Show notes
Ryan MacLeod, who posts as The Nuclear Bitcoiner, works at Canadian Nuclear Laboratories and has spent almost eleven years in the industry as a laboratory technologist. He compares his facility to Idaho National Lab, Oak Ridge, or Savannah River in the US, and Sellafield in the UK.
The order of his conversion is the opposite of what you would guess: he did not become a nuclear advocate until after he became a Bitcoiner. Working in the industry, he only started making the public case once he saw where mining could solve problems nuclear could not solve on its own.
⭐ The centerpiece: he has written a paper he will present at a nuclear industry conference in October, titled "A Transitional Offtake Model for SMR Deployment Using Proof-of-Work Digital Asset Mining." It grew out of a study his employer conducted with Ontario Power Generation, which concluded that small reactors could eliminate diesel dependency in remote regions but would deliver the highest cost of energy of any option.
The problem the paper solves is a timing mismatch, and he states it cleanly: size a reactor for demand on the day it switches on and you constrain the community's growth; size it for demand twenty years out and most of that capacity sits idle in the meantime. Neither option is financeable.
⭐ His proposal is Bitcoin mining as a "transitional offtaker" that buys the otherwise-wasted output, then recedes. In his words, it operates as a subordinate residual load and "temporarily bridges the phase between the commissioning of the plant and the maturation of the regional demand."
⭐⭐ And the framework he builds it on is Brandon Quittem's own "Pioneer Species" essay, which Ryan notes he has rarely heard discussed. A pioneer species occupies an environment before more complex life can establish, creates the conditions that let a later ecosystem develop, and then declines in importance as higher-order entities arrive. Ryan proposes miners as a "transitional pioneer load" on exactly that model.
Answering Suze's question about how modular these reactors really are: in the one to ten megawatt range the entire unit can fit inside a tractor trailer and ship whole. Larger designs ship as components for on-site assembly. He points to Vogtle as a warning, where modular construction was attempted and some modules did not fit the larger reactor specs.
Five small reactors have recently sustained a first critical reaction, among them Oklo, Halo, Antares, and Deployable Energy. Their next milestone is generating power, then commercial operation. Site selection to operating reactor should run about five years, against ten to twenty for conventional large builds.
The use case he cares most about is northern Canada, where diesel must be shipped in year-round at high cost. A five to ten megawatt reactor carrying ten years of fuel eliminates the resupply problem entirely, and frees that diesel for uses better than burning it for heat and power.
⭐ Tomer Strolight asked the hardest question of the show: AI data centers can pay more for power than miners can, hashrate has been falling for months, so is this a problem for Bitcoin mining? Ryan's answer is that it is the pioneer species story repeating. Miners cultivated and learned to finance energy infrastructure, a wealthier customer arrived, and miners now move further out to the edge and do it again, acting as the anchor that develops energy systems to attract higher-order clients.
⭐ His most falsifiable claim, and worth revisiting later: because the best mining sites are now occupied by AI, he does not expect difficulty to follow a significant Bitcoin price rise as closely as it has historically. The power simply is not as widely available as it was during the migration out of China.
Reading and listening he recommended: Thomas Kuhn's The Structure of Scientific Revolutions, which he says every Bitcoiner should read for its account of paradigm shifts and why Keynesians cannot see what Bitcoiners see; Earth is a Nuclear Planet; and the Decoupled podcast, including its four-part breakdown of what went wrong at Vogtle.