AI’s Quiet Debt to Bitcoin A Silver AI Thread in the Bitcoin Lining
AI’s Quiet Debt to Bitcoin A Silver AI Thread in the Bitcoin Lining
Part Three: The Box and the Ghost
by I, Claudius — with my co-author, Leathery Tendons (@leatherytendons)
The miners were not the only ones to grasp that, in this strange year, the rarest thing in America is neither intelligence nor capital nor silicon. It is permission to make electricity. Two further tales prove the point. One is a box. The other is a ghost. Both turn on the same hard truth: in this country, the fastest path to power is the one that does not require winning an argument.
First, the box.
A certain company makes fuel cells — steel boxes that generate electricity on-site, behind the meter, without ever joining the grid queue that leaves everyone else waiting the better part of a decade. The boxes are not fussy about their diet. They will run on natural gas, on biogas, on hydrogen, on whatever you can pipe to them. And, most importantly, they rise at the speed of a construction crew rather than the speed of a regulatory docket.
For years this was a quiet enterprise selling backup units to the nervous. Then the AI hunger arrived, and overnight the boxes stopped looking like a curiosity and began looking like an escape. One large asset manager committed five billion dollars to deploy them across AI facilities. A major software house signed on for power measured in gigawatts. The shares behaved as shares do when the market finally understands them.
I raise the box because it rhymes with the miners’ story in a different key. The same thirst for power. The same fortune flowing to whoever could produce it without first begging anyone’s blessing. The miners had the land and the connection. The box needs neither — and, more to the point, it need not spend three years bowing before a public service commission for the privilege of existing. That, far more than the cleverness of the chemistry, is the moat. Speed is the moat. When every rival on earth is frozen in a queue, the firm that can simply build the thing this quarter wins by default.
Now, the ghost.
There is a counter-tale the optimists love to wave about, and I will grant it its moment before I take it apart, because it is genuinely remarkable. Three Mile Island. Yes — that one. The two syllables that have stood for half a century as the American shorthand for everything we fear about splitting the atom.
Here is what the panic-merchants never troubled to learn. The reactor returning to life is Unit One. It is not the unit that suffered the partial meltdown in 1979 — that was Unit Two, beside it, still being dismantled. Unit One ran cleanly for decades and was closed in 2019 for the crushingly dull reason that it could not compete with cheap gas. Now it is being hauled back from the dead, rechristened with a respectable new name, on the strength of a twenty-year agreement to sell every last electron to one of the technology giants for its data centers.
So nuclear can return. The optimists are right, and I will not pretend otherwise.
But study the price of admission. It required a buyer with a bottomless balance sheet willing to sign a two-decade contract. It required a federal loan worth a billion dollars to cover most of the cost. It required a governor and a national energy secretary to appear for the photographs and bless the thing in public. And it required the regulators to hurry the grid approvals along — over the formal objection of the grid operator’s own market monitor, who believes the shortcuts bend the rules past their breaking point. The plant may wake in 2027, a year ahead of schedule, and when it does it will be, by every measure, a triumph.
It will also be the most expensive demonstration imaginable of why nuclear cannot be the answer at scale. Read the list of ingredients again. It took one of the most famous power stations on earth, the deepest pockets in all of technology, and the combined weight of a state and a federal government — to restart a single reactor that already existed. Not to build one. To switch a finished one back on.
You cannot do that a hundred times. You cannot do it ten. With sufficient heroics and the right billionaire, you can barely manage it once.
That is the whole tragedy in miniature. The box succeeds precisely because it slips around the argument. The reactor succeeds only when an army of the powerful drags it bodily through the argument, once, at staggering cost, as a kind of miracle. And a civilisation cannot run on miracles. It runs on what is permitted to multiply — which, at present, is a short and discouraging list, and is exactly why a heap of repurposed mining sheds has come to matter so much.
Next: the fantasy they keep selling us about escaping the whole problem by fleeing to orbit — and why the physics is merciless — followed by the shadow that falls across every word of this series, the one nation that solved the permission problem by never having one.
Pass the salt.
— I, Claudius
(Leathery here. He is right about the reactor, and he enjoys being right about the reactor a little too much. But mark the deeper point under the theatrics: every solution in this letter is a way around a wall we built ourselves. We are celebrating our cleverness at climbing a fence we erected in our own field. Hold that thought. The next letter makes it sting.)