South African gold output has fallen from 995 tonnes in 1970, two thirds of world production, to about 100 tonnes a year now, yet the mines that remain are the deepest and among the most technically demanding on earth. The industry’s future is a cost equation between three variables: depth, grade and the gold price. Right now the price is winning, and it is worth being precise about what that buys and what it cannot.
What does depth actually cost?
Every kilometre down compounds four cost lines at once. Virgin rock temperature at Mponeng’s working levels approaches 60 degrees Celsius, so refrigeration and ventilation become a permanent energy bill measured in megawatts. Rock stress at four kilometres produces seismicity that demands support, monitoring networks and conservative mining layouts, all of which cut the tonnes a crew can pull per shift. Hoisting men, material and rock up a four-kilometre shaft system caps daily throughput. And travel time eats the shift itself: an hour or more each way from bank to face means a fraction of paid hours are productive. This is why depth is not just an engineering marvel but a structural cost disadvantage that only high grades or high prices can carry.
Is deep-level mining profitable right now?
Yes, decisively, because the price term of the equation exploded. With gold trading at record levels, deep-level operators whose all-in costs once flirted with the gold price are printing margins, and the clearest proof is capital allocation: Harmony committed roughly R7.9 billion in 2024 to extend Mponeng out to about 2040 at a steady quarter-million-plus ounces a year, a mine that a decade ago was scheduled for closure this decade. The caveat matters: nothing structural changed. Grades still deplete, Eskom tariffs still climb, and the depth costs above are permanent. A sustained price retreat re-asks every closure question at once, which is why the operators are banking cash and, in Harmony’s case, spending it on copper diversification rather than new gold shafts.
Why is nobody building new gold mines?
Follow the exploration money: South African gold exploration collapsed by roughly 95 percent over two decades, to a few tens of millions of dollars a year by 2025, against around 900 million dollars in 2006. Nobody funds a new four-kilometre shaft with a 10-year build before first ore. The two live alternatives both dodge depth. Qala Shallows, brought into production by West Wits Mining in late 2025 as the first new underground gold mine in fifteen years, works old shallow ground on the West Rand where development is measured in millions, not billions. And the biggest gold “mines” around Johannesburg process no rock at all: DRDGOLD’s tailings retreatment produced about 157,653 ounces in a recent half-year from dumps holding billions of tonnes, at surface, with no ventilation bill. The read: South African gold growth is shallow, surface, or nothing. Full detail in our Gauteng mines guide.
Who still mines gold, and with what strategy?
Four names, four different answers to the depth equation. Harmony doubles down where it has the skills, extending Mponeng and Moab Khotsong while using gold cash flow to buy copper growth. Gold Fields keeps exactly one South African asset, South Deep, a mechanised three-kilometre mine that took years to tame, and builds everywhere else in the world. Sibanye-Stillwater sweats the mature Kloof and Driefontein shafts for cash and lets them shrink. AngloGold Ashanti, the company that grew out of the Chamber of Mines itself, exited South African mining entirely. When the industry’s own founders allocate nothing to new domestic gold, believe the capital, not the nostalgia.
What this means for suppliers
Deep-level gold is now a well-funded maintenance market with a decade of visibility. The Mponeng extension anchors long-cycle demand in refrigeration, ventilation, hoisting, seismic monitoring, backfill and shaft steelwork, and record margins mean deferred maintenance is being caught up. The growth adjacency is surface: tailings retreatment plants, pumping, pipelines and water treatment. Pitch depth-cost reduction, anything that cuts kilowatt-hours per tonne cooled or minutes lost per shift travels straight to the bottom line. AMIQ tracks 2,200+ African mining projects with verified owner and engineer contacts, every deep-level operation included. Join at AMIQ, and see gold mining in South Africa for the full picture.
Frequently asked questions
Which mine is the world’s deepest?
Mponeng, Harmony Gold’s operation on the West Wits line, working close to four kilometres below surface. Its R7.9 billion extension, approved in 2024, plans mining to about 2040 at around 260,000 ounces a year.
How much gold does South Africa produce now?
About 100 tonnes a year, a tenth of the 995 tonnes of 1970. The decline reflects depleted shallow reefs, the compounding costs of extreme depth, and a 95 percent collapse in exploration spending.
Will South Africa run out of gold?
Not geologically. Economically it already has in most places. What remains is very deep, low-grade, or in tailings dumps, so record prices extend existing mines while surface retreatment grows into an industry of its own.
