Solar PV Through 2035: Where Our Copper and Silver Sit in the Demand Curve

Last month a procurement lead from a Chinese module maker asked me a question I didn't have a clean answer for: how much silver can you commit annually by 2028? I said I'd come back to him after we finished the current sampling round on two of our polymetallic blocks. That conversation is what pushed me to write this.

Because the solar buildout numbers are getting hard to ignore. IEA's latest STEPS case has annual PV additions running above 650 GW by 2030, and BloombergNEF's transition scenario pushes past 800 GW in the same window. Whatever number you trust, the mineral intensity behind it lands on a short list of metals — and three of them matter for our portfolio in Gilgit-Baltistan.

Copper: the boring one that actually moves the needle

Copper solar demand doesn't get the headlines silver does, but per gigawatt of utility-scale PV you're looking at roughly 2.8 to 5 tonnes of copper depending on tracker configuration, inverter type and how far the array sits from the substation. Rooftop is lower. Utility scale with long DC runs is higher.

Multiply that by 700 GW a year and you get somewhere between 2 and 3.5 million tonnes of annual copper demand from solar alone by the early 2030s. That's on top of grid expansion, EVs and the general electrification story. S&P Global's 2023 copper study — the one Daniel Yergin's team put out — flagged a structural deficit opening up around 2028 that nobody in the industry has a credible plan to close.

Our position on copper sits across four concessions, with the most advanced work on a porphyry-style target in the Chalt-Chaprot belt. Grades from surface sampling and shallow trenching have come back in the 0.4 to 0.9% Cu range with molybdenum credits — I've written about the moly side separately. Honestly, we're still at the stage where I won't publish a resource number I can't defend, and we haven't drilled enough metres to defend one yet. But the mineralised footprint at surface is large, and the alteration signatures line up with what you'd expect from the Kohistan arc geology.

For a buyer or JV partner, what matters is this: the copper story in Pakistan doesn't begin and end with Reko Diq. The Karakoram arc has the same fundamental ingredients — subduction-related magmatism, the right age window, and multiple gossans that have never seen a drill rig. Reko Diq gets the attention. The north gets ignored. That's the opportunity.

Silver — where the solar PV story gets uncomfortable

Silver is the one that keeps the module engineers awake. Even after a decade of thrifting, PV still consumes roughly 15 to 20 milligrams of silver per cell, and the industry now eats up around 14% of global silver mine supply. TOPCon and heterojunction cells actually use more silver than the older PERC designs — HJT can pull 100+ mg per cell before thrifting.

The Silver Institute's 2024 numbers had solar demand at 193 million ounces. Push PV to 700 GW annually and even with aggressive thrifting you're looking at 250 to 280 million ounces from solar alone. Meanwhile mine supply has been essentially flat since 2016.

Silver in our portfolio is mostly a by-product story — associated with the lead-zinc-silver veins on two concessions near Bunji and with the polymetallic sulphides further north. Assays on the better vein samples have come back at 180 to 340 g/t Ag with lead running 4 to 9%. Not a primary silver play. But at current prices and forward silver solar PV demand, the by-product credit rewrites the economics of the base-metal work.

Here's the thing I got wrong two years ago. I used to think silver was a nice-to-have on top of a lead or copper project. Looking at the module-maker demand curves now, I've flipped that view. On the polymetallic blocks, silver is potentially the co-product that determines whether the whole thing gets built.

Tellurium — the one we probably can't help with

I'll be direct. Tellurium is a byproduct of copper refining, specifically from anode slimes at electrolytic refineries. First Solar's CdTe thin-film modules are the main consumer, and while CdTe holds maybe 5% of global module share, that share is stickier than people think because of the US IRA and domestic-content rules pushing First Solar's order book out past 2030.

We don't have a refinery. Pakistan doesn't have an electrolytic copper refinery at meaningful scale — concentrate gets exported. So tellurium recovery from our copper wouldn't happen onshore under current infrastructure. If a JV partner brought a downstream refining component into the structure, tellurium credits become part of the conversation. Otherwise it's a footnote.

What we're actually offering

Look, I'm not going to pretend our copper concessions compete with Chilean tier-one assets on grade. They don't. What they offer is:

• Undrilled or barely-drilled ground in a proven arc setting
• Polymetallic assemblages where silver, moly and gold credits carry real weight
• Concession security through the GB Mines and Minerals Department with title clean enough to satisfy Chinese and Gulf partners we've already hosted on site
• Logistics through the Karakoram Highway to Kashgar, or south to Karachi for seaborne concentrate — I've broken those costs down in an earlier post

For copper solar demand and silver solar PV exposure specifically, the interesting structure is off-take-linked project finance. A module maker or a wire-and-cable buyer takes a tonnage commitment out to 2032, and that commitment underwrites the drilling and development capex. We've had two preliminary conversations along those lines, one with a Guangdong group and one out of the Gulf.

If you're modelling your 2030 solar supply chain and you don't have Pakistan on the map yet — worth asking why not.