Analysis of Powerwall Battery Retention - Part 3
Introduction
This is Part 3 of the Powerwall battery retention analysis. For an introduction, methodology, and earlier datasets, see Part 1 and Part 2.
This update includes:
- Another year of data. The oldest Powerwall 2 units in the dataset, built in late 2017, are now 9 years old, one year short of the 10-year warranty period. As before, data come from anonymized statistics collected through Netzero, an app for managing home energy systems that also provides access to Powerwall diagnostic data.
- Trend lines for both Powerwall 2 and Powerwall 3.
- The first utilization analysis for Powerwall 3.
- A direct comparison of Powerwall 2 and Powerwall 3 degradation at matching ages.
Methodology Changes
The plots follow the same conventions as Parts 1 and 2: manufacturing date on the X axis (reversed, newest units on the left), nominal full-pack energy on the Y axis, and a small amount of random noise to prevent identification of individual batteries.
New in this part, each part-number plot carries a trend line showing the median capacity by manufacturing date.
The Powerwall 3 data now include Expansion packs (part number 1807000), which were not part of the Part 2 dataset.
Powerwall 2 Update
The updated scatter plot, color-coded by part number, now includes a trend line. The trend line excludes the 2018-2020 outlier cluster (part number 2012170, orange points at the bottom, discussed in Part 1), since its cause is not related to normal degradation, as well as the post-2024 replacement units discussed in Part 2. It therefore describes regular production only. Powerwall 2 production ended in late 2024.
The trend line makes the shape from Part 1 explicit:
- The newest units, now almost 2 years old, have a median capacity of 14.1 kWh.
- Capacity falls to about 13.6 kWh by 3-4 years and to 13.1 kWh by 5 years.
- From about 5.5 years on, the trend is flat. Units from 2017 through 2021 all sit at a median of 12.7-13.1 kWh, regardless of age.
- Excluding the 2018-2020 outlier cluster, only about 3% of units older than 5 years are below 12 kWh, and 2 of about 1,500 are below the 9.45 kWh warranty threshold.
Median capacity by age, excluding the outlier cluster and replacement units:
| Age | Median capacity | Below 13.5 kWh | Below 12 kWh |
|---|---|---|---|
| 1-2 years | 14.05 kWh | 6% | 0% |
| 2-4 years | 13.71 kWh | 38% | 1% |
| 4-6 years | 13.25 kWh | 64% | 2% |
| 6-8 years | 12.79 kWh | 90% | 3% |
| 8-10 years | 12.82 kWh | 83% | 5% |
Battery Retention by Discharged Energy
The color scale runs from 0 MWh/year (purple) to 7 MWh/year (yellow) of average discharged energy. 5 MWh/year corresponds to a full 13.5 kWh discharge every day.
The utilization gradient is clear. Among units older than 3 years (excluding the outlier cluster), the median capacity by average yearly discharge is:
| Average discharge | Median capacity |
|---|---|
| Under 1 MWh/year | 14.04 kWh |
| 1-2 MWh/year | 13.50 kWh |
| 2-3 MWh/year | 13.12 kWh |
| 3-4 MWh/year | 12.86 kWh |
| Over 4 MWh/year | 12.79 kWh |
Units that barely cycle, typically those with the backup reserve set to 100%, retain almost their full initial capacity after more than 3 years.
Powerwall 3 Update
The Powerwall 3 dataset has grown substantially, and the oldest units are now over 3 years old. The Y axis is narrower than in the Powerwall 2 plots to spread out the points, so the same capacity loss looks steeper here.
Observations:
- New units report 14.4 kWh, about 0.3 kWh more than new Powerwall 2 units did, and with much less spread.
- The spread widens with age. At 2-3 years, 16% of units are below the 13.5 kWh rating, compared with under 1% in the first year.
| Age | Median capacity | Below 13.5 kWh |
|---|---|---|
| 0-1 years | 14.38 kWh | 0% |
| 1-2 years | 14.29 kWh | 5% |
| 2-3 years | 14.03 kWh | 16% |
Battery Retention by Discharged Energy
This is the first utilization plot for Powerwall 3. The color scale runs from 0 to 8 MWh/year.
The same utilization gradient appears, on a smaller scale. Among units older than 1.5 years:
| Average discharge | Median capacity |
|---|---|
| Under 1 MWh/year | 14.38 kWh |
| 1-2 MWh/year | 14.29 kWh |
| 2-3 MWh/year | 14.12 kWh |
| 3-4 MWh/year | 13.99 kWh |
| Over 4 MWh/year | 13.95 kWh |
Lightly used units have lost essentially nothing, and heavily cycled units about 0.45 kWh. Powerwall 3 units are also cycled harder: the median unit in this age range discharges 2.7 MWh/year, compared with 2 MWh/year for Powerwall 2 units older than 3 years.
Powerwall 2 vs Powerwall 3
With three years of Powerwall 3 data, the two generations can be compared at matching ages. Powerwall 3 uses lithium iron phosphate (LFP) cells, which are expected to tolerate more cycles than the nickel-manganese-cobalt (NMC) cells in Powerwall 2.
| Age | Powerwall 2 median | Powerwall 3 median | Powerwall 2 below 13.5 kWh | Powerwall 3 below 13.5 kWh |
|---|---|---|---|---|
| 1.5-2 years | 14.05 kWh | 14.16 kWh | 5% | 9% |
| 2-2.5 years | 14.07 kWh | 14.03 kWh | 13% | 14% |
| 2.5-3 years | 13.61 kWh | 13.75 kWh | 41% | 29% |
Observations:
- Relative loss over the first three years is similar. Powerwall 3 drops from 14.4 kWh new to 13.75 kWh at 2.5-3 years, a loss of about 4.5%. Powerwall 2 units of the same age are at 13.61 kWh, about 3.5% below the 14.1 kWh that new Powerwall 2 units reported in Part 1. The Powerwall 3 units at that age are a small cohort of early hardware revisions, so the difference may not hold for later production.
- Powerwall 3 is more uniform. At 2.5-3 years, 29% of Powerwall 3 units are below the rated capacity, compared with 41% of Powerwall 2 units of the same age. The advantage so far is fewer early fallers, not a slower median decline.
- Both generations lose capacity in proportion to cycling. Comparing the utilization tables above, the capacity difference between lightly and heavily cycled units is about 1.25 kWh for Powerwall 2 units older than 3 years and 0.45 kWh for Powerwall 3 units older than 1.5 years. Per unit of energy discharged, the two are closer than these numbers suggest, since the Powerwall 3 units are younger and have cycled less.
- The long-term question remains open. Powerwall 2 loss is front-loaded: most of it happens in the first 4-5 years, after which capacity stabilizes around 12.8 kWh, 95% of the rated capacity. Whether Powerwall 3 flattens the same way, and at what level, will only become clear as the 2023-2024 units age further.
Conclusions
Nine years after the first Powerwall 2 units were built, the picture is stable: units lose about 1.2 kWh over their first five years, then hold steady near 12.8 kWh. Capacity loss tracks utilization closely, and lightly cycled units retain almost all of their initial capacity. Outside the 2018-2020 outlier cluster, units below the warranty threshold remain rare.
Powerwall 3 starts higher and more consistent, holds its capacity for the first year, and then begins to decline at a rate that, so far, is comparable to Powerwall 2 at the same age. The oldest Powerwall 3 units are still young by warranty standards, and we will continue to track them.
Contact
For questions or comments contact feedback@netzero.energy.