How Much Bitcoin Has Bhutan Mined? Do the Math

In 2026 the story wrote itself: Bhutan, the little kingdom that mined Bitcoin with its rivers, was dumping it. On-chain trackers watched the attributed wallets shrink, and the headlines followed — sold 70%, offloaded $200 million, the quiet sovereign seller. Bhutan does not publish its holdings, so inference rushed in to fill the vacuum.

But there is a cleaner way to ask how much Bitcoin Bhutan has than squinting at wallet clusters. You do not need its books. You need its power — which is public — and arithmetic, which is free. Mining is not alchemy; it is physics with an electricity bill. Give a calculator the megawatts, the network’s difficulty, and the going rate of a modern machine, and it will tell you, within a believable band, how much Bitcoin those rivers have made.

So let us do the math, out loud, with numbers anyone can check.

A note on timing: the arithmetic below is a snapshot as of June 2026. Bitcoin’s network hashrate, its block reward, and the efficiency of the going hardware all shift over time — so the figures are anchored to that moment, and re-running them with a later month’s numbers will move the answer.

The only input that really matters: power

Everything starts with capacity. Bhutan’s sovereign mining runs on its hydropower surplus, and the public anchor is the joint venture between Druk Holding & Investments and the mining firm BitDeer, announced in 2023 and reported to scale toward roughly 600 megawatts. Call it 600 MW of mining load — the figure BitDeer itself has put on the public record. For a country of under 800,000 people, it is a staggering amount of computing power to point at one problem. It is also, conveniently, all we need.

From megawatts to coins: the yearly rate

A mining machine has one job: turn electricity into hashes — guesses at the next block. The best 2026 hardware does it at about 11 joules per terahash (the latest hydro-cooled units run 10–12.5 J/TH). Push 600 MW through machines that efficient and you get roughly 55 exahashes per second of computing power.

The entire Bitcoin network, in mid-2026, runs at about 950 EH/s (it has bounced between 918 and 974 this year). So Bhutan’s 600 MW is about 5.7% of all the Bitcoin mining on Earth — exactly the kind of share you would expect from the first sovereign miner built at industrial scale.

The network mints 450 fresh bitcoin a day (3.125 per block × 144 blocks). Five-point-seven percent of that is about 26 BTC a day, or ~9,400 a year at full throttle.

Nothing runs at full throttle. Machines underperform, hydropower is seasonal, capacity ramps. So halve it — a deliberately steep 50% haircut that absorbs every real-world friction at once. That still leaves roughly 4,700 BTC a year from 600 MW at today’s difficulty.

Walking the ramp: the all-time total

Bhutan didn’t switch on 600 MW yesterday — and here is the part the headlines miss: earlier years mined richer. Before the April-2024 halving, every block paid 6.25 BTC, not 3.125. And when the global network was smaller, the same megawatts captured a bigger slice of it. A modest rig in 2022 out-earned a big one in 2026.

So walk the curve, keeping the same 50% haircut throughout:

Roughly sixteen thousand bitcoin, mined across the program’s life — and that is the conservative reading. (The capacity ramp is the one genuinely soft input; nudge it, or soften the haircut, and the total climbs past 18,000.) The point is the order of magnitude, not the third digit.

Here is the tell that the model is honest. It lands right on top of the ~12,000 BTC peak that on-chain analysts independently reported in 2024 — and then keeps climbing, as two more years of mining should. Built from nothing but public power, public difficulty, and public hardware specs, the arithmetic reproduces the number the trackers found by a completely different route. When two unrelated methods agree, you are usually near the truth.

Where the “sell-off” actually went

Now lay the public numbers side by side:

• Bhutan has mined on the order of 16,000 BTC.
• In its December 2025 National Day pledge, it publicly committed up to 10,000 BTC to the long-term build of Gelephu Mindfulness City — moved into a ring-fenced reserve the trackers no longer count as “active.”
• 16,000 − 10,000 ≈ 6,000 left in the wallets analysts still watch — which is precisely the “mid-thousands” they report.

Read it the other way and the famous sell-off evaporates. The tracked stack “fell” from a ~12,000 peak to a few thousand — a drop of roughly ten thousand coins. The pledge to Gelephu was for up to ten thousand coins. The amount that “disappeared” from the trackers is, almost to the thousand, the amount Bhutan publicly earmarked on national television. That is not a liquidation. It is a transfer — and a disputed one at that, with Druk Holding & Investments stating it does not recall selling.

Read this as a back-of-envelope — because it is one

Holdings are undisclosed; the capacity ramp is estimated; the haircut is a judgment call. Move the efficiency to 12.5 J/TH or the network to 974 EH/s and the yearly figure shifts a few hundred coins. Assume a gentler haircut and the cumulative pushes well past 18,000. None of that changes the shape of the answer.

And the shape is what matters: a country mining ~4,700 BTC a year cannot also be a country that has quietly run its reserves down to nothing. To believe the sell-off story, you have to believe Bhutan sells essentially everything it earns, every year, while its head of state announces the opposite.

So — how much Bitcoin has Bhutan mined? The arithmetic says roughly sixteen thousand, give or take a few thousand: a fortune for a country its size, made from water that would otherwise have been sold south at a discount. How much does it still hold? Enough that the only way to believe the sell-off is to ignore the power bill.

The trackers see a shrinking pile and infer a seller. The calculator sees a miner that earns more in a year than the trackers think it owns — and a public pledge that accounts for the gap almost exactly. One of those readings requires Bhutan to be lying on national television. The other just requires a calculator.

Do the math yourself. Every input is public — that is the whole point.