Alluvial and Placer Gold in the Indus, Gilgit and Hunza Rivers: What the Numbers Actually Look Like

Last March I stood on a gravel bar near Jaglot, just below where the Gilgit meets the Indus, watching three men work a sluice that hadn't changed in design since the 1960s. They pulled roughly 4.2 grams that day. From maybe nine tonnes of gravel. That's the reality of placer gold Pakistan — it's there, it's recoverable, and almost nobody is doing it at scale.

That's what this post is about. Not the romance of it. The geology, the grades we're actually seeing, and what the economics look like if you bring proper equipment in.

Where the gold comes from, and why these rivers carry it

The source rocks matter more than most buyers realise. The Indus drainage above Skardu cuts through the Kohistan-Ladakh island arc and the Karakoram batholith. The Gilgit and Hunza tributaries drain hydrothermally altered zones along the Main Karakoram Thrust and the Shyok suture. We've got lode gold in quartz veins through the Bagrote, Haramosh and parts of the Hunza valley feeding directly into these systems.

So the gold isn't mysterious. It's eroding out of veins we can actually map. The rivers are doing 40 million years of free comminution and gravity concentration for us.

What that produces is classic placer stratigraphy. Fine to medium flour gold dominates — typically 80 to 150 microns — with occasional flakes up to 2mm and the odd nugget. Most of what we recover from the Indus bars near Chilas and Jaglot is fine. The Hunza system, because it's a shorter transport distance from source, gives us coarser material. And honestly, that changes everything about recovery design.

Fineness runs high. We've assayed bars at 870 to 920 fine on average, with some Hunza concentrate coming back at 940. That's better than a lot of West African placer product and it matters when you're negotiating refining terms.

What the grades actually look like

I'll give you real numbers because that's what serious buyers want.

From our sampling programme across three concessions on the Indus between Raikot and Sazin, surface gravel runs 0.15 to 0.42 g/m³. The paystreak — and there's almost always a paystreak sitting on or just above bedrock or a false bedrock clay layer — runs 1.8 to 4.6 g/m³. We've had isolated bedrock samples come back at 11 g/m³ but I won't pretend those are representative.

Gilgit River bars are leaner on average. Call it 0.08 to 0.25 g/m³ in surface gravel, paystreak around 1.2 to 2.9 g/m³. But the overburden is thinner and the gravels are looser, so the strip economics are better.

Hunza is the interesting one. Smaller volume, but grades on the better bars near Aliabad and below the Hispar confluence have come back at 0.6 to 1.4 g/m³ in surface material. That's recoverable with a 100 cubic metre per hour wash plant without much fuss.

I got the volumetrics wrong on our first concession, by the way. I assumed paystreak continuity that wasn't there. Took us a full season of test pitting on a 1.2 km bar to figure out the paystreak braided and pinched in three places. Lesson learned — auger drilling on a 25 metre grid before you commit to plant siting.

The economics, honestly

Here's the part most overseas buyers want to understand. Can you actually make money pulling alluvial gold Pakistan out at scale?

Short answer — yes, but the model is different from artisanal.

A properly specced 150 m³/hr trommel-and-sluice plant with centrifugal concentrators (think Knelson or Falcon style) lands in Karachi for roughly 380,000 to 450,000 USD depending on configuration. Trucking to Gilgit-Baltistan adds another 18,000 to 24,000 USD and three weeks if the Karakoram Highway behaves. It usually doesn't in winter.

Diesel is the killer. We're running at roughly 1.15 USD per litre at site, and a plant that size burns 40 to 55 litres per hour. Labour is cheap — skilled plant operators at 600 to 900 USD per month, general labour at 180 to 250. Royalty to the GB government is 2% of gross plus a fixed concession fee. Federal export clearance through SBP is straightforward once you've done it once.

Working the numbers on a paystreak at 2.5 g/m³ with 88% recovery (realistic with centrifugal kit on fine gold), at current gold around 2,650 USD per ounce, you're looking at roughly 187 USD recovered value per cubic metre of paystreak. All-in cash cost per cubic metre processed — including stripping a 3:1 overburden ratio — sits around 14 to 19 USD depending on fuel and how far you're trucking water.

The margin is real. What kills projects here isn't the geology or the economics. It's seasonality (you get maybe seven workable months on the Indus before water levels make the bars inaccessible), logistics, and trying to run the operation without a serious local partner who understands the licensing and community side.

A note on what we're not doing

We don't use mercury. None of our concessions, none of our partners. Gravity concentration only, with intensive cyanidation on concentrate done at a licensed facility downcountry. I mention it because every credible buyer in the EU and Japan asks, and it should be the first question.

Dredging is another conversation. We've looked at small cutter-suction dredges for the deeper Indus channels below Jaglot where the historical paystreaks are submerged. The capex jumps to 1.4 million USD plus and the permitting is harder. Workable, but only at JV scale.

If you're a refiner, trader or mining group looking at Gilgit-Baltistan gold mining seriously — what stage are you at, and what tonnage of doré per quarter would actually move the needle for your book?