KoBold Shifts to Mine Ownership

KoBold stops looking like a software company once it carries a discovery into mine buildout, because the value shifts from selling picks to owning the whole copper factory. At Mingomba, that means paying for the shaft, plant, and site infrastructure up front, then keeping the upside from roughly 300,000 tonnes of targeted annual copper output and an 80% project stake if the mine reaches production.

• Wholly owned claims and earn in deals let KoBold cap spending at drilling and exploration. Mingomba is different. KoBold and ZCCM-IH broke ground on shaft construction in April 2026, moving from finding ore to financing and building the physical mine.
  1 koboldmetals 2 sacra 4 pdu
• The economics get much bigger at this stage. Public disclosures put Mingomba development cost at more than $2 billion, with KoBold targeting first output in the early 2030s. That creates exposure to mine cash flow, but also to delays, cost overruns, and commissioning risk before any metal is sold.
  1 koboldmetals 4 pdu 5 mmmd
• This makes KoBold closer to a capital intensive industrial builder like Redwood Materials than to a pure exploration software vendor like Terra AI. Both KoBold and Redwood pair technical IP with heavy physical assets, while Terra AI is positioned to monetize subsurface modeling through software fees instead of mine ownership.
  2 sacra 6 sacra 7 sacra

If Mingomba is built on schedule, KoBold will have shown that its AI can do more than improve drill targeting, it can originate tier one deposits worth developing in house. That would support a split model where the very best discoveries become owned mines, and smaller targets are farmed out through joint ventures and asset sales.