A short primer on the ideas behind the map and the data. No background assumed.
Lithium locked in rock — mostly the mineral spodumene. You mine the ore, crush it, and concentrate it into spodumene concentrate (a powder ~6% lithium oxide). That concentrate is shipped to a converter that turns it into battery-grade lithium chemicals.
Fast to build and ramp; concentrate is the product that gets sold and priced. Most of Australia, Canada, Brazil, Africa and China's mines on the map are hardrock (shown in teal).
Lithium dissolved in salty groundwater under salars (salt flats). You pump the brine into evaporation ponds for months to concentrate the lithium, then process it into lithium carbonate.
Slower and weather-dependent, but often very low cost once running. The big salars of Chile and Argentina — the "Lithium Triangle" — are brine (shown in blue).
SC6 = spodumene concentrate at 6% Li₂O. It's how hardrock output and the headline lithium price ("spodumene CIF China") are usually quoted.
LCE = lithium carbonate equivalent. It's the common yardstick that lets you compare hardrock and brine on one axis, because they sell different products.
Rough rule of thumb: 1 tonne of LCE ≈ 7 tonnes of SC6. That's why the map sizes every mine by its LCE-equivalent output — so a brine salar and a hardrock mine are comparable at a glance.
AISC (all-in sustaining cost) is roughly what it costs to produce one tonne of product, including the spending needed to keep the mine running. It's the single best number for "how cheap is this mine?"
Line every mine up from cheapest to most expensive and you get the cost curve. It matters because lithium is cyclical: when prices crash, the high-cost mines lose money and cut output first, while the low-cost mines keep producing. So the cost curve is really a survival ranking — who can outlast a downturn.
The map's free layer shows where mines sit; the paid Vault data gives the actual AISC figures and each mine's percentile on the global curve, every number cited to a filing.
AISC tells you whether the mine survives a downturn. The Musk Score asks a different question for a long hold: do the people running the company have skin in the game — do they win and lose alongside you?
It's named after Elon Musk, who sits at both extremes at once: an enormous ownership stake (real skin), but also heavy share-pledging — borrowing against his stock, which is "fake skin" that can unwind violently in a crash. So a single number doesn't tell the whole story; how the skin is held matters as much as how much.
The score blends three things: how management is paid (is their pay actually at risk, or guaranteed?), how much stock they truly own, and whether they're recently buying or selling — weighted toward the people at the very top. A high score isn't a "buy" signal; it's a trust signal for holding something for years.
The instructive case is a major producer with textbook pay-for-performance design — almost all of management's pay is "at risk" on paper — but almost no real ownership and net insider selling. Great incentives, little actual skin. That gap is exactly what the score is built to expose.
A project moves through stages before it produces. On the map, solid dots are producing; hollow rings are still in development.
The supply pipeline shows which developers are expected to start producing, and when.
Everything here is extracted from primary-source filings — company reports to regulators (SEC, ASX, TSX, SEDAR) and investor disclosures — then converted to US dollars and cited back to the source document. No web scraping, no guesses. Free pages show the descriptive facts; the financial detail is the paid Vault API.