What Curtailment and Demand Response Mean in Practice
Curtailment is the deliberate reduction of a facility’s power consumption, typically in response to grid conditions or price signals. In mining, this means reducing fleet hashrate — either partially or completely — for a defined period. The grid operator or power purchaser signals the need for load reduction; the facility reduces consumption and is compensated for the reduction.
Demand response is the formal program structure through which this compensation is administered. Grid operators create demand response programs to give large electricity consumers a structured way to participate in grid balancing, with defined triggers, compensation mechanisms, and performance requirements. For Bitcoin mining facilities, participation typically involves registering with the grid operator or a demand response aggregator, agreeing to respond to curtailment signals within a specified timeframe, and receiving payment for the load reduction delivered.
The two are related but distinct. Curtailment is the operational act of reducing load. Demand response is both the program and the market structure that compensates for it. A facility can curtail without participating in a formal demand response program — for example, when spot electricity prices rise above the cost of mining, and it is more profitable to sell power back to the grid than to mine. Formal demand response participation adds structured compensation on top of this economic curtailment logic.
Why Bitcoin Mining Is Well-Suited to Demand Response
Most large industrial electricity consumers — steel mills, chemical plants, data centers running committed workloads — face significant operational or product-quality risks if they reduce their load quickly. Bitcoin mining, by contrast, is uniquely flexible. Operators can reduce or shut down mining load whenever requested, knowing that the only tradeoff is lower hashrate and reduced Bitcoin production during the curtailment window. No customer service is interrupted, no data is lost, and mining resumes normally once power is restored.
This characteristic makes mining facilities among the most responsive and reliable demand response resources available to grid operators. Response times can be measured in seconds for power reduction and in minutes for a full restart, depending on the facility configuration. No other industrial load of comparable scale can respond as quickly with as little operational consequence.
ERCOT, the Texas regional grid operator, recognized this early. The organization created a voluntary curtailment program specifically for Large Flexible Load customers — a classification that includes Bitcoin mining facilities — to reduce power use during periods of high demand. The program formalizes what miners were already doing opportunistically and creates a structured payment mechanism for doing it on request.
This Bitcoin mining curtailment workflow shows how ERCOT and AESO dispatch signals trigger an API-based fleet ramp-down, earning demand response credits before UMC OS automatically recalibrates miners via Perpetual Tune upon restart.
The ERCOT Market: Where Curtailment Revenue Has Been Most Significant
Texas has become the most developed market for mining curtailment participation, primarily because ERCOT’s deregulated structure and real-time price volatility create frequent and financially significant curtailment opportunities.
ERCOT allows Large Flexible Load facilities to participate in energy and ancillary services markets, treating miners as part of the energy-balancing toolkit. Miners in Texas have pre-arranged voluntary curtailment agreements with ERCOT to scale back consumption during extreme events — including heat waves and winter storms — to help prevent blackouts.
The financial scale of these arrangements can be substantial. In August 2023, Riot Platforms earned roughly $31 million in a single month by not producing Bitcoin — taking in curtailment credits and demand response payments from ERCOT that exceeded its typical mining revenue for the period. More recently, Riot reported $30.6 million in power curtailment credits during Q3 2025 alone, representing a 147% increase compared to the same period in 2024.
These figures reflect a specific market structure. ERCOT’s Large Flexible Load classification applies to facilities with an expected peak demand of 75 megawatts or greater, and it is assumed that these facilities will reduce part of their electricity consumption during hours when potential wholesale power prices exceed $100 per megawatthour. Facilities below that threshold may participate through demand response aggregators, which pool smaller loads to meet market participation minimums.
The U.S. Energy Information Administration projected that approved Large Flexible Load facilities in ERCOT would consume approximately 54 billion kWh in 2025 — roughly 10% of ERCOT’s total electricity demand — up nearly 60% from expected demand in 2024. The scale of mining’s role in the Texas grid has grown substantially, increasing both scrutiny of the programs and the competitive value of being a reliable and responsive curtailment participant.
Alberta: A Developing Market With Structural Advantages
Alberta operates under a deregulated electricity market administered by the Alberta Electric System Operator (AESO), which shares some structural similarities with ERCOT. The province has attracted significant mining investment in part because of access to low-cost natural gas generation and a regulatory environment that has remained accommodating to large industrial electricity consumers.
Operations in Alberta can enter into agreements with AESO that allow them to curtail computing workloads and sell power back to the provincial grid during periods of peak demand or grid stress. This sell-back capability — available to facilities with their own generation assets — represents a more direct form of grid participation than enrollment in demand response programs, and it is attracting investment from operators seeking both low-cost power and the option to monetize flexibility when grid conditions make it profitable.
Alberta has also eased cost pressures on energy-intensive industries by freezing its carbon price at $95 per tonne in 2025, reducing one of the fixed cost components of operations in the province. Combined with access to competitive natural gas pricing and available land for large-scale facilities, Alberta’s position as a curtailment-friendly mining market is strengthening.
The AESO’s approach to large load integration is evolving. In mid-2025, AESO established an interim limit of 1,200 MW of additional large-load capacity that the grid can accommodate without negatively impacting reliability, signalling both the pace of growth and the grid operator’s effort to manage it in a structured way. For mining operators evaluating Alberta as a location, understanding the interconnection process and curtailment program structure is a prerequisite for planning.
Two Types of Curtailment: Economic and Program-Based
In practice, mining operations encounter curtailment in two distinct forms, and the operational requirements for each differ somewhat.
Economic curtailment
occurs when the spot price of electricity rises above the effective cost of mining per unit of revenue. When this happens, it is more profitable to stop mining and either sell power back to the grid (if the facility has a power purchase agreement with sell-back provisions) or simply reduce load to avoid paying high spot prices. Economic curtailment is reactive — it responds to price signals in real time — and requires the ability to monitor electricity prices and adjust fleet performance quickly.
Program-based curtailment
occurs when the facility has enrolled in a formal demand response program and is responding to a dispatch signal from a grid operator or aggregator. This type of curtailment may be scheduled in advance, triggered by a real-time signal, or both. Program-based curtailment typically has defined performance requirements — response time from signal to load reduction, minimum curtailment duration, and accuracy of the load reduction delivered — and compensation is tied to reliably meeting those requirements.
Both types place real operational demands on the mining facility. A fleet that cannot reduce power quickly, predictably, and to a specific target cannot participate effectively in either form. This is where firmware capabilities become a practical requirement rather than a performance enhancement.
What Firmware Needs to Do for Curtailment to Work
Stock Bitmain firmware was not designed with curtailment participation in mind. It offers a small number of fixed power presets and limited API for programmatic fleet control. Transitioning between presets requires manual interaction via the miner’s web interface, and the transition is neither instantaneous nor granular enough to meet the response time requirements of most demand response programs.
For curtailment participation to be operationally viable at scale, firmware needs to support three capabilities:
Programmatic power control via API.
The fleet management system — or a direct integration with a demand response aggregator — needs to be able to send a power reduction command and receive confirmation of execution without operator involvement. UMC OS exposes a full REST API that allows operators to set hashrate targets, adjust power modes, and monitor fleet state programmatically across all connected miners simultaneously.
Rapid and accurate ramp-down.
When a curtailment signal arrives, the fleet needs to quickly reduce power to a specific target level. The speed and accuracy of this reduction affect both the operator’s ability to meet program requirements and the event’s financial outcome. Firmware that can ramp down smoothly to a defined power target — rather than toggling between fixed presets — gives operators more control over exactly how much load is reduced and how quickly.
Reliable ramp-up and performance recovery.
After a curtailment event ends, the fleet needs to return to its pre-curtailment performance level. With stock firmware, ramp-up is straightforward because settings are static. With custom firmware running continuous optimization, the system recalibrates to current hardware and thermal conditions on restart, so the fleet reaches its optimal operating point after ramp-up without manual intervention or a new calibration run.
UMC OS’s Perpetual Tune feature handles this recalibration automatically. When miners come back online after a curtailment window, the system adjusts to current chip and thermal conditions and restores optimized performance without operator input. For facilities that participate in frequent curtailment events — which is common in ERCOT during the summer months — this automation meaningfully reduces the operational burden of repeated ramp cycles.
Curtailment as a Component of the Power Rate Equation
The financial case for curtailment participation is most clearly understood when it is modeled as part of the effective electricity rate rather than as a separate revenue line. A facility paying $0.055/kWh for power and earning $0.008/kWh equivalent in demand response credits over the course of a month is effectively operating at a net rate of approximately $0.047/kWh for that period. At mining scale, a reduction of that magnitude has a material impact on profitability.
The degree to which curtailment revenue offsets power costs depends on several factors: the number and duration of curtailment events in a given month, the compensation rate per MW reduced, the facility’s enrolled capacity, and its reliability in meeting dispatch requirements. In high-volatility markets like ERCOT, curtailment events during summer peak demand periods can be both frequent and financially significant. In markets with lower volatility, curtailment may be a more modest supplement to other cost-reduction efforts.
What makes curtailment particularly valuable in a post-halving environment is that it reduces effective electricity cost without requiring capital expenditure or changes to the mining hardware. Combined with firmware-level efficiency improvements that reduce power consumption during normal mining hours, curtailment participation represents one of the few levers available to enterprise operators that affect both sides of the operating cost equation simultaneously.
Practical Considerations for Curtailment Program Participation
For operators evaluating demand response participation, several practical questions shape the approach:
Market eligibility.
In ERCOT, facilities with 75 MW or greater of demand can participate directly. Smaller facilities typically participate through aggregators such as Voltus, which pool loads to meet market participation thresholds. In Alberta, participation mechanisms vary depending on whether the facility is grid-connected or operating with behind-the-meter generation.
Performance requirements.
Demand response programs specify the response time from signal to load reduction, the minimum and maximum duration of curtailment events, and the accuracy of load reduction required. Facilities that fail to meet these requirements may face penalties or removal from the program. Firmware that enables precise, programmable power control reduces the risk of performance failures.
Interconnection and metering.
Accurate, real-time metering of facility load is a prerequisite for demand response participation. Grid operators and aggregators need to verify that the committed load reduction was actually delivered.
Operational integration.
For most enterprise mining operations, the demand response signal needs to be integrated into the existing fleet management workflow. Whether this is done through direct API integration with the curtailment aggregator, a manual process triggered by a notification, or an automated system that monitors price signals and triggers ramp-down automatically depends on the facility’s operational structure and the program requirements.
UMC OS’s API is designed to support this integration layer, allowing fleet power targets to be set programmatically in response to external signals. For operators building a curtailment-ready fleet management stack, the UMC OS product documentation covers API capabilities and integration requirements in detail.