Utility program deep dive · Colorado and Minnesota
Xcel Energy C&I battery, demand charge and VPP economics
How Colorado’s on-peak demand charges and aggregator VPP, and Minnesota’s demand charges, curtailment discounts and hosted batteries, stack up for a business battery.
VPP Economics has no affiliation with, sponsorship from, or endorsement by any utility, aggregator, platform, or device maker named here.
The short answer
- Colorado’s new rate design does most of the work. Since April 2026, large business customers pay generation and transmission demand charges only on their highest demand in the 4–9 p.m. summer or 5–9 p.m. winter weekday window (Xcel Energy). A battery can hold that window down reliably, which is much easier than capping a 24-hour peak. On secondary voltage, that demand is billed at $15.14 per kW in summer and $9.11 otherwise (Xcel Colorado tariff).
- The aggregator VPP is the largest possible payment. Colorado’s Aggregator Virtual Power Plant values capacity at up to $264.56 per kW-year to aggregators (Xcel Energy). In our central Colorado case, bill savings of about $35k a year plus a 70% share of aggregator pay bring a 250 kW / 1,000 kWh battery to a payback of about 7.5 years with the 30% federal credit.
- Minnesota relies on demand charges and curtailment discounts. General Service demand costs $16.49 per kW in summer and $11.90 otherwise (Xcel Minnesota rate book). Electric Rate Savings discounts demand the customer agrees to curtail (Xcel Energy). Without a storage rebate or market payment, the same battery takes about 13 years.
- Hosting is the capital-free alternative. Under Capacity*Connect, Xcel owns 1–3 MW batteries on business sites in Minnesota and pays the host for the space (Utility Dive, Sparkfund). The host gets lease income but none of the bill savings.
What each Xcel program is
Time-differentiated demand (CO)
Generation and transmission demand is billed on the peak inside the on-peak window; distribution demand is still billed on the monthly maximum, with a 50% ratchet on the prior 12 months (Xcel Energy, Xcel Colorado tariff).
Aggregator Virtual Power Plant (CO)
Aggregations of at least 100 kW, up to 80 events of up to four hours a year, paid per kW-year for generation, transmission and, on eligible feeders, distribution capacity (Xcel Energy). Orchestrated by the aggregator rather than scheduled by the utility.
Peak Partner Rewards
Colorado: $2 per kW-month reserved plus $0.70 per kWh in events, with a 25 kW minimum (Xcel Energy) and up to 15 events of up to four hours (Xcel Energy). Minnesota runs a similar program (Xcel Energy).
Electric Rate Savings (MN)
Lower demand charges on load the customer agrees to curtail, for sites with at least 50 kW of controllable load (Xcel Energy). Load not shed during a control period carries a $10 per kW penalty (Xcel Energy).
Capacity*Connect (MN)
Up to 200 MW of Xcel-owned, front-of-the-meter batteries at host sites by 2028, with a $430 million three-year budget, deployed with Sparkfund (Sparkfund, Canary Media). Critics argued customer-owned VPPs would cost less; one analysis put the Minnesota cost at $2,150 per kW against $624 in Colorado (Avanza Energy).
Battery incentives
Colorado’s Renewable Battery Connect pays $250 per kW up to $5,000 for small businesses (Xcel Energy); Minnesota’s storage incentive covers batteries up to 50 kWh (Xcel Energy). Both are small relative to a C&I project.
Value per kW-year
Value per kW-year by source
Bill savings and program pay for one kW of reliable battery capacity, before any sharing with an aggregator.
Swipe to see the full chart →
These are avoided costs, not payments. Demand savings appear on each month’s bill only if the battery holds the peak every day of the month. Colorado’s on-peak window makes that practical for generation and transmission demand. Distribution demand is harder: it is billed on any-time peaks with a 50% ratchet, so the central case counts only 75 of 250 kW.
Stacking needs care, depending on the measurement and verification (M&V) method. A Colorado site cannot enroll in the aggregator VPP and another Xcel demand management program at the same time. The aggregator VPP and demand-charge savings can stack only if event performance is measured at the battery itself. If it is measured against a meter baseline, a battery that already shaves the peak every weekday shows little extra reduction on event days. The program documents published so far do not settle which method applies to batteries.
The capital stack
Ten-year capital stack
Federal credit plus 10 years of avoided costs and program pay as a share of installed cost plus 10 years of maintenance. Past the dashed line, the package more than covers its cost.
Swipe to see the full chart →
In Colorado, on-peak demand savings and the federal credit cover about 83% of the battery’s 10-year cost, and an aggregator VPP payment takes it past break-even. Peak Partner Rewards adds much less. In Minnesota, the same battery covers about 85% of its cost with the credit and about 55% without it. That leaves hosting under Capacity*Connect or a load-specific case as the realistic paths.
Simple payback
Years to repay net cost from avoided costs and program pay, net of maintenance.
Swipe to see the full chart →
Assets and combinations
Colorado: shave the window
A four-hour battery fits the four- to five-hour on-peak window. A two-hour battery can cover only part of it, and the demand charge is set by the highest 15 minutes in the window.
Solar shifts the peak later
Solar reduces midday load, but the Colorado window runs to 9 p.m., so solar alone barely touches on-peak demand. Pairing it with storage covers the evening. Net metering credits exports in both states (EnergySage).
Minnesota: curtail and get the discount
Chillers, process loads and EV chargers can provide the controllable demand under Electric Rate Savings; a battery makes the commitment firmer.
Demand charges dominate
Managed fleet charging outside the Colorado window avoids G&T demand without a battery.
Financing and ownership
Own, borrow or host
A 250 kW / 1,000 kWh battery owned by a Colorado site on secondary general service, against hosting an Xcel-owned battery under Minnesota’s Capacity*Connect.
Swipe to see the full chart →
The developer’s side: Colorado shared savings
Hosting is one capital-free route; a developer-owned battery on the customer’s side of the meter is the other. The developer pays the full $600k, carries maintenance, insurance and performance risk, and prices all of it into the contract to earn a return on its capital (Morgan Lewis). Third-party storage contracts usually run 15 years or more (Peak Power), and lenders treat 15- to 20-year terms as financeable (Illinois Power Agency). The chart below values each deal from the developer’s side.
What the developer earns
Present value of what the developer receives, as a share of what it spends, by contract terms, for a Colorado site.
Swipe to see the full chart →
Colorado is a hard case for a developer. The Aggregator Virtual Power Plant is the closest thing to a capacity sale in Xcel territory, but its contracts run up to five years, so a 15-year deal cannot count on it after year 5; the model assumes Peak Partner Rewards afterward. Even with the aggregator pay, about $35k a year of demand savings on a $600k battery leaves the developer short. At the modeled cost, a developer cannot earn 8% and still leave the site a share of demand savings. Battery-service offers are not priced the other way around, as a fee on top of handing over the savings: shared-savings providers promote a guaranteed share from day one (Gridwealth Electric), and Peak Power’s projects in Westchester, New York, split bill savings and market revenue with the building owner (NYSERDA case study). A deal that cannot fund a share is usually not offered at all. A guaranteed 15% share would need an installed cost about 32% lower, near $410k, or a larger site where cost per kWh falls; providers such as Peak Power focus on facilities with more than 1 MW of load (Peak Power’s overview). The last row of the chart shows that 15% share. Minnesota’s answer is different: under Capacity*Connect, Xcel owns the battery itself and recovers its cost from all customers, so no developer needs a return from the host.
Additional Finance Options and Considerations
Captures the bill and the credit
Net cost about $420k after the 30% credit. Projects must meet the 2026 construction-start and foreign-entity sourcing rules (Sidley).
Shared savings or lease
A developer owns the battery, claims the credit and enrolls it with an aggregator. In Colorado, the site would get a share of demand savings only where a lower installed cost or a larger site lets the developer earn 8%.
Lease income, no capital
Capacity*Connect pays host sites monthly for space, typically for 15–20 years; Sparkfund describes the income as thousands of dollars a month (Sparkfund, Sparkfund). The battery sits in front of the meter, so it does not lower the host’s bill. There is no open enrollment: sites are chosen by grid need.
Up to five years
Aggregator VPP contracts with Xcel run up to five years, so the customer share should be fixed in writing for the same term.
Why a lease works here. A Capacity*Connect battery is distribution infrastructure, not a customer device. Xcel owns it, dispatches it from its control room and counts it as capacity in its resource plan (Sparkfund). The host supplies land and receives rent; it takes no bill, performance or equipment risk. The model needs three conditions, all present in Minnesota in 2026: a vertically integrated utility that plans its own capacity; a commission willing to let it own front-of-the-meter batteries and recover the cost from customers (Xcel’s bill analysis projects net savings of about 13¢ a month for an average home over 20 years; Sparkfund); and a large load willing to pay. Google committed $50 million through a large-load tariff in exchange for capacity credits (Sparkfund). Sparkfund’s fact sheet lists rate-basing, large-load investment and third-party infrastructure capital as funding routes (Sparkfund).
What ifs
The largest swing in Colorado
At 40% instead of 70%, the Colorado aggregator case loses about $10,000 a year and the payback lengthens by about two years.
Device versus meter baseline
If aggregator events are measured against a baseline, daily peak shaving and event pay overlap, and the stack shrinks toward the bill-only case.
November 2027
The commission will revisit Capacity*Connect and the VPP model (Utility Dive). A customer-owned aggregator program would change the Minnesota stack.
Construction-start deadlines
Losing the 30% credit adds about $180k to net cost and roughly three to six years to payback.
Demand charges move with rate cases
Colorado’s on-peak G&T demand design is new. Changes to the window or the charge would move the largest layer.
Eligible feeders only
Sites on constrained Colorado feeders can earn the distribution layer of the aggregator rate, up to the $264.56 ceiling (Xcel Energy).
Compare: Xcel Energy residential for households, and ComEd DG Rebate for C&I for a rebate-driven stack in PJM.
Method and caveats
An independent, hypothetical analysis for business owners, facility managers, program designers, developers and other readers, covering Xcel Energy’s Colorado and Minnesota service areas. Terms checked October 26, 2026.
Assumptions: 250 kW / 1,000 kWh battery at $600 per kWh; $15 per kW-year maintenance; 30% federal credit; Colorado secondary general service with 200 kW reliably shaved inside the on-peak window and 75 kW off distribution demand; Minnesota General Service with 150 kW shaved and 100 kW under Electric Rate Savings; aggregator share 70%; Peak Partner Rewards at six four-hour events a year.
Riders, taxes and franchise fees are excluded. Figures are illustrative avoided costs and program payments before financing, degradation and augmentation.
VPP Economics has no affiliation with, sponsorship from, or endorsement by Xcel Energy, Sparkfund, any aggregator or any company named here.