GeoMines: What We're Actually Building for Pakistan's Mining Sector

I own 15 mines in Gilgit Baltistan. That's where this story starts.

Because every single one of those mines cost me money, time, and honestly a lot of guesswork before I ever saw a gram of return. I'd send teams up to Chilas, Skardu, Astore — sometimes based on a tip from a local, sometimes based on an old geological survey map from the 1980s. Half the time we'd dig and find nothing worth extracting. The other half we'd hit something, but not where we expected it.

So I got tired of guessing.

That's why I built GeoMines.

What GeoMines Actually Does

We pull data from four satellite sources — Sentinel-2, ASTER, Sentinel-1 SAR, and the SRTM DEM — and run it through our own AI models to figure out where minerals are likely sitting underground. Not "maybe somewhere in this district." I mean specific polygons on a map. Coordinates you can hand to a drill team.

The minerals we focus on: gold, copper, lithium, chromite, emerald, marble, granite. Basically the seven that matter most for Pakistan's economy right now.

And here's the thing most people don't realize. Pakistan is sitting on roughly $6 trillion in untapped mineral reserves. Reko Diq alone is estimated at $90 billion in copper and gold. But the vast majority of the country hasn't been properly surveyed since the British left. There are mountain ranges in Balochistan and KP that have never had a modern geochemical survey done on them. Not one.

Satellites fix that problem cheaply. A ground survey team in Chagai District can cost you $40,000 to $80,000 just to get preliminary data on a 50 sq km block. Our platform gives you a first-pass assessment on the same area for a fraction of that — often in under 72 hours.

Why Satellites, and Why Now

I used to think satellite data was too coarse to be useful for real exploration. That was my honest view three years ago. I figured it was fine for academic papers but not for someone actually about to spend $200K on a drilling program.

I was wrong about that.

What changed my mind was ASTER's shortwave infrared bands. ASTER can pick up specific mineral signatures — clay alteration, iron oxides, carbonates — because different minerals absorb and reflect light at different wavelengths. Combine that with Sentinel-2's 10-meter resolution optical imagery, and then layer SAR data (which sees through cloud cover and can detect surface deformation), and you start getting a picture that a single ground team would need months to build.

Then the AI does the heavy lifting. Our models have been trained on known deposits across Pakistan — Saindak, Reko Diq, the chromite belts in Muslim Bagh, the emerald zones in Swat. So when we scan a new area, the system compares the spectral signatures against those known productive zones and flags anomalies with a confidence score.

It's not magic. It won't tell you exactly how many tonnes are underground. But it'll tell you where to drill first, which for most mine owners is the single most expensive decision they make.

Who's Actually Using This

Four groups so far.

Mining companies use us for early-stage target generation before they commit to a full exploration budget. Investors use our reports to do due diligence on properties they're being pitched — because a lot of "mineral rich" claims in Pakistan turn out to be nonsense once you look at the actual spectral data. Provincial mining departments use us for regional assessments (KP's mineral department has been particularly forward-thinking on this). And individual mine owners, people like me, use it to figure out whether their existing lease has more potential than they've tapped.

That last group is the one I care about most. Because there are thousands of small mine owners across Gilgit Baltistan, Balochistan, and KP who inherited their leases or bought them cheap, and they have no idea whether they're sitting on $50,000 or $50 million. A proper geomine analysis report costs less than a single week of diesel for their excavator.

One of my own mines near Skardu — I'd been working the marble deposit for years. Ran our own breeze geo mineral analysis on the surrounding 12 sq km last year. Turns out there's a probable copper-gold indication about 2.3 km northwest of the marble quarry. I hadn't touched that ridge in a decade. Currently getting ground samples pulled for verification.

That's the kind of thing I want every mine owner in this country to have access to.

What's Next

We're expanding coverage. Right now geomines.org satellite mineral exploration covers the entire country, but we're building out province-specific models — Balochistan's geology is completely different from GB's, and the AI performs better when it's tuned for local rock formations.

We're also adding hyperspectral data (from newer satellites like EnMAP) which will push accuracy further. And we're working with two universities in Islamabad on training data — because the more verified ground-truth samples we feed the models, the sharper the predictions get.

Look, I'm not going to pretend geo mining is easy or that a satellite report replaces boots on the ground. It doesn't. You still need geologists, still need drilling, still need chemistry labs. What we do is make sure you're spending that money in the right place instead of the wrong one.

If you own a mine, or you're thinking about buying one, or you're a company scouting Pakistan for the first time — reach out. Even if you don't end up using geo mine reports from us, I'll tell you honestly what I think about your area.

Because we've got too much mineral wealth in this country to keep guessing at it. Don't you think?