Converting One Solar Kilowatthour Yields 3,412 Btu or 8,617 Depending on the EIA Table

Table 10.1 of the Monthly Energy Review puts United States solar energy consumption at 180 trillion Btu for June 2026. That figure is not a measurement of heat. It is a count of kilowatthours multiplied by a conversion factor, and the Energy Information Administration publishes two different factors for that multiplication inside the same monthly report. One is 3,412 Btu per kilowatthour. The other, for a June 2026 period, is 8,617. Both are official, neither is a rounding artifact, and the choice between them moved the noncombustible-renewable share of energy consumption for utility-scale electricity generation from 6% to 15% in 2018. The agency explains the choice twice, in Appendix A and again in Appendix E. That explanation is what most citations of the 180 leave out.

Where the 3,412 Comes From

Start with the footnote attached to the solar column of Table 10.1. In EIA, Monthly Energy Review, Section 10, Renewable Energy, footnote h reads: "Solar photovoltaic (PV) and solar thermal electricity net generation (converted to Btu by multiplying by the heat content of electricity in Table A6), and solar thermal direct use energy." The wind footnote has no such clause. So the 180 is converted generation plus a direct-use quantity that was never a kilowatthour. The same table also states that "For hydroelectric power, geothermal, solar, wind, and biomass waste, production equals consumption," which is why the production and consumption panels of Table 10.1 print the same solar figure rather than two.

Follow the footnote to its destination. EIA, Monthly Energy Review, Appendix A, British Thermal Unit Conversion Factors explains why the factor is a constant at all: "There is no generally accepted practice for measuring the thermal conversion rates for power plants that generate electricity from hydro, geothermal, solar thermal, photovoltaic, and wind energy sources." The appendix then resolves it in one sentence: "Therefore, EIA uses the heat content of electricity, 3,412 Btu per kilowatthour." A second entry in the same appendix is blunter still: "The value of 3,412 Btu per kilowatthour, which is the heat content of electricity, is a constant."

At that factor, 180 trillion Btu divides out to about 52.8 billion kilowatthours, a derived figure that overstates generation by whatever the direct-use component contributes. Table 10.1 prints whole trillion Btu, so the 180 is rounded and anything computed from it inherits that rounding.

Where the 8,617 Comes From

Appendix A ends its noncombustible-renewable entry with a pointer: "See Appendix E for more information." That is where the second factor lives. EIA, Monthly Energy Review, Appendix E, Alternative Measures for the Energy Content of Noncombustible Renewables describes the alternative directly: "The fossil fuel equivalency approach converts the consumption of noncombustible renewable electricity (in kWh) to Btu by applying a fossil fuel equivalency factor, based on the fossil-fuels heat rate (Table A6)." The next sentence defines that rate: "The fossil-fuels heat rate is equal to the average thermal efficiency across fossil-fueled fired generating plants based on fuel consumption and net generation data reported to EIA."

The numeric series behind Table A6 carries the Description string "Electricity Net Generation, Fossil-Fueled Plants" and the Unit string "Btu per Kilowatthour" in EIA, Total Energy data browser extract of Table A6. Its annual values are 8,767 for 2020, 8,844 for 2021, 8,813 for 2022, 8,630 for 2023, and 8,617 for 2024. The Appendix A source note is headed "Approximate Heat Rates for Electricity Net Generation, Total Fossil Fuels." while the data series says Fossil-Fueled Plants: two strings for one concept.

Appendix E grants the alternative one use. The equivalency consumption "is useful for analysis when considering the amount of primary fossil fuel energy displaced by renewable energy sources." It then says that "The fossil fuel equivalency approach measures neither primary energy consumption nor fossil fuels actually displaced," and that "However, unlike the captured energy approach, the fossil fuel equivalency approach is not as directly tied to any real market or physical quantity." It adds a forward-looking caution: "Additionally, its use becomes increasingly problematic as noncombustible renewables begin to displace other renewables instead of fossil fuels." The fossil-fuel-equivalency tables are published anyway, as EIA, Total Energy data browser, Table E3, Primary Energy Consumption by Source, Fossil Fuel Equivalency Approach and its companion overview table.

Count the Branches: There Are Three, Not Two

A reader who stops after the two factors has miscounted the appendix. Appendix E lays out a third approach, incident energy, defined as the energy measurable at the input of the conversion device rather than at its output, and it enumerates that input resource by resource. For the resource at issue here the definition is: "For solar, the energy contained in the sunlight that strikes the panel or collector mirror." That would be a third Btu value for the same month, and a far larger one.

The third branch has no table, and the appendix says the data are not available. The captured energy approach, it notes, "is preferred by the UN International Recommendations for Energy Statistics (IRES) because the detailed data needed to estimate quantities of incident energy are not available now and are not likely to develop soon." The published pair is therefore a pair by data availability, not by conceptual completeness. Appendix E also gives the positive case for the Table 10.1 factor: "Captured energy reflects the primary energy captured for economic use and does not include losses."

The Measured Number Falls When Nothing Renewable Changes

Here is the consequence that the two-factor framing hides. The 3,412 is a constant. The 8,617 is not. It is a weighted average of a fossil generating fleet that keeps getting more efficient, and EIA, Today in Energy, EIA offers two approaches to compare renewable electricity generation with other sources traces the decline: "This increase in efficiency from combined-cycle units has led to a decline in the average heat rate of fossil fuel generation from 10,333 Btu per kilowatthour in 2001 to 9,213 Btu per kilowatthour in 2017, an 11% improvement, as shown above."

Run that through the conversion. That 2017 figure now reads 9,208 in the series. At 9,213, a kilowatthour of solar was worth about 2.70 times its captured-energy value. At 8,617 it is worth about 2.53 times. Both multiples are derived. Applied to the rounded 180 for June 2026, the 2024 factor yields about 455 trillion Btu and the 2017 factor would have yielded about 486 trillion Btu. That is a derived difference of roughly 31 trillion Btu produced by gas turbine engineering, with no change whatsoever in how much sunlight hit a panel.

Under fossil fuel equivalency, the better the fossil fleet performs, the less primary energy an unchanged renewable output is credited with. The measure moves against the thing it is used to advertise, which is the problem that turns up whenever a published ratio hides a moving denominator, as in the note on capacity utilization as a ratio.

The 2026 Factor Repeats the 2024 Value

The Table A6 extract shows 8,617 for the 2024 annual period, 8,617 for 2025, and 8,617 for 2026. The same flatness appears in the companion rows: coal reads 10,777 for each of those three periods after 10,745 for 2023, petroleum reads 11,200 after 11,465, and natural gas reads 7,754 after 7,721. Four heat-rate series freezing at once across three consecutive annual periods reads as a factor held pending data, though no EIA note says so.

That matters for how a June 2026 fossil-fuel-equivalency figure should be read. Today in Energy, not Appendix E, is where the factor's periodicity is stated: "EIA calculates the fossil fuel equivalency factor as an annual weighted-average heat rate for all generators to account for geographic and operational differences." An annual weighted average cannot be final for a year still in progress, so the current-period figure carries an older fleet until the fuel and generation data close. Section 10 points the same way, with Table 10.5 noting that "Small-scale solar energy data for all years, and utility-scale solar energy data for the current two years, are estimates." Two vintages are stacked in one printed number: an estimated generation quantity and a carried-forward conversion factor. Vintage problems of this kind recur across the EIA calendar, as in the case of one release day carrying two vintages.

EIA Already Published the Size of the Gap

The cleanest number for the gap is EIA's own, and it is not derived. For 2018 the agency wrote: "In 2018, the fossil fuel equivalence approach showed that primary energy consumption for electric power generation from noncombustible renewable sources was equal to consuming 5,989 trillion British thermal units (Btu), or 15% of the total energy consumption for utility-scale electricity generation in that year." It then decomposed that total: "This value includes 2,218 trillion Btu of energy transformed into electricity based on the constant heat content of electricity and an adjustment of 3,771 trillion Btu to reflect the weighted average heat rate for fossil fuel-fired power plants."

The adjustment is larger than the measured quantity, and the share consequence is stated outright: "Without adjusting primary energy for fossil fuel equivalence (see above chart), the noncombustible renewable share of total energy consumption for utility-scale electricity generation would be 6% in 2018 instead of the 15% share under the fossil fuel equivalency approach." Note also that 5,989 divided by 2,218 is 2.70, reproducing the 9,213 heat rate against the 3,412 constant. EIA's published pair and its published factor agree, which checks the arithmetic above.

The same release explains why the older convention was the headline one: "Traditionally, EIA used the fossil fuel equivalency approach to report noncombustible renewables' contribution to total primary energy, in part because the resulting shares of primary energy are closer to the shares of generated electricity." The front tables now run on the constant while the equivalency tables sit in an appendix: a presentational reversal, not a measurement change.

A Fourth Reading, With a Different Denominator

Step outside the primary-energy frame and the same kilowatthours land in a share built on a different base entirely. World Bank, World Development Indicators Metadata Glossary, series EG.FEC.RNEW.ZS defines its indicator as follows: "Renewable energy consumption is the share of renewable energy in total final energy consumption." Its methodology field states that "the denominator is the total final energy consumption of all energy products." A final-energy denominator excludes the conversion losses a primary-energy denominator includes, so a share computed that way is not comparable with either EIA figure, and the series at World Bank, Renewable energy consumption as a percent of total final energy consumption carries an IEA source attribution rather than an EIA one.

A third federal treatment weights the same output by neither Btu convention. Federal Reserve Board, G.17 Industrial Production and Capacity Utilization, Explanatory Notes states that "In the IP index, series that measure the output of an individual industry are combined using weights derived from their proportion in the total value-added output of all industries." That sentence governs how whole industries are combined, so the utilities industry enters the index on a value-added weight. How solar and gas-fired output are weighted inside that component is not stated on the page. Convention choices of this shape run through the national accounts too, as in the piece on market-based PCE and imputed prices.

Concrete Framework

Four questions settle which Btu figure a given renewable number is.

First, ask which table it came from. If it came from Table 10.1 or any Section 10 sector table, the solar footnote applies and the factor is the 3,412 constant. If it came from a Table E series, the factor is the fossil-fuels heat rate. A renewable Btu figure with no table citation is unusable, because the two readings differ by a factor near 2.53.

Second, ask what the number is a number of. Table 10.1 solar is electricity net generation converted, plus solar thermal direct use, and Table 10.5 splits it by facility size at a one megawatt nameplate boundary: "Data are for utility-scale facilities (combined generator nameplate capacity of 1 megawatt or more)." A quoted solar total that omits the small-scale component is a different quantity from one that includes it.

Third, ask which year the conversion factor belongs to, not which year the generation belongs to. For a 2026 month the factor currently reads 8,617, the same value shown for 2024. A year-over-year comparison of fossil-fuel-equivalency Btu mixes a change in generation with a change, or a non-change, in fossil efficiency.

Fourth, convert back to kilowatthours before comparing anything. Division by 3,412 recovers kilowatthours from an electricity-only captured-energy figure, and that layer is where the three approaches stop disagreeing.

What to Watch Next Week

Watch the next Monthly Energy Review for the Table A6 annual rows. If the current-period heat rate moves off the value it now shares with the earlier annual period, the carry-forward reading here is confirmed, every fossil-fuel-equivalency figure for the current year shifts with it, and the Section 10 front tables hold still.

Watch also the estimate flags in the Section 10 notes. Utility-scale solar is flagged as estimated for the current two years, so the generation quantity and the conversion factor can revise in the same release for different reasons. Separating those two revisions is the task when a renewable Btu total changes.

What Would Invalidate This

The printed body of Table A6 was not read. The appendix pages that carry the table served their source notes but not the table grid, so the column headers, the units line and the per-year values on the printed page remain unchecked. The heat rates used here come from EIA's own data browser extract of that table, read twice with identical results for the three most recent annual periods. If the printed table assigns those values to a different column than the extract's Description implies, the per-kilowatthour comparison would need restating.

The bodies of the Table E series were not read either. Those pages served titles and notes only. So the 455 trillion Btu for June 2026 is derived by applying 8,617 to a rounded 180, not read from Table E3. If the published Table E3 solar figure for that month differs from 455, the published figure governs and the derivation here is the thing that is wrong.

The carry-forward claim rests on pattern, not on an EIA note. Four heat-rate series showing identical values across three consecutive annual periods is strong evidence of a factor held pending data, but an explicit statement to that effect was not located, and the two footnotes on the Appendix E page were not read. An EIA note describing those repeated values as independently estimated would undercut the vintage argument in the section above.

The G.17 point is narrow. The notes quoted here establish the value-added weighting for the index as a whole; the physical unit behind the electric utilities component, and the agency supplying it, were not established within the collection budget. No non-EIA United States statistical agency was found that republishes these kilowatthours in Btu, so the cross-agency comparison rests on a different denominator rather than on a competing Btu value. Finally, the whole comparison assumes Table 10.1 solar and a Table E solar figure describe the same underlying generation. If the two presentations differ in scope as well as in factor, the ratio is not a clean 2.53, and the scope question in the framework above becomes the first one to ask rather than the second.