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

Last winter I stood on a gravel bar near Chilas watching four men work a sluice box with a diesel pump. In six hours they pulled somewhere around 11 grams of fine flour gold. No drilling, no permits above the district level, no geology degree between them. That's the informal reality of Indus river gold, and it's also why the formal opportunity here is so badly understood by outside investors.

Let me give you the numbers as we actually see them on our concessions, and where the economics stop making sense.

The geology, briefly

The gold in the Indus, Gilgit and Hunza river systems is overwhelmingly fine — what placer miners call flour gold or flood gold. Particle sizes typically run 50 to 300 microns, occasionally coarser near tributary mouths where the Gilgit meets the Indus at Jaglot, or where the Hunza dumps into the Gilgit near Danyore. You won't find many nuggets. What you will find is huge tonnages of auriferous gravel, replenished every monsoon season as the rivers tear down freshly weathered material from the Kohistan island arc, the Karakoram batholith, and the various shear zones cutting through them.

The source rocks are the interesting part. We're downstream of orogenic gold systems that haven't been properly explored in the hard rock. Every gram in the river came from somewhere upstream. That's a story for another post.

On the placer side, grades across our sampled bars sit in this range:

• Indus mainstem between Sazin and Chilas: 120 to 340 mg per cubic metre of processed gravel
• Gilgit river near Bunji confluence: 180 to 420 mg/m³, with occasional high bars pushing 600
• Hunza river below Chalt: more variable, 90 to 280 mg/m³, but coarser particles
• Terrace deposits (older, elevated benches) above the active channel: 200 to 550 mg/m³ but far more overburden

These are our numbers from panned concentrate and small-scale sluice tests, cross-checked against work the Geological Survey of Pakistan did in the 1990s and some more recent university sampling. I'd treat anything above 500 mg/m³ as a lucky bar, not a resource.

Honestly, I got the terrace economics wrong when we started. I assumed higher grade meant better project. It doesn't — the stripping ratio kills you unless the bench is thin and the gold is concentrated at the basal contact.

Dredging economics, without the hype

Here's where most pitch decks I've seen from Pakistan fall apart. People quote a grade, multiply by cubic metres, apply the LBMA gold price, and produce a fantasy revenue number. That's not how placer works.

Run the actual math on a mid-sized cutter suction or bucket-ladder dredge operating 20 hours a day in the Indus during the low-flow season (November through March, when you can actually work the channel safely):

• Throughput: 180 to 250 m³/hour of gravel through the sluice, realistically
• Recovery on fine gold with a properly set-up sluice plus centrifugal concentrator (Knelson or Falcon type): 62% to 78%. Anyone quoting 90%+ on 100-micron gold is selling you something
• Effective operating days per year in GB: 110 to 140. The monsoon shuts you down, and winter access above 2,500 m gets ugly
• Diesel burn, labour, security, camp, spares, royalty to the GB government: our modelling comes out at USD 14 to 19 per cubic metre processed, all-in

So on a 250 mg/m³ bar at 70% recovery, you're recovering 175 mg per cubic metre. At current gold prices around USD 84 per gram wholesale (net of refining), that's roughly USD 14.7 of gold per m³. Against USD 16 in costs.

You see the problem. The average bar doesn't pay. The good bars pay handsomely.

This is why placer mining here has to be selective. Bulk sampling first, then targeted dredging of proven high-grade sections, with the flexibility to move equipment. Anyone showing up with one big dredge parked in one spot is going to lose money. We've watched it happen twice — one Chinese contractor near Thakot in 2019, one domestic operator near Bunji more recently.

Licensing, royalty and how the gold actually leaves

Mineral rights in Gilgit-Baltistan fall under the GB Mines and Minerals Department, not the federal government. This trips up foreign investors constantly. The regime is separate from Punjab, Sindh, Balochistan and KP. Placer gold requires a mining lease specifically endorsed for alluvial workings, plus environmental clearance from GB-EPA, plus a district-level NOC. Timelines run 4 to 9 months if your paperwork is clean.

Royalty on gold is currently 3% of wholesale value, paid to the GB government. There's also a district cess and a small welfare contribution. Nothing punitive by international standards.

Export goes through State Bank of Pakistan channels. Refined dore or high-purity concentrate can be shipped legally with the right documentation — we've done it, it works, but it requires a proper compliance setup. The informal cross-border flows you may have heard about are a separate universe and not something a serious JV partner should touch.

One thing worth flagging for anyone modelling this: the Karakoram Highway is your logistics spine. From our concessions to Islamabad dry port is 14 to 20 hours by road depending on weather and the Attabad tunnel situation. From Islamabad to Karachi port or onward air freight is straightforward. The bottleneck is always the KKH, not the paperwork.

What we're actually looking for

We're not selling shovels. We hold the ground, we've done the sampling, and we're open to JV structures where a partner brings dredging capital and technical operating experience — particularly on fine-gold recovery, where the Chinese and Australian placer operators are frankly ahead of anyone else. Off-take on dore is also on the table for refiners who want traceable, single-origin Pakistani gold with proper chain-of-custody documentation.

If you're modelling this seriously, ask for our bar-by-bar sampling data before you build any spreadsheet. Averages will mislead you here. The whole game is knowing which 200 metres of river to work, and which 20 kilometres to leave alone.