The Closest I’ve Ever Been to Changing the World (10) — Who Are the Clients, and How Do They Pay?

Continuing from the last post, I asked Troy whether those AI data centers, big tech companies, and hyperscalers had given any feedback yet.

He said they’re extremely eager. “Just this week, we were invited to bid on a 5-km project from one of the world’s largest tech companies; several well-known AI hyperscalers have also visited our site many times — mainly those based in San Francisco and Seattle. They’ve been sending teams of engineers in batches. Right now, everyone’s waiting for our tunnel robot field test, which will start in three weeks.”

I said, “That’s exactly the case of ‘time is money’ — electricity supply just can’t keep up!”

He said, “Exactly. It’s pure supply shortage. As long as you can deliver power, you’ll never have trouble selling it. You don’t even need to generate new power — simply transmitting electricity that’s already produced but stranded on the congested grid today creates enormous value.”

I asked how their pricing works. He said there are two models:

(1) BOOM model — they dig and own the tunnel themselves, don’t sell it, and charge a “toll” Under this model, they can buy solar or wind power at about $40 per MWh and sell it at around $100 per MWh under long-term fixed-price contracts. “We can charge that because that’s what big tech firms are already paying the grid today anyways! We don’t have to pass on the discount and savings, and we deliver ten times faster — not to mention safer, cleaner, quicker, and more stable.”

(2) Badass service model — they charge per cubic meter of material moved, usually $50–$300 depending on the material composition, with margins of 60%–100%.

I said, “That margin is huge.”

He said, “Many solar farms’ problem isn’t that they can’t generate electricity, but that they can’t deliver it — the grid’s too congested. Sometimes transmitting power actually costs more money than they can afford, so they just dump it (called curtailment). That wasted energy costs billions each year, and we can fix that.”

I asked what other variable costs they have. He said, “Almost none. Power is the main variable cost. We need some labor, but it’s not labor-intensive: one tunnel needs two robots, each with 2-3 operators — the operators are above ground, the robots underground. The excavated rock and dust can be sold as raw materials for concrete or for road paving — that’s an extra revenue source. Spoils removal is outsourced; other costs are minimal.”

I said, “From what I hear, you’ve got the technology, engineering experience, team, and capital — and your moat sounds really deep.”

He said, “Yes, extremely deep. We hold multiple global patents — over a hundred claims across various patent clusters — basically covering everything related to plasma-based tunneling and excavation. Anyone trying to use plasma underground would infringe on our patents. We also have more than 1,200 trade secrets, each tied to unique rock-melting formulas developed over years of experiments. If someone started from scratch, they’d need 2–4 years of R&D plus 1–2 years of testing — at least 3–5 years behind us. Our key suppliers are bound by exclusive contracts — they can’t sell to potential competitors. Many components and devices were co-developed with those suppliers, requiring heavy manpower and funding. Altogether, from patents, technology, and supply chain to human capital, it would take anyone 4–6 years to replicate what we’ve built — and by then, we’ll already be far ahead 😂.”

I asked, “Given how big the market is, and how close you are to monopoly, how will you scale? Aren’t you worried about project deliveries?”

Troy said, “First, yes, we’re already oversubscribed. Right now our client database has about 20,000 verified contacts — including governments, hyperscalers, data-center operators, and major infrastructure companies. Once field-testing starts in three weeks and the robot passes that milestone, demand will explode. We’ll have to prioritize contracts and rapidly scale manufacturing. Our current facility is controlled by my family office — it’s a former WWII tank factory. At full capacity, it can produce 10 tunnel robots per month. We’re only using 10% of the space now — 90% is rented out. Once we ramp up, we can reclaim the whole building and expand production lines. With more capital, we can add even more lines.”

I asked, “What about manpower and safety measures — are those potential bottlenecks?”

He said, “Not really. We have a very detailed safety manual and program — about 30–40 pages — fully utility-compliant. Large power utilities are usually conservative and hate being the first to try new things, but in this case, once the tech is validated, they all want to be among the first customers. We’ve already done early pilots with them and built a complete compliance and safety system.”

I said, “You really have thought of everything 😂.”

He laughed, “Yeah, hopefully — I’ve been thinking through this everyday for eight years. And this isn’t my first rodeo. I know utilities inside out and have done plenty of similar projects.”

I said, “Then can you share more about your previous track record? And how others can participate?”

Till next time — the finale!