Yes — even in the most coal/gas-reliant state grids, EVs cut emissions vs. gasoline cars

Summary

Yes. Even in the two most coal-reliant state electric grids in the country -- West Virginia (87% coal, 93.6% coal+gas combined) and Kentucky (65% coal, 92% coal+gas combined) -- an EV charged on that state's own grid produces meaningfully less CO2 per mile than a gasoline car. Using each state's actual 2025 generation mix and EIA's own fuel-specific CO2 emission rates, a grid-charged EV in West Virginia works out to about 282 grams CO2/mile, and in Kentucky about 240 g/mi -- both well below the EPA's average U.S. gasoline car (400 g/mi) and even below a 27-mpg efficient compact gasoline car (329 g/mi). Every other coal/gas-heavy state tested here (Missouri, Wyoming, Indiana, Utah, Wisconsin, Ohio, Pennsylvania) shows the same pattern, with the margin widening as gas (lower-carbon than coal) makes up more of the fossil mix. This matches EPA's own published finding for West Virginia specifically (~270 g/mi via the Beyond Tailpipe Emissions calculator) and the Union of Concerned Scientists' 2022 finding that EVs beat even the most efficient gasoline compacts on the dirtiest U.S. grids.

Which states actually burn the most coal and gas

Using EIA's state-level electricity generation data (energy.eia_electricity_generation, filtered to the annual/all-sector rollup rows), the states with the highest share of 2025 generation from coal + natural gas combined were:

West Virginia and Kentucky are the two genuinely coal-heavy grids; several others on this list (Delaware, Rhode Island, Mississippi, Florida, Louisiana, Massachusetts, Pennsylvania) are fossil-heavy mainly through natural gas rather than coal, which matters because gas emits roughly 40% as much CO2 per kWh as coal at the plant (EIA 2023: coal 2.31 lb CO2/kWh generated, natural gas 0.96 lb/kWh).

Converting grid mix to an actual EV-vs-gasoline comparison

A percentage of coal on the grid does not by itself answer the question -- what matters is grams of CO2 per mile driven, comparably for both vehicle types. The calculation:

1. Grid carbon intensity = (state coal % x 2.31 lb CO2/kWh) + (state gas % x 0.96 lb CO2/kWh), using EIA's 2023 national average fuel-specific generation emission rates (nuclear, hydro, wind, solar treated as zero-carbon at the point of generation, consistent with EIA's own convention).
2. EV consumption: DOE's average EV efficiency of about 0.30 kWh per mile (~3.3 miles/kWh), applied at the wall/vehicle level.
3. Gasoline benchmarks: EPA's average U.S. passenger vehicle, 400 g CO2/mile (22.2 mpg fleet average); and a more efficient 27-mpg compact gasoline car (8,887 g CO2/gallon / 27 mpg = 329 g/mile), the same benchmark the Union of Concerned Scientists uses in its own regional EV-emissions comparisons.

Results for the coal/gas-heaviest states:

In every single case, even the dirtiest grid tested (West Virginia) still produces a lower-emissions mile for an EV than for a gasoline car of either benchmark. The margin shrinks as coal's share rises (coal is the dominant driver, since it emits roughly 2.4x as much CO2/kWh as gas), but it never reverses.

Cross-checking against independent sources

This computed figure is not the only evidence. Two independent checks corroborate it:

• EPA's own Beyond Tailpipe Emissions calculator (fueleconomy.gov), which EPA explicitly built for this exact question, reports that in West Virginia "where a lot of electricity is generated from coal-fired plants," a representative EV is responsible for about 270 grams of CO2 per mile driven -- within about 4% of the 282 g/mi computed independently here from raw EIA generation data.
• Union of Concerned Scientists (2022, "State of Charge" / EV emissions research) found that nationwide, EVs produce lower global-warming emissions than the average compact gasoline car (27 mpg) "even when the electricity is produced primarily from coal in regions with the 'dirtiest' electricity grids" -- the same conclusion this analysis reaches independently for West Virginia and Kentucky specifically.
• EPA's own explainer ("Electric Vehicle Myths") states plainly: "Even accounting for these electricity emissions, research shows that an EV is typically responsible for lower levels of greenhouse gases than an average new gasoline car," citing Argonne National Laboratory's GREET model lifecycle analysis (which also finds EV manufacturing emissions are more than offset over the vehicle's life).

All three independent sources agree with the state-level calculation above: the answer does not flip even in the states that burn the most coal and gas.

What this does and doesn't cover

Data gap disclosed: environment.egrid_emission_rates -- this corpus's own EPA eGRID state/subregion CO2-rate table, purpose-built for exactly this kind of question -- returned zero rows when queried (declared coverage 2016-2024, but the table is empty). This is a genuine sourcing gap, not a finding of "no coal-heavy state data exists"; it was reported and the analysis was built instead from raw EIA generation-mix data (energy.eia_electricity_generation) combined with EIA's own published national fuel-specific CO2 rates, then cross-checked against EPA's calculator and UCS's published research as shown above.

Vintage caveat: energy.eia_electricity_generation's monthly detail is only actually loaded for 2024 (a single month, December) and 2025 (10 of 12 months -- February and March are missing). The state fuel-mix percentages above use the 10 available 2025 months; this is close to but not a complete calendar year, and monthly generation mix shifts modestly with weather-driven demand (e.g., more gas peaking in summer). This would move individual state percentages by low single digits, not enough to change which states rank as coal/gas-heavy or to flip the EV-vs-gasoline conclusion.

Scope not covered: this compares tailpipe-equivalent CO2 per mile only (the standard "does the electricity source make EVs dirtier than gas" question). It does not cover full lifecycle emissions (battery/vehicle manufacturing), which EPA and Argonne National Laboratory's GREET model separately find still favor EVs over a vehicle's lifetime even after accounting for higher manufacturing emissions, nor does it cover non-CO2 pollutants (NOx, SO2, particulates) or costs.

Every query behind this report

9 warehouse calls ran in this session, in order. Each is reproducible against the same snapshot.

SELECT subregion_acronym, subregion_name, "year", co2_output_rate_lb_per_mwh, coal_generation_pct, gas_generation_pct, coal_co2_rate_lb_per_mwh, gas_co2_rate_lb_per_mwh FROM environment.egrid_emission_rates WHERE "year" = (SELECT MAX("year") FROM environment.egrid_emission_rates) ORDER BY co2_output_rate_lb_per_mwh DESC

SELECT "year", COUNT(*) FROM environment.egrid_emission_rates GROUP BY "year" ORDER BY "year"

SELECT state_abbr, state_name, fossil_pct, renewable_pct, total_gen_thousand_mwh, fossil_gen_thousand_mwh, months_covered FROM energy.state_energy_mix WHERE generation_year = 2023 AND months_covered = 12 ORDER BY fossil_pct DESC LIMIT 15

SELECT COUNT(*) FROM environment.egrid_emission_rates

SELECT generation_year, months_covered, COUNT(*) FROM energy.state_energy_mix GROUP BY generation_year, months_covered ORDER BY generation_year

SELECT generation_year, generation_month, COUNT(DISTINCT state_abbr) AS n_states FROM energy.eia_electricity_generation WHERE geo_level='state' AND sector_code='99' AND generation_year=2024 GROUP BY generation_year, generation_month ORDER BY generation_month

SELECT generation_month, COUNT(DISTINCT state_abbr) AS n_states FROM energy.eia_electricity_generation WHERE geo_level='state' AND sector_code='99' AND generation_year=2025 GROUP BY generation_month ORDER BY generation_month

SELECT energy_source_code, fuel_is_rollup, COUNT(*) FROM energy.eia_electricity_generation WHERE geo_level='state' AND sector_code='99' AND generation_year=2025 GROUP BY energy_source_code, fuel_is_rollup ORDER BY energy_source_code

WITH agg AS (
  SELECT state_abbr, state_description,
    SUM(CASE WHEN energy_source_code='ALL' THEN generation_thousand_mwh ELSE 0 END) AS total_mwh,
    SUM(CASE WHEN energy_source_code='COW' THEN generation_thousand_mwh ELSE 0 END) AS coal_mwh,
    SUM(CASE WHEN energy_source_code='NG' THEN generation_thousand_mwh ELSE 0 END) AS gas_mwh
  FROM energy.eia_electricity_generation
  WHERE geo_level='state' AND sector_code='99' AND generation_year=2025
    AND energy_source_code IN ('ALL','COW','NG')
  GROUP BY state_abbr, state_description
)
SELECT state_abbr, state_description, total_mwh, coal_mwh, gas_mwh,
  ROUND(100.0*coal_mwh/NULLIF(total_mwh,0),1) AS coal_pct,
  ROUND(100.0*gas_mwh/NULLIF(total_mwh,0),1) AS gas_pct,
  ROUND(100.0*(coal_mwh+gas_mwh)/NULLIF(total_mwh,0),1) AS coal_gas_pct
FROM agg
WHERE total_mwh > 0
ORDER BY coal_gas_pct DESC
FETCH FIRST 15 ROWS ONLY