Issue 032 - Battery materials - Material-flow scaling
How many EV batteries could one Chinese lithium mine support?
CATL's Jianxiawo lithium mine remains closed while awaiting environmental approval. Reuters reports that, when operating normally, the mine has annual production capacity equivalent to about 46,000 metric tons of lithium carbonate, roughly 3% of 2025 global output.
The problem
Estimate how many electric-car battery packs could ultimately be supported by one year of lithium production from the Jianxiawo mine.
Then use your result to estimate the number of days or weeks of Chinese EV production that this represents, and how many additional mines of roughly this scale would be needed to supply lithium for a hypothetical fleet of 100 million new EVs.
Is a mine supplying a few percent of global lithium production large enough that its closure should materially affect the EV industry, or is the supply chain diversified enough that other producers can readily absorb the loss?
Because Fermi problems target an order of magnitude, I normally use no more than two significant digits and write most calculations in scientific notation; the Fermi reference explains both conventions.
Before checking sources
Matt's first pass
I assumed the chemical formula for lithium carbonate was LiCO2. Since the atomic mass of carbon is 12 and oxygen is 16, I estimated that lithium accounts for about 12% of the mass of lithium carbonate mined.
mine output ~= 4.6 x 10^4 metric tons lithium carbonate/year
assumed lithium share ~= 12%
lithium mass ~= 4.6 x 10^4 x 0.12
~= 5.5 x 10^3 metric tons
~= 5.5 x 10^6 kg lithiumFor the batteries, I assumed we are interested in the lithium ion rather than the whole lithium carbonate molecule. I assumed very high purification yields and guessed that only a small portion of the battery mass is lithium ion. I approximated that as 1 kg per battery, though I thought that might still be high.
battery packs ~= 5.5 x 10^6 kg Li / 1 kg Li/pack
~= 5.5 x 10^6 EV batteriesAt that rate, if annual production by Chinese EV manufacturers is about 50 million vehicles, then that is about 1 million per week. So this mine's annual yield would support about 5.5 weeks worth of EV batteries.
For a hypothetical 100 million EVs, I estimated at least 18 to 20 mines of this scale would be required.
Calibration Score
Matt's Calibration Score: 40 / 100
Higher is better: earn points for accurate pegs, sound models, correct math, and a result close to the sourced answer. The image shows percent full of it: 100 minus the Calibration Score.
Pegs: 0/30. Lithium carbonate chemistry, battery lithium intensity, and China production pegs were all meaningfully off.
Model: 15/30. The material-flow model was basically right, but converting through pure lithium added avoidable confusion.
Math: 5/10. The chemistry mistake affected the setup, though the main arithmetic was still usable.
Result: 20/30. The final pack and mine-count answers were within an order of magnitude because errors canceled.
Grounding facts
The headline "3% of global output" sounds modest, but in battery terms it is close to a million full-size EV battery packs per year. That is several weeks of China-scale NEV production, and a meaningful shock to a market that can already be sensitive to temporary disruptions.
The answer also shows why material substitution, recycling, pack-size discipline, sodium-ion batteries, and diversified lithium supply matter. A 100-million-EV fleet is not impossible from a lithium standpoint, but it requires many mine-years, processing capacity, and stable supply chains.
After checking sources
Check and recalibrate
The first correction is chemical. Lithium carbonate is Li2CO3, not LiCO2. Its molar mass is about:
Li2CO3 mass ~= 2 x 7 + 12 + 3 x 16
~= 14 + 12 + 48
~= 74 atomic mass units
lithium share ~= 14 / 74
~= 0.19But the cleanest battery calculation uses lithium carbonate equivalent directly. IRENA cites a rough estimate of about 160 g of lithium metal per kWh, equal to about 850 g of LCE per kWh. IEA reports that in 2025, average battery electric car pack size in China was below 60 kWh.
average BEV pack ~= 60 kWh
LCE per kWh ~= 0.85 kg/kWh
LCE per pack ~= 60 x 0.85
~= 51 kg LCE/packNow divide the mine's annual LCE output by the battery-pack LCE demand:
mine output ~= 4.6 x 10^4 metric tons LCE
~= 4.6 x 10^7 kg LCE
packs supported ~= 4.6 x 10^7 kg / 51 kg/pack
~= 9 x 10^5 packsSo one year of Jianxiawo output supports on the order of 900,000 average BEV battery packs. A smaller average pack, plug-in-hybrid-heavy mix, or lower lithium intensity could push this above 1 million. Larger packs or buffer losses could push it lower.
China's new-energy vehicle production exceeded 13 million units in the first ten months of 2025 and was projected around the mid-teens of millions for the year. Use about 16 million NEVs/year as the current memory peg, while remembering that plug-in hybrids have smaller packs than BEVs.
China NEV production ~= 1.6 x 10^7 vehicles/year
weekly production ~= 1.6 x 10^7 / 52
~= 3.1 x 10^5 vehicles/week
weeks supported ~= 9 x 10^5 / 3.1 x 10^5
~= 3 weeksFor 100 million new EVs with roughly this battery size:
mines ~= 1 x 10^8 EVs / 9 x 10^5 EVs per mine-year
~= 1.1 x 10^2 mine-yearsThat is roughly 100 mine-years at Jianxiawo scale. If the question is annual supply for 100 million EVs per year, it means about 100 mines operating at this scale. If the fleet is built over a decade, divide the annual mine requirement by about 10.
A mine that represents 3% of global lithium carbonate output is absolutely large enough to move prices and planning, especially if inventories are thin. But it does not mean the entire EV industry stops. The practical effect depends on stockpiles, substitute suppliers, lithium chemistry, refining capacity, long-term contracts, and how quickly other mines or processors can fill the gap.
Post-check reflection
Matt's reflection
As is often the case, errors in estimates canceled out and I ended up with answers that were somewhat close to the correct answer, Fermi-correct within an order of magnitude.
I did not remember the chemical formula for lithium carbonate correctly. I thought it was LiCO2, and the actual formula is Li2CO3, so the percent mass of lithium I used was off. Then my guess about how much lithium is found in each battery was wrong by about an order of magnitude: it is closer to 10 kg of lithium metal, or roughly 40 to 50 kg of lithium carbonate equivalent, than it is to 1 kg. Also on this point, lithium carbonate equivalent is already the material-accounting unit we want, so I did not need to purify the calculation down to lithium metal first.
Next, my guesstimate for how many EV vehicles are manufactured annually in China was about 3x too high. I should have anticipated this; the problem indicates this one lithium mine was producing about 3% of global annual output, so China alone would have been using a huge amount of the total lithium output.
Ultimately, I calculated that about 5.5 million batteries could be supported annually by this one mine. In reality, it should have been closer to a million, which is still correct in Fermi terms. I guessed the annual output of the mine would support about 5.5 weeks worth of EVs in China, and a more accurate guess is closer to 3 weeks. I guessed 18 to 20 mines of that scale to support 100 million EVs, while the actual count would be about 90 to 110, still within an order of magnitude.
It was sloppy, and my assumptions were a bit off because my mental models were not quite right, but I got close enough for the purposes of this exercise. I think my intuitions on the topic should be tightened up, but in terms of the scale of the issue, they are decent. I also think this news item is an important one: lithium is an incredibly important mineral resource for modern technologies, and a mine that produces 3% of global output is tremendously important to lots of manufacturing pipelines.
Recommended memory peg
For EV lithium estimates, remember lithium carbonate is Li2CO3, 1 kg Li ~= 5.3 kg LCE, EV batteries need about 0.85 kg LCE/kWh, and a 60 kWh pack needs about 50 kg LCE.
Reader results
Bars show how submitted estimates sort into the answer choices from the gut-check prompt.