Why this test could not be run before
Proof of workMaking block producers compute hashes until one comes in under a published threshold, so that only the volume of guessing decides the winner. is a lottery. Each hashOne run of the hashing arithmetic over a header: a single guess. is a ticket, and the winner of each draw collects the coinbase. The ticket competition is what makes an attack expensive, and that expense is what Bitcoin calls security.
The argument rests on something it cannot check. It assumes honest, online hashpower dominates, and that idle or offline capacity will never be mobilised against it. No rule in any protocol can tell you how big that reserve is, and nothing can stop online hashpower from being paid to go somewhere else.
If your first thought is that somebody has already tried this, you are right — several projects have. They solved the hard half: paying in a coin whose value does not depend on the miners staying.
What differs from those earlier attempts is who can mine it. The previous projects set their own puzzle formats, so the machines that answered were GPUs and CPUs — hardware whose behaviour was never the question. What is new here is that Bitcoin's installed base can mine WHAT, unmodified.
Paying miners in the coin they secure collapses on itself
The usual shape of an attempt is this: a new proof-of-work coin pays for hashpower. The problem is what the payment is made of. If the coin's own security is its hashrateHow many hashes a second are being spent on a chain., then the hashpower it attracts is also the thing holding its value up — and the moment that hashpower leaves, the coin it was paid in is worth less. That is the proof-of-work loop, and each step of it holds up the next.
Pay miners inside the loop and the payment works against itself: the power bill is due in money, so the coins get sold, and the selling pushes down the price that was paying for the work.
This has already happened twice, in public. After Bitcoin Cash forked in 2017, miners moved between the two chains according to which paid more that hour; researchers named it fickle mining. Bitcoin SV, split off in 2018, was never worth enough to keep its hashrate expensive, so an attacker could rent enough hashpower to reorganise it — and did, repeatedly, in 2021. Both are the loop running backwards: the price fell, so the hashpower left, so the chain got cheaper to attack, so the price fell further.
So the test has to be run from outside the loop. Pay in a token somebody else secures and the arrow from work to security is gone: Cardano supplies the security instead. Hashpower arrives and leaves without moving what the coin is worth.
Miners still sell to pay the power bill, but a cheaper WHAT cannot make this chain easier to attack: the securing was never bought with WHAT. Cardano's stake pools are paid in ADA, its own coin.
Every market for SHA-256d work has been inside that loop
A secondary market for SHA-256d work has run continuously since about 2013: spot rental at CEX.IO, then Mining Rig Rentals and NiceHash; cloud contracts at Hashnest, Genesis Mining and Bitdeer; hashrate futures in 2020 and hashprice forwards from 2022. None of that is new.
What every one of those markets has in common is what the work buys. The hashing produces an 80-byte header, that header is a block of BTC or BCH or BSV, and the coinbase of that block pays — in the coin of the chain the hashing just secured. Renting changes who receives the coinbase. It has never moved the work outside the loop.
Here the header is a receipt presented to a different ledger, and that ledger issues under its own rules. The work secures nothing on the ledger that pays it, so the reward's value does not move with the hashrate. Nobody has bought SHA-256d that way, so the figure in the hashprice method is the first price anyone has put on work from outside the loop.
Nobody is attacked by any of this. No chain is targeted and nothing is double-spent — hashpower chooses where to go, and the blocks say where it went. Which leaves the question of what the work is actually doing once it gets here.