For most of my career, I thought I was working on the front lines of climate innovation. I studied chemistry and materials science, worked on solar fuels and energy systems, and spent years inside deep science venture building. Like many people in climate tech, I assumed the hardest problems would be about generation: how we produce cleaner energy, how we store it, how we decarbonise power systems.

It took a long time (and several startups) to realise that I was looking in the wrong place.

The real bottleneck isn’t energy. It’s materials.

If we want electrification, renewables, batteries, industrial infrastructure, data centres, or any of the technologies that underpin modern life, we need vastly more metals. Not incrementally more. Orders of magnitude more. And the way we extract those metals today simply does not scale, economically, environmentally, or politically to meet what’s coming.

That realisation was the moment Thunderstone crystallised.

The future demand problem no one can model

One of the most uncomfortable truths about mining is that no one actually knows what the demand curve will look like 10, 20, or 30 years from now.

We talk about lithium, nickel, copper, rare earths, materials that are clearly critical today, but history shows us that material demand shifts quickly. Technologies change. Chemistries evolve. Entire supply chains are disrupted by breakthroughs no one saw coming.

Inversely the back-bone of all this: mining, dependable supply depends on certainty. 

Assets are developed with 20–40 year horizons. Capital is deployed based on assumptions that must hold for decades. Infrastructure is optimised for specific ores, grades, and processing routes. When demand shifts, mining can’t respond quickly. That’s why the industry experiences violent boom-and-bust cycles, stranded assets, and geopolitical fragility.

When I was at Deep Science Ventures, one of the hardest questions we kept running into was deceptively simple: which commodity do you build for?

If you optimise for one metal and the world moves on, you’ve built an elegant dead-end.

This is why Thunderstone exists, because we decided that betting on a single answer to an unknowable future is the wrong strategy.

Designing for uncertainty, not optimisation

Instead of asking, “Which metal should we mine?”, we asked a different question:

“How do we design an extraction technology that still works even when we’re wrong about the future?”

That shift changed everything.

Thunderstone’s approach of using high-voltage underground electrical stimulation to increase permeability and extract metals without excavation isn’t about a single commodity. It’s about creating a platform that can adapt across different ore bodies, grades, and metals as demand evolves.

We designed the system to be transferable. The physics doesn’t care whether you’re extracting copper today or something else tomorrow. What matters is controlled fluid flow and metal extraction underground. This is because we’re using fluids to dissolve metals where they are without fracking and without excavating. 

That flexibility is intentional. It’s what allows mining to respond to uncertainty rather than being crushed by it.

Why access matters more than ideas

Building something this radical inside an industry that’s hungry for innovation is not something you can do from a lab alone.

One of the reasons Thunderstone has moved as fast as it has is access. Being part of the Founders Factory and Rio Tinto Mining Tech Accelerator fundamentally changed our trajectory. Not because of advice decks or theory, but because we were embedded in the real ecosystem.

We spent time in Perth, WA. We spoke directly with operators, engineers, and decision-makers. We learned what equations actually matter when a mine considers new technology. We understood where innovation gets blocked by risk, permitting, and operational constraints rather than lack of interest.

That exposure shaped our product. It forced us to prioritise deployability, not just performance. It’s easy to design a clever system; it’s much harder to design one that can survive the real world.