Where the numbers come from
Every figure on this site is listed below with its source, page, whether it is an actual or a forecast, and how confident we are. Data last built 2026-11-11.
Sources used
| ID | Document | Publisher | Type |
|---|---|---|---|
| S01 | 2025 Annual Report | GRIDCo | Primary document |
| S09 | EV registrations statement (DVLA via UN Ghana) | UN Ghana | Secondary |
| S14 | 2013 Annual Report | GRIDCo | Primary document |
| S15 | 2014 Annual Report | GRIDCo | Primary document |
| S16 | 2015 Annual Report | GRIDCo | Primary document |
| S17 | 2016 Annual Report | GRIDCo | Primary document |
| S18 | 2017 Annual Report | GRIDCo | Primary document |
| S19 | 2018 Annual Report | GRIDCo | Primary document |
| S20 | 2019 Annual Report | GRIDCo | Primary document |
| S21 | 2020 Annual Report | GRIDCo | Primary document |
| S22 | 2021 Annual Report | GRIDCo | Primary document |
| S23 | 2022 Annual Report | GRIDCo | Primary document |
| S24 | 2023 Annual Report | GRIDCo | Primary document |
| S25 | 2024 Annual Report | GRIDCo | Primary document |
Other sources are catalogued but not yet used for any figure: the Electricity Supply Plans, PURC monthly fact sheets and Energy Commission outlooks are being verified.
Where sources disagree
When two reports give different values for the same year, both are kept. The value used on the site is marked. Nothing is averaged. A check that settled most cases: loss volume divided by loss percentage gives the energy total a report must have used.
| Metric | Year | Status | Values (source, page) | Why |
|---|---|---|---|---|
| Energy transmitted | 2019 | Resolved | 17,887.00 used (S20, p. 18) 17,034.00 rejected (S21, p. 16) | Loss GWh / loss % = 17,898 GWh, matches the 2019 report (17,887); the 2020-report comparator (17,034) does not fit |
| Energy transmitted | 2020 | Resolved | 19,716.59 used (S24, p. 26) 18,819.00 rejected (S21, p. 16) | Loss GWh / loss % = 19,711 GWh, matches the 2023-report series (19,716.59); the 2020 report headline (18,819) does not fit |
| Energy transmitted | 2022 | Resolved | 22,478.46 used (S23, p. 24) 21,545.00 rejected (S24, p. 25) | Loss GWh / loss % = 22,488 GWh, matches the 2022 report and the 2023 report p26 series; the 21,545 comparator does not fit |
| Energy transmitted | 2024 | Unresolved | 23,692.00 used (S25, p. 25) | Published headline, and the 2025 growth of 9.15% is computed from it. But the same report says "over 24,600" and loss GWh / 3.8% implies about 25,000. Unresolved |
| Energy transmitted | 2025 | Unresolved | 25,860.00 used (S01, p. 25) | Published headline. Loss GWh (1,082) / 4.02% implies about 26,900, so the loss volume and the energy total may use different scopes. Unresolved |
| System availability (do not trend) | 2014 | Resolved | 99.79 used (S15, p. 18) 99.17 rejected (S17, p. 15) | Same-year narrative used; the 2016 report restates 99.17. Definitions may differ between reports |
| System availability (do not trend) | 2018 | Resolved | 99.48 used (S19, p. 17) 99.85 rejected (S22, p. 20) | Same-year narrative used; the 2021 report restates 99.85. Definitions may differ |
| System availability (do not trend) | 2020 | Resolved | 99.15 used (S21, p. 17) 99.69 rejected (S22, p. 20) | Same-year narrative used; the 2021 and 2024 charts show 99.69. Definitions may differ |
| System availability (do not trend) | 2021 | Resolved | 99.22 used (S22, p. 19) 98.89 rejected (S22, p. 20) | Same-year narrative used; the chart on p20 of the same report shows 98.89 |
| Transmission losses | 2021 | Resolved | 5.01 used (S23, p. 25) 5.15 rejected (S22, p. 20) | 5.01% x 21,466 GWh = 1,076 GWh, which equals the audited loss volume; 5.15% does not |
National electricity access (1 values)
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2025 | 89.40 | percent | actual | S01 | 11 | high | Rural 76.7%; urban 100% |
Energy transmitted (13 values)
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2013 | 12,930 | GWh | actual | S14 | 14 | medium | 2013 report states 12.93 TWh transmitted; no loss GWh available to cross-check |
| 2014 | 13,071.28 | GWh | actual | S15 | 25 | high | Single reported value, no conflict |
| 2015 | 11,690 | GWh | actual | S16 | 15 | high | Single reported value, no conflict |
| 2016 | 13,700 | GWh | actual | S17 | 15 | high | Single reported value, no conflict |
| 2017 | 14,310 | GWh | actual | S18 | 14 | high | Single reported value, no conflict |
| 2018 | 15,960 | GWh | actual | S19 | 16 | high | Single reported value, no conflict |
| 2019 | 17,887 | GWh | actual | S20 | 18 | high | Loss GWh / loss % = 17,898 GWh, matches the 2019 report (17,887); the 2020-report comparator (17,034) does not fit |
| 2020 | 19,716.59 | GWh | actual | S24 | 26 | high | Loss GWh / loss % = 19,711 GWh, matches the 2023-report series (19,716.59); the 2020 report headline (18,819) does not fit |
| 2021 | 21,466 | GWh | actual | S22 | 19 | high | Loss GWh / 5.01% = 21,477 GWh, matches 21,466; "2022" comparator label in the source is a typo for 2021 |
| 2022 | 22,478.46 | GWh | actual | S23 | 24 | high | Loss GWh / loss % = 22,488 GWh, matches the 2022 report and the 2023 report p26 series; the 21,545 comparator does not fit |
| 2023 | 23,550.99 | GWh | actual | S24 | 26 | high | Single reported value, no conflict |
| 2024 | 23,692 | GWh | actual | S25 | 25 | low | Published headline, and the 2025 growth of 9.15% is computed from it. But the same report says "over 24,600" and loss GWh / 3.8% implies about 25,000. Unresolved |
| 2025 | 25,860 | GWh | actual | S01 | 25 | low | Published headline. Loss GWh (1,082) / 4.02% implies about 26,900, so the loss volume and the energy total may use different scopes. Unresolved |
Substations and bulk supply points (6 values)
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2019 | 63 | count | actual | S20 | 5 | high | Single reported value, no conflict |
| 2021 | 68 | count | actual | S22 | 5 | high | Single reported value, no conflict |
| 2022 | 69 | count | actual | S23 | 6 | high | Single reported value, no conflict |
| 2023 | 66 | count | actual | S24 | 5 | high | Single reported value, no conflict |
| 2024 | 72 | count | actual | S25 | 9 | high | Single reported value, no conflict |
| 2025 | 72 | count | actual | S01 | 8 | high | Single reported value, no conflict |
System availability (do not trend) (12 values)
Note. The definition appears to change between reports: the same year is restated differently in later reports, and 2024 is blank. Do not read it as a trend until GRIDCo confirms the definition.
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2013 | 97.59 | percent | actual | S14 | 15 | high | Single reported value, no conflict |
| 2014 | 99.79 | percent | actual | S15 | 18 | low | Same-year narrative used; the 2016 report restates 99.17. Definitions may differ between reports |
| 2015 | 99.46 | percent | actual | S16 | 25 | high | Single reported value, no conflict |
| 2016 | 99.39 | percent | actual | S17 | 15 | high | Single reported value, no conflict |
| 2017 | 99.48 | percent | actual | S18 | 15 | high | Single reported value, no conflict |
| 2018 | 99.48 | percent | actual | S19 | 17 | low | Same-year narrative used; the 2021 report restates 99.85. Definitions may differ |
| 2019 | 99.43 | percent | actual | S20 | 20 | high | Single reported value, no conflict |
| 2020 | 99.15 | percent | actual | S21 | 17 | low | Same-year narrative used; the 2021 and 2024 charts show 99.69. Definitions may differ |
| 2021 | 99.22 | percent | actual | S22 | 19 | low | Same-year narrative used; the chart on p20 of the same report shows 98.89 |
| 2022 | 99.26 | percent | actual | S23 | 25 | high | Single reported value, no conflict |
| 2023 | 99.53 | percent | actual | S24 | 27 | high | Single reported value, no conflict |
| 2025 | 99.61 | percent | actual | S01 | 11 | high | Single reported value, no conflict |
System availability excluding the J2K line (1 values)
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2024 | 99.86 | percent | actual | S25 | 27 | medium | Different scope: Explicitly excludes J2K; not interchangeable with overall system availability. |
Annual system peak demand (13 values)
Note. Peaks up to 2016 were held down by load shedding, so they show what was delivered rather than what was wanted.
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2013 | 1,942.90 | MW | actual | S14 | 14 | high | Single reported value, no conflict |
| 2014 | 1,970 | MW | actual | S15 | 17 | high | Single reported value, no conflict |
| 2015 | 1,933 | MW | actual | S16 | 16 | high | Single reported value, no conflict |
| 2016 | 1,997 | MW | actual | S17 | 15 | high | Single reported value, no conflict |
| 2017 | 2,192.15 | MW | actual | S18 | 15 | high | Single reported value, no conflict |
| 2018 | 2,525 | MW | actual | S19 | 16 | high | Single reported value, no conflict |
| 2019 | 2,804 | MW | actual | S20 | 19 | high | Single reported value, no conflict |
| 2020 | 3,090 | MW | actual | S22 | 20 | high | Chart value read from the 2021 report p20 and the 2025 report p26 series; both agree |
| 2021 | 3,246 | MW | actual | S22 | 20 | high | Single reported value, no conflict |
| 2022 | 3,469 | MW | actual | S23 | 26 | high | Single reported value, no conflict |
| 2023 | 3,618 | MW | actual | S24 | 27 | high | Single reported value, no conflict |
| 2024 | 3,952 | MW | actual | S25 | 26 | high | Single reported value, no conflict |
| 2025 | 4,283 | MW | actual | S01 | 26 | high | Single reported value, no conflict |
Transformation capacity (network equipment, not generation) (6 values)
Note. This is transformer capacity on the transmission network (MVA). It is not generation capacity and says nothing about how much power plants can supply.
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2019 | 8,959.60 | MVA | actual | S20 | 5 | high | Single reported value, no conflict |
| 2021 | 9,152.80 | MVA | actual | S22 | 5 | high | Single reported value, no conflict |
| 2022 | 9,642 | MVA | actual | S23 | 6 | high | Single reported value, no conflict |
| 2023 | 9,784.80 | MVA | actual | S24 | 5 | high | Single reported value, no conflict |
| 2024 | 9,799.80 | MVA | actual | S25 | 9 | high | Single reported value, no conflict |
| 2025 | 10,493.80 | MVA | actual | S01 | 8 | high | Single reported value, no conflict |
Transmission energy losses (GWh) (11 values)
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2015 | 440 | GWh | actual | S16 | 15 | high | Single reported value, no conflict |
| 2016 | 610 | GWh | actual | S17 | 15 | high | Single reported value, no conflict |
| 2017 | 590 | GWh | actual | S18 | 14 | high | Single reported value, no conflict |
| 2018 | 707 | GWh | actual | S19 | 55 | high | Single reported value, no conflict |
| 2019 | 843 | GWh | actual | S20 | 59 | high | Single reported value, no conflict |
| 2020 | 887 | GWh | actual | S21 | 65 | high | Single reported value, no conflict |
| 2021 | 1,076 | GWh | actual | S22 | 62 | high | Single reported value, no conflict |
| 2022 | 922 | GWh | actual | S23 | 75 | high | Single reported value, no conflict |
| 2023 | 908 | GWh | actual | S24 | 75 | high | Single reported value, no conflict |
| 2024 | 951 | GWh | actual | S25 | 72 | high | Single reported value, no conflict |
| 2025 | 1,082 | GWh | actual | S01 | 84 | high | Single reported value, no conflict |
Transmission line length (6 values)
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2019 | 6,229.60 | circuit_km | actual | S20 | 5 | high | Single reported value, no conflict |
| 2021 | 6,472.20 | circuit_km | actual | S22 | 5 | high | Single reported value, no conflict |
| 2022 | 6,472.20 | circuit_km | actual | S23 | 6 | high | Single reported value, no conflict |
| 2023 | 6,639.20 | circuit_km | actual | S24 | 5 | high | Single reported value, no conflict |
| 2024 | 6,719.60 | circuit_km | actual | S25 | 9 | high | Single reported value, no conflict |
| 2025 | 6,777.71 | circuit_km | actual | S01 | 8 | high | Single reported value, no conflict |
Transmission losses (13 values)
| Year | Value | Unit | Type | Source | Page | Confidence | Note |
|---|---|---|---|---|---|---|---|
| 2013 | 4.49 | percent | actual | S14 | 14 | high | Single reported value, no conflict |
| 2014 | 4.32 | percent | actual | S15 | 17 | high | Single reported value, no conflict |
| 2015 | 3.79 | percent | actual | S16 | 15 | high | Single reported value, no conflict |
| 2016 | 4.43 | percent | actual | S17 | 15 | high | Single reported value, no conflict |
| 2017 | 4.10 | percent | actual | S18 | 14 | high | Single reported value, no conflict |
| 2018 | 4.43 | percent | actual | S19 | 16 | high | Single reported value, no conflict |
| 2019 | 4.71 | percent | actual | S20 | 19 | high | Single reported value, no conflict |
| 2020 | 4.50 | percent | actual | S21 | 19 | high | Single reported value, no conflict |
| 2021 | 5.01 | percent | actual | S23 | 25 | high | 5.01% x 21,466 GWh = 1,076 GWh, which equals the audited loss volume; 5.15% does not |
| 2022 | 4.10 | percent | actual | S23 | 25 | high | Single reported value, no conflict |
| 2023 | 3.86 | percent | actual | S24 | 27 | high | Single reported value, no conflict |
| 2024 | 3.80 | percent | actual | S25 | 26 | medium | Rounded chart label; do not infer more precision |
| 2025 | 4.02 | percent | actual | S01 | 26 | high | Single reported value, no conflict |
Method: baseline peak-demand projection
We fit a constant yearly growth rate to the 13 annual peak-demand figures from 2013 to 2025 (log-linear regression), which gives about 7.6% a year. The shaded band on the chart is an 80% prediction interval, so it widens the further ahead we look. It is wide because the history is not one steady trend: peaks were held down by load shedding to about 2016, rebounded quickly to 2019, and have grown at about 7.3% a year since. The projection does not include new demand sources such as electric-vehicle charging, extra cooling load, or supply limits.
How the method was chosen. We tested six simple methods by pretending to forecast earlier years, using only the data available at the time. The one with the lowest average miss is the one used. There are only 15 such tests and they overlap, so the ranking is a rough guide, not proof.
| Method | Tests | Average miss |
|---|---|---|
| loglinear (used) | 15 | 3.0% |
| cagr_endpoints_w7 | 15 | 3.7% |
| linear_w7 | 15 | 3.9% |
| last_yoy | 15 | 4.7% |
| loglinear_w7 | 15 | 6.9% |
| loglinear_w5 | 15 | 7.0% |
Names ending _w7 or _w5 use only the last 7 or 5 years. They did worse in these tests, partly because early test windows still contained the rebound years, but that is an explanation, not proof. The 7-year version is shown below as a sensitivity because it describes only the steadier recent years.
Projection values (forecast, baseline-0.2)
| Year | Low (p10) | Central (p50) | High (p90) | Steady-years trend only |
|---|---|---|---|---|
| 2026 | 4,203 | 4,582 | 4,994 | 4,525 |
| 2027 | 4,512 | 4,931 | 5,389 | 4,837 |
| 2028 | 4,842 | 5,307 | 5,816 | 5,172 |
| 2029 | 5,196 | 5,712 | 6,279 | 5,529 |
| 2030 | 5,573 | 6,147 | 6,780 | 5,911 |
| 2031 | 5,977 | 6,616 | 7,323 | 6,320 |
| 2032 | 6,410 | 7,121 | 7,910 | 6,757 |
Peak demand is the highest load actually delivered. If supply shortages held some demand back, the true underlying peak is higher. Capacity and reserve margin are not shown because verified plant capacity has not been loaded yet.
Data rules
- Every number records its value, unit, year, source, page, and whether it is an actual or a forecast.
- A monthly peak is never mixed with an annual peak. A forecast is never shown as an actual.
- Where sources disagree, both figures are kept side by side and flagged. We do not average them.
- Primary documents come first. News and AI summaries are only used to find primary sources.
- Gaps are left as gaps, not filled from memory.