What Evergy’s Large-Load Tariff Means for Your Electric Bill
Data centers on Evergy’s system now pay a premium rate, sign contracts of up to 17 years and post two years of collateral. The first measurable effect on a household bill is about $3.20 a month in your favor, in a rate case that has not been decided yet.
On Monday, October 26, Evergy emailed its customers to explain how it charges data centers. The message was short: its rates exist so that current customers are protected “while data centers pay their fair share.” A day later, Evergy’s chief customer officer, Chuck Caisley, told KCTV5 that large-load customers pay about 20% more than similar customers, and that Evergy has likely turned away “dozens” of projects that did not meet its terms.
The timing matters. Data centers came up in Kansas City races this year, from Congress to Jackson County Executive, and voters go to the polls on Tuesday, November 3. Prices are rising everywhere: the U.S. Energy Information Administration expects the average residential price to reach 18.2 cents per kilowatt-hour in 2026, up nearly 5% from 2025.
So does the tariff protect your bill? I went through both states’ decisions, the filings in the pending Missouri Metro rate case and Evergy’s latest earnings call, then built a small model you can run with your own assumptions. The short answer: the tariff is well built for the first 12 to 17 years, the first dollars are already flowing toward households, and the real risk sits in the years after the contracts end.
The short version
- The tariff covers any new or expanded load of 75 MW or more. It has applied in Kansas since November 6, 2025 and in Missouri since March 9, 2026. Customers sign for 12 years after a ramp of up to 5, pay a minimum bill based on 80% of contracted demand and post collateral equal to two years of minimum bills.
- Evergy has signed about 2.5 GW under it. Five agreements: two with Google and one each with Meta, Digital Realty and Beale. Another 500 MW of smaller large loads sit outside the tariff.
- The first bill effect is in the Missouri Metro rate case. Evergy cut its request by $25 million for expected data center revenue. In proportion, that lowers the average residential increase from about $20.91 a month to the $17.70 Evergy filed, and to about $15.39 if the offset grows to $43 million.
- A 300 MW data center’s premium is worth about $2.57 a month to each household in a proposed model of a Kansas City-sized utility. Every $10 million a year of cost the tariff fails to cover would add about $1.34.
- The contracts cover roughly 61% to 79% of a new power plant’s cost in present value. What happens after year 17 is the open question, and it is a question about load growth more than about this tariff.
What the Evergy large load tariff actually says
Evergy calls it Large Load Power Service, or LLPS. It is one rate plan approved twice, once by each state, with nearly the same terms. Evergy filed in Kansas in February 2025. Missouri’s version followed Senate Bill 4, the 2025 law that tells the Public Service Commission to set large-load rates that reflect those customers’ share of costs and keep unjust costs off everyone else.
- Cases: KCC Docket 25-EKME-315-TAR in Kansas, approved as a unanimous settlement on November 6, 2025. PSC File EO-2025–0154 in Missouri, decided November 13, 2025, with tariffs effective March 9, 2026.
- Who it covers: any new load of 75 MW or more, or an existing customer that expands by 75 MW, in both states.
- Contract term: up to 5 years of ramp, then at least 12 years at full contract capacity, for up to 17 years in total.
- Minimum bill: based on 80% of contracted demand, owed even in months the site uses less.
- Collateral: two years of minimum bills, as cash, a letter of credit or, for very large companies, a parent guarantee. Kansas lets strong credit reduce it.
- Leaving early: in Kansas, after year 5 a customer can cut the lesser of 25 MW or 10% at no charge with 24 months’ notice; bigger cuts need 36 months’ notice and a fee. In Missouri, 24 months’ notice avoids the early termination fee, but an exit fee still applies.
- Price premium: KCC staff estimated 7% to 10% above existing industrial customers. In Missouri, Google’s rate design witness puts the settled rates about 15% to 17% above standard large-power rates.
- Optional riders: Kansas adds clean energy, customer-owned capacity and demand response riders; Missouri lets large customers fund extra nuclear, clean energy or capacity programs at their own expense.
On top of these terms, the customer pays 100% of the substations and lines built only to serve it. In Kansas, any project of 25 MW or more also enters Evergy’s “Path to Power” review, with at most four projects in the active queue at a time. Caisley says Evergy can also “turn down or turn off” large loads during peaks to protect supply for everyone else.
The two states got there differently. In Kansas, every party signed the settlement, including the Citizens’ Utility Ratepayer Board, Sierra Club, NRDC, Google and the Data Center Coalition. In Missouri, the commission’s own staff and the Office of the Public Counsel backed a different design. Chair Kayla Hahn rejected it as too confusing, saying it carried more than 25 possible charges.
Two groups sit outside the tariff entirely. Meta’s first Kansas City phase came online before the tariff existed, though Evergy says future Meta expansion falls under it. And city-owned utilities and electric cooperatives are not regulated by either commission, which is why the 800 MW Nebius project in Independence is a separate story.
The one equation behind “fair share”
A utility’s rates recover a revenue requirement: a return on everything it has built, depreciation on it, and operating costs. In a rate case, the commission splits that requirement among customer classes. When a large new customer arrives, two numbers move: the cost the utility takes on to serve it, and the revenue it pays. Your bill moves by the difference, scaled to your class’s share.
ΔB = s × (ΔC − ΔR) / (N × 12)
Here ΔB is the monthly change on one household’s bill. ΔC is the annual cost the new load adds to the shared system, meaning cost not billed directly to it. ΔR is the annual revenue it pays toward shared costs, premium included. s is the share of any change allocated to residential customers, and N is the number of households. A negative ΔB is a lower bill.
The tariff works on both terms. Direct assignment pushes ΔC toward zero for the wires and substations built only for the data center. The premium and the minimum bill hold ΔR up even when the site runs below plan. Generation is the hard part. A new gas plant serves everyone, so its cost lands in ΔC, and the data center’s demand charges have to outrun it for the full life of the plant.
The script below turns that into numbers. It uses only the standard library.
"""llps_bill.py: what a large-load tariff can do to a household electric bill.
Proposed model, not Evergy's method. Every input is labeled in ASSUMPTIONS
or comes from a public filing (see the article's sources).
"""
from dataclasses import dataclass
@dataclass
class Utility:
residential_customers: int # households sharing the residential allocation
residential_share: float # share of a revenue change allocated to residential
def household_monthly(annual_dollars: float, u: Utility) -> float:
"""Monthly bill change per household from an annual revenue change.
Positive = households pay more, negative = they pay less.
"""
return annual_dollars * u.residential_share / u.residential_customers / 12
def scale_increase(filed_increase: float, filed_bill_impact: float,
offset_change: float) -> float:
"""Re-scale a filed residential bill impact when the offset changes.
Assumes the residential impact moves in proportion to the total request.
"""
return filed_bill_impact * (filed_increase + offset_change) / filed_increase
def contract_coverage(contract_years: int, asset_life: int, rate: float) -> float:
"""Share of a level annual cost's present value that falls inside the contract."""
pv = lambda n: (1 - (1 + rate) ** -n) / rate
return pv(contract_years) / pv(asset_life)
if __name__ == "__main__":
# Evergy Missouri Metro rate case ER-2026-0143 (filed Feb 6, 2026)
filed = 137.9e6 # requested increase, $
filed_bill = 17.70 # average residential increase, $/month
llps_offset = 25e6 # expected large-load revenue already netted out
trued_up_offset = 43e6 # large-load revenue at true-up, per Google's position statement
no_dc = scale_increase(filed, filed_bill, +llps_offset)
true_up = scale_increase(filed, filed_bill, -(trued_up_offset - llps_offset))
print(f"Without large-load revenue: ${no_dc:.2f}/month")
print(f"As filed ($25M offset): ${filed_bill:.2f}/month")
print(f"At $43M offset: ${true_up:.2f}/month")
print(f"Offset share of gross ask: {llps_offset / (filed + llps_offset):.1%}")
# ASSUMPTIONS: a Kansas City-sized utility (change these)
u = Utility(residential_customers=280_000, residential_share=0.45)
# Premium a 300 MW transmission-voltage load pays above standard large-power
# rates, per Google's rate design witness (settled rates vs Evergy's proposal)
for label, premium in [("settled", 19.2e6), ("Evergy proposal", 22.1e6)]:
print(f"300 MW premium, {label}: {household_monthly(-premium, u):+.2f} $/month")
# Exposure if $10M a year of cost is left uncovered
print(f"Per $10M/yr uncovered: {household_monthly(10e6, u):+.2f} $/month")
# How much of a generator's cost the contract term covers (level cost, 7% discount)
for years in (12, 17):
for life in (30, 35):
c = contract_coverage(years, life, 0.07)
print(f"{years}-yr contract, {life}-yr asset: {c:.1%} of PV covered")Run it with python3 llps_bill.py and you get:
Without large-load revenue: $20.91/month As filed ($25M offset): $17.70/month At $43M offset: $15.39/month Offset share of gross ask: 15.3% 300 MW premium, settled: -2.57 $/month 300 MW premium, Evergy proposal: -2.96 $/month Per $10M/yr uncovered: +1.34 $/month 12-yr contract, 30-yr asset: 64.0% of PV covered 12-yr contract, 35-yr asset: 61.3% of PV covered 17-yr contract, 30-yr asset: 78.7% of PV covered 17-yr contract, 35-yr asset: 75.4% of PV covered
Two inputs are mine, not Evergy’s: 280,000 households and a 45% residential share. Evergy Missouri Metro serves about 313,000 customers of all classes, so treat those as a stand-in for a Kansas City-sized utility and replace them with your own. The rate case figures and the 300 MW premiums come from filings, and the next two sections walk through them.
Where the money shows up first: the Missouri Metro rate case
The first place data center revenue touches a household bill is Evergy Missouri Metro’s rate case, ER-2026–0143. Evergy filed it on February 6, 2026, asking for $137.9 million, a 14.92% increase, or about $17.70 a month for an average residential customer. If approved, new rates take effect in January 2027.
Evergy says the request includes no costs of serving data centers. It is driven by about $500 million of grid investment since 2022. What the filing does include is a credit: Evergy cut the request by $25 million for revenue it expects from new large data centers in 2026. Without that credit the ask would have been $162.9 million, so data center revenue offsets 15.3% of the gross need.
The credit can grow. In a position statement filed September 25, Google says the large-load revenue adjustment could true up to $43 million or more. Google also says $3.8 million of the current $25 million is the premium itself, the part above standard large-power rates, which flows to the other customer classes.
Those two end bars are estimates. They assume the residential increase moves in proportion to the total request, and the commission’s final rate design may not split it that way. Still, the direction is clear: in this case, data center revenue is lowering what households are asked to pay, not raising it.
The same case holds a fight over how big the premium should be. Evergy proposes raising every class by 15.19%, the large-load rates included. Google’s rate design witness, Carolyn Berry, argues against it: for a 300 MW transmission-voltage site she puts the premium at about $19.2 million a year under the settled rates and $22.1 million under Evergy’s proposal. Run through the model, that is the difference between $2.57 and $2.96 a month per household. For scale, $2.57 is about 1.9% of an average Evergy Kansas Metro bill, which was $136.02 at 1,000 kWh in 2026.
Evergy’s own outlook on its August earnings call: residential increases “in line with or below inflation” for most customers. The exception is Missouri West, its smallest utility with the lowest rates, where increases may run above inflation for five years as new generation comes online.
What each protection guards against, and the 17-year gap
Each clause in the tariff answers one way a data center deal can go wrong. Read them as insurance policies against specific failures.
The minimum bill covers a slow ramp. If a campus promised 300 MW and draws 150 MW for two years, it still pays demand charges on 240 MW, so its demand revenue falls by a fifth instead of a half. Collateral covers a developer that disappears: two years of minimum bills sit in cash, a letter of credit or a parent guarantee. Notice periods and exit fees cover a customer that leaves on purpose, and they give Evergy two to three years to redirect capacity. Direct assignment covers dedicated wires. The demand response rider and Evergy’s right to curtail cover the few hot afternoons a year when supply is tightest.
What no clause covers is time. A new gas plant is a 30 to 35 year asset, and its cost sits in rates for that whole life. The longest contract is 17 years. The contract_coverage function in the script measures how much of a plant's cost the contract term spans. It assumes a level annual cost and a 7% discount rate, both mine.
The result: a 12-year contract spans 61.3% to 64.0% of the plant’s cost in present value, and a full 17-year contract spans 75.4% to 78.7%. The remaining 21% to 39% falls in years when no data center contract is required to pay it. Real utility revenue requirements are front-loaded, because the return shrinks as the plant depreciates, so the true coverage is somewhat higher than this level-cost version.
That tail is only a problem if nobody else needs the plant. Evergy’s numbers say demand will be there. It has about 3 GW of large loads signed: 1.3 GW already operating and ramping, and 1.7 GW in agreements with minimum bills running 16 to 17 years. It expects retail sales to grow 7% to 8% a year through 2030, and it describes a pipeline of “well over 10 additional gigawatts,” mostly after 2030. To serve that, it plans more than 5 GW of new generation through 2032, about 3.9 GW of it natural gas, inside a $21.6 billion five-year capital plan plus about $1 billion more.
If that growth arrives, a plant that outlives its first contract simply serves the next customer, and the shared fixed cost is spread over more kilowatt-hours. If it does not, the model gives a quick rule: every $10 million a year of plant cost left without a payer adds about $1.34 a month to each household bill.
Where the tariff, and this model, can break
The proportional estimates are the weakest link. The $20.91 and $15.39 figures assume the residential share of the Missouri Metro increase moves with the total. Class cost studies, rate design and any settlement can move that split, and the case is still in front of the commission.
The premium is negotiated, not computed from risk. The 7% to 10% Kansas estimate and the 15% to 17% Missouri figure come from settlements and testimony, not from a model of stranded-asset exposure. Missouri’s commission staff and the Public Counsel wanted a design built from many separate charges and lost. Whether the chosen premium is large enough is an empirical question that only the next decade answers.
The post-contract tail is real. Coverage of 61% to 79% means a share of every plant built for this growth outlives the contracts that justified it. Load growth makes that harmless. A sharp drop in AI demand after 2035 would not.
Regional transmission is a separate bill line. The tariff makes large customers pay for upgrades needed to serve them. It does not govern the wider Southwest Power Pool build-out, whose costs reach households through transmission riders. On an average Evergy Kansas Central bill, that rider was $25.61 a month in 2026. Whether large-load growth across the region pushes that line up is worth watching on its own.
Not every data center is on this tariff. Meta’s first Kansas City phase predates it. City-owned utilities and cooperatives answer to their own boards, so the Nebius campus in Independence, expected to need more than 800 MW, falls outside both commissions. Its contract requires power from a new, privately funded plant.
What to watch, and how to check it yourself
Four dates matter over the next few months. Voters decide local and congressional races on November 3. Evergy reports third-quarter results the same week; its CEO said in August he was “highly confident” of at least one more large-load agreement in 2026, with details due on that call. The commission’s decision in the Missouri Metro case sets the rates that take effect in January 2027. And Evergy is taking new gas, solar and battery plants to both commissions, including Mullin Creek #2, a 440 MW gas turbine in Missouri whose hearings began October 19.
You do not need to take anyone’s word for the numbers, mine included. Missouri filings are public in the PSC’s EFIS system: search ER-2026–0143 for the rate case and EO-2025–0154 for the tariff. Kansas filings sit under KCC Docket 25-EKME-315-TAR, and the Citizens’ Utility Ratepayer Board publishes plain-language summaries for residential customers. When the Missouri Metro order comes out, put its residential figure and its final large-load credit into scale_increase and you will see how much of the promise held.
My read: on paper, the tariff has the protections other states are still debating, from long terms and an 80% minimum bill to collateral and exit fees, and the first evidence in a live rate case points the right way for households. The honest open question is not whether data centers pay their share for the next 17 years. It is who pays for the plants in the years after that, and that depends on whether today’s forecasts of growth come true.
Keep reading. Last week I argued that Evergy could add capacity without building all of it, through a curtailable compute tariff that pays flexible AI loads to step down at peak. On November 6, I’ll break down Evergy’s third-quarter data center update in five numbers. Subscribe on Substack to get each piece when it publishes.
Brian Curry writes about applied machine learning, causal inference and forecasting for healthcare, energy, commodities and marketing at Vector1 Research in Kansas City.
Sources
- Evergy says data centers won’t raise your rates. Here’s how that works, KCTV5, October 27, 2026.
- Evergy executive answers questions about data centers after email to customers, KCTV5, October 28, 2026.
- KCC Approves Large Load Rate Plan for Evergy, Docket №25-EKME-315-TAR, Citizens’ Utility Ratepayer Board newsletter, Q4 2025.
- Kansas, Michigan regulators approve large load rules for Evergy, Consumers Energy, Utility Dive, November 10, 2025.
- Missouri passes new rules, costs for data centers and large power users, Kansas Reflector, November 14, 2025.
- Order Approving Tariffs, File No. EO-2025–0154, Missouri Public Service Commission, February 25, 2026.
- Large Load Tariff Fact Sheet, Missouri Public Service Commission, April 2026.
- Evergy asks Missouri Public Service Commission to allow recovery of investments for Missouri Metro customers, Evergy, February 6, 2026.
- Intervention Deadline Set in Evergy Missouri Metro Rate Case, Missouri Public Service Commission, 2026.
- Public 2026 Missouri Metro Rate Case Summary, Evergy, February 6, 2026.
- Google LLC Position Statement, File No. ER-2026–0143, Missouri Public Service Commission, September 25, 2026.
- Direct Testimony of Carolyn A. Berry on rate design for large load customers, File No. ER-2026–0143, Missouri Public Service Commission, July 14, 2026.
- Evergy (EVRG) Q2 2026 Earnings Call Transcript, The Motley Fool, August 2026.
- Evergy has 3 GW of large loads signed and a 5-GW generation pipeline, Utility Dive, August 11, 2026.
- Evergy Kansas Central residential rates, 2026 and Evergy Kansas Metro residential rates, 2026, Citizens’ Utility Ratepayer Board, July 2026.
- KSHB report on Evergy’s large load rate and electric bills, March 6, 2026.
- KSHB voters guide on Missouri candidates’ data center stances, July 16, 2026.
- Utility Dive report on EIA’s May 2026 Short-Term Energy Outlook, May 15, 2026.
Method notes. The household model is a proposed model, not Evergy’s or either commission’s method. Assumed inputs: 280,000 households, a 45% residential share, a level annual plant cost, a 30 to 35 year plant life and a 7% discount rate. The $20.91 and $15.39 figures scale Evergy’s filed $17.70 in proportion to the total request. The $43 million true-up figure and the 300 MW premiums come from Google’s filings in ER-2026–0143. Figures current as of October 28, 2026.
