW of generation and storage capacity will be seeking grid connection in the United States alone—more than double the existing total installed power capacity (1,280 GW). This backlog has become a major bottleneck for clean energy deployment.

Demand-side pressure

• Data Centers: Global electricity demand is projected to double to 1,000-1,200 TWh by 2030 (equivalent to Japan's total electricity consumption).
• PJM Capacity Market: Prices surged from $28.92/MW·day (2024-25) to $329.17/MW·day (2026-27), an increase of more than 10 times, primarily driven by data center commitments.
• US power grid planners' five-year demand forecasts have nearly doubled; AI data centers require 99.999% uptime and consume enormous amounts of electricity.
• Grid upgrade costs: The EU will need €730 billion in distribution investment and €477 billion in transmission investment by 2040; flexibility can provide 30-40% cost savings compared to infrastructure construction.

Flexible trading and pricing

Grid operators (such as PJM, ERCOT, CAISO, and other ISO/RTOs) need to balance supply and demand in real time, but they cannot communicate directly with millions of distributed assets (thermostats, batteries, industrial loads). Therefore, aggregators act as intermediaries.

The aggregators we analyzed (Enel X, CPower, Voltus) are positioned between the two parties:

1. Grid operators/utilities requiring flexible capacity
2. End customers with flexible loads or assets

Aggregators package thousands of small, distributed resources into a single “virtual power plant” and participate in wholesale market bidding as traditional power plants.

Settlement mechanism

Unlike power generation (measuring MWh output), demand response measures unconsumed MWh. This requires establishing a "baseline"—the amount of electricity a customer would have consumed in the absence of a demand response (DR) event. Common baseline methods include:

• 10-of-10 method: Take the average consumption of the same period over the past 10 similar days.
• Weather adjustment method: Adjust the baseline according to the temperature difference.
• Pre-event/In-event measurement method: Comparing consumption before and during the event.

Settlement example:

The aggregator then pays the client according to the contract (usually 50-80% of the total revenue), and the balance is the aggregator's revenue.

Flexibility is monetized through a variety of market mechanisms, each with different timeframes, product forms, and pricing structures. Suppliers can engage in "revenue stacking" across multiple markets to maximize returns on assets.

Furthermore, Energy Communities—localized citizen and small business collaborations empowered by EU policy—are becoming a significant force for aggregating flexibility. There are approximately 9,000 communities across the EU, representing around 1.5 million participants.

• By aggregating off-balance-sheet assets such as solar power, batteries, and controllable loads, these communities overcome the scale and coordination barriers that typically prevent individual households from accessing multiple flexible income streams.
• This aligns directly with research findings that flexibility providers can “overlay” value across capacity markets, ancillary services, energy arbitrage, demand response, and local DSO markets. The energy community has created the organizational and operational framework needed for reliable participation across markets, transforming fragmented DERs into coordinated portfolios, democratizing flexibility revenues, and supporting grid decarbonization and resilience.

Why Flexibility Matters

Flexibility services offer a faster and cheaper alternative to building new generation and transmission facilities. Virtual power plants can be "built" at the same speed as customer registration—without grid connection queues. Brattle Group estimates that VPP peaking capacity is 40-60% cheaper than gas-fired peaking plants or utility-grade batteries. ENTSO-E estimates that flexibility can save €5 billion annually in generation costs in the EU alone.

For grid operators: Real-time balancing of supply and demand; reduced reliance on expensive peak-shaving power plants and transmission upgrades; improved integration of renewable energy; enhanced grid resilience in extreme weather conditions.

For asset owners: generate new revenue streams from existing assets (batteries, EVs, HVAC, industrial loads); multi-service integration can increase returns by 30-50%; participation has minimal disruption to operations.

For consumers: Reduced electricity costs through demand response incentives; costs avoided by postponing infrastructure investment; improved reliability and reduced power outages.

For energy transition: Achieve higher renewable energy penetration without curtailing wind and solar power; decarbonized grid services (replacing gas-fired peaking power plants); and accelerated deployment of alternatives to infrastructure-constrained solutions.

Structural tailwind

1. Regulatory momentum: FERC Orders 2222/2023 (US), EU Demand Response Network Regulation (2027), and UK BSC P483 have engaged 345,000 households. Flexibility markets are being introduced in over 45 countries globally.
2. A surge in grid investment: US utilities are projected to invest $1.1 trillion in grid infrastructure by 2029. The EU will need €730 billion in distribution and €477 billion in transmission upgrades by 2040. Flexibility is a more economical alternative.
3. Data center demand: Global data center electricity consumption is projected to double to 1,000-1,200 TWh by 2030. PJM capacity prices will increase tenfold (2024-2027). This will simultaneously create both flexibility demand (grid pressure) and supply.
4. DER growth: 4 million+ US residential PV systems; 240,000+ residential batteries; 1 million+ EV sales in 2023. Critical scale has been reached, empowering aggregators and DER economics.

Key risks to be aware of

1. Oversupply after 2030: Large-scale battery storage investment may compress profit margins in the flexibility market. Pumped storage is experiencing a resurgence in some markets.
2. Cybersecurity: Millions of distributed assets expand the attack surface. The EU AI Act classifies grid operations as "high-risk." NFPA 855 increases the cost of urban battery storage by 15-25%.

V. Aggregator Business Model

Sources of income

1. Capacity fees ($/MW·year or $/MW·day): The largest and most predictable revenue stream. Customers are paid for availability, even if it is never dispatched. Example: PJM capacity prices reached $329/MW·day in the 2026-27 auction.
2. Energy payment ($/MWh): Payment for actual load reduction during the event. More volatile, depending on dispatch frequency and market prices.
3. Ancillary services ($/MW + $/MWh): frequency regulation, spinning reserve, etc. These are higher value but require faster response times (seconds to minutes). Voltus pioneered access to these higher-margin products.

Cost structure

Example of a unit economic model (C&I customer)

Benefit Allocation: How Aggregators Can Maximize Value

The most sophisticated aggregators "overlay" multiple revenue streams from the same asset:

Example: 10 MW industrial load in PJM

This is precisely why Enel's DER.OS and Tesla's Autobidder emphasize "co-optimization"—their AI determines at every moment which market to participate in to maximize total return.

VI. In-depth analysis of key players in the aggregated business layer

Enel X – A Global Market Leader

#Company Overview

Enel X is the demand response and distributed energy business unit of Enel Group, one of the world's largest utility companies (annual revenue exceeding €86 billion). Its origins trace back to EnerNOC—a demand response pioneer founded in 2001 and acquired by Enel in 2017. Today, Enel X operates the world's largest industrial and commercial virtual power plant, with over 9 GW of demand response capacity and 110+ active projects in 18 countries.

#Scale and Coverage

• Global capacity: 9+ GW under management (Q1 2025), with a target of 13 GW.
• North America: ~5 GW, covering 10,000+ sites in 31 US states and 2 Canadian provinces.
• Projects: 80+ demand response projects, 30+ utility partnerships (11 exclusive bilateral agreements)
• Customer payments: Nearly $2 billion has been allocated to DR participants since 2011.
• Technology Investment: Over $200 million invested in platform development

#Strategic Partnership

In September 2024, Enel X partnered with Google to aggregate 1 GW of flexible load from data centers—the world's largest enterprise VPP. This collaboration demonstrates the convergence of growing data center demand and flexible supply: hyperscale cloud providers driving grid stress can simultaneously become significant providers of demand-side flexibility through their UPS batteries and load shifting capabilities.

#Technology Platform: DER.OS

Enel X's DER.OS platform employs machine learning-driven scheduling optimizations, which, according to internal audits, can improve profitability by 12% compared to rule-based strategies. The platform streams data from over 16,000 enterprise sites and operates a 24/7/365 network operations center for real-time scheduling management and monitoring.

#Core Customers: Commercial & Industrial Facilities

These are large electricity consumers with interruptible loads—processes that can be temporarily reduced without causing major disruptions:

Key Insights

These customers already own "assets" (their electricity load). Enel X simply helps them monetize the flexibility they were unaware of. Enel X is explicitly positioned on the demand side and is asset-light, not building or owning any power generation assets. Reducing demand is equivalent to increasing supply in grid terms.

#The deeper meaning of Google's partnership

The Google deal in September 2024 is worth watching because it disrupts traditional models:

• Traditional model: Enel X recruits facilities → aggregates them into VPPs → sells them to the power grid
• The Google Model: Google data centers become flexible assets → Enel X operates VPPs → Grid operators purchase flexibility

Google data centers feature massive UPS battery packs (typically used for backup), flexible cooling loads, and some workload scheduling flexibility. Google is no longer consuming grid flexibility, but rather providing it—Enel X is the orchestration layer. This is a real-world application of the "data center is a grid asset" argument.

#Income Model Breakdown

#Competitive Position

• Advantages: Largest global scale, deep utility relationships, integrated clean energy ecosystem (11 GW renewable energy + 1 GW energy storage), mature platform, and financial backing from the Enel Group.
• Disadvantages: Traditional enterprise sales model, slower innovation cycle compared to pure startups, and higher enterprise management costs.
• Strategy: Focus on C&I niche markets, utility-grade renewable energy integration, and flexible data center partnerships.

Voltus – A Software-First Challenger

#Company Overview

Founded in 2016 by former EnerNOC executives Gregg Dixon and Matt Plante, Voltus positions itself as a technology-first alternative to traditional demand response providers. The company's argument is that superior software and broader market coverage can overcome scale disadvantages. As of September 2025, Voltus has ranked first in managed GW in Wood Mackenzie's North American VPP report for the third consecutive year.

#Size and Financing

• Capacity: 7.5+ GW under management (September 2025), a significant increase from 2 GW in 2021.
• Market Coverage: Active in all nine U.S. wholesale electricity markets and Canada—the broadest geographical coverage among pure-play startup aggregators.
• Funding: Total funding raised: $121 million (investors include Equinor Ventures, Activate Capital, and Prelude Ventures)
• SPAC Attempt: Announced a $1.3 billion SPAC merger in December 2021 (valued at $1.3 billion), but the transaction did not go through.

#Differentiation Strategy

Voltus differentiates itself in three dimensions: (1) pioneering innovation – the company pioneered access to operational reserve projects among multiple grid operators; (2) the broadest market coverage – active in projects that competitors avoid due to complexity; and (3) DER partnerships – instead of competing with equipment manufacturers, it partners with OEMs such as Resideo and Carrier to aggregate their installation bases into VPPs.

#Data Center Focus

In 2025, Voltus launched its "Bring Your Own Capacity" (BYOC) product, designed specifically for data centers and hyperscale cloud service providers. BYOC allows data center developers to deploy VPP-driven grid flexibility during project construction, offsetting capacity requirements by procuring flexibility from Voltus' distributed network, thereby reducing uptime. Partners include Cloverleaf Infrastructure.

#Core Client: C&I Facilities (similar to Enel X)

#OEM Partnership

#Why the OEM Model is Important

Customer acquisition cost (CAC) is the largest expense for aggregators. Through OEM partnerships:

• OEM is responsible for customer relations
• Voltus provides software and market access.
• Revenue is distributed among OEMs, Voltus, and end customers.
• CAC is significantly lower than direct enterprise sales.

Differences in Revenue Sources: Voltus vs Enel X

#Enel X: Primarily focused on the capacity market

• Predictable (Annual Auction)
• The unit price per kW is relatively low, but the volume is large.
• Large-scale MW commitment required

#Voltus: Deliberately pursuing ancillary services that competitors avoid.

#Why choose ancillary services?

Higher per unit (2-3 times the capacity market); fewer competitors (complexity creates a barrier); requires sophisticated software (Voltus's strength); but demands assets with faster response.

Competitive position

• Strengths: Technological precision, broadest market coverage, regulatory influence (former FERC Chairman Jon Wellinghoff serves as Chief Regulatory Officer), OEM partner strategy, data center positioning.
• Disadvantages: Smaller scale than Enel X, lack of utility-grade assets, high burn rate due to venture capital backing, failed SPAC.
• Strategy: Software monetization of third-party DERs, first-mover advantage in ancillary services, and data center partnerships

VII. VPP/Aggregator Investment Evaluation Standards

EU vs. US market

With its robust supportive regulations and highly interconnected infrastructure, the EU has outpaced the US in expanding system-wide flexibility. Eurelectric points out that the liberalized EU market has effectively incentivized both producers and consumers to participate, continuously enhancing the flexibility of supply; meanwhile, the widespread adoption of smart meters has facilitated time-of-use pricing, laying the foundation for demand-side shifts.

• Market Design: A liberalized market mechanism drives active participation from both the supply and demand sides, while smart meters, combined with time-of-use pricing, enable load shifting.
• Interconnected power grids: The EU's robust cross-border interconnected power grids have significantly reduced the frequency and duration of power outages, providing industrial users with a stable and reliable power supply.

The United States has enormous untapped potential for customer-side flexibility, and research suggests that large-scale load reductions (such as 100 GW) can be achieved with minimal impact on users.

• Grid Edge Focus: The rapid proliferation of distributed energy resources (DERs) makes flexible management at the "grid edge" increasingly critical for U.S. utilities.

"The inherent vulnerability of the power grid requires us to treat every connected asset with caution, ensuring that reliable supply matches forecasted demand. The rapid growth of intermittent power sources (unstable supply) and the simultaneous emergence of the electrification wave (peak demand) are posing serious challenges to the power system." — a16z

VIII. Conclusion

To date, flexibility has been dominated by "macro-flexibilities"—large industrial-grade assets (>200 kW) connected to the transmission or high-voltage distribution layers. These assets are attractive due to their ease of identification, contracting, and dispatch. However, this model is reaching structural bottlenecks. Macro-flexibilities are no longer sufficient, leading to power supply shortages and cascading problems such as grid connection delays. This increases system vulnerability and is becoming a key bottleneck for AI-driven load growth.

Therefore, the next frontier is inevitable: micro-flexibilities. These refer to small off-meter assets in the 1-10 kW range connected to low- and medium-voltage power grids, including EV chargers, heat pumps, HVAC systems, batteries, and home appliances. These assets, when aggregated, represent capacity several orders of magnitude higher than macro-level sources, but are significantly more difficult to acquire.

Current methods of acquiring this flexibility largely leave a significant amount of untapped value, creating opportunities for flexibility owners to fill this gap and participate in the ecosystem. An aggregator that directly reaches owners at critical scale, independent of suppliers or equipment brands, can create a powerful pull effect. Once users are horizontally aggregated, energy companies and OEMs will be economically incentivized to participate proactively, rather than trying to control customer relationships from the outset.

At the heart of all this, I believe DePIN has the greatest opportunity to disrupt this space and create long-term value through crypto-native infrastructure and incentive mechanisms. By increasing capacity and opening up new avenues to access flexibility, this segment will revolutionize the current electricity market, enabling AI to continuously reshape the world in an unconstrained manner.