Hashprice, written out

The headline figure is a day's worth of WHAT block rewards at market price, divided by WHAT's estimated hashrateHow many hashes a second are being spent on a chain.. The result is in ADA per TH/sA trillion hashes; hashrate is usually quoted in terahashes per second, written TH/s. — a trillion hashes a second — per day, published beside Bitcoin's.

Three of the four factors come out of WHAT's own blocks — out of the data each one writes to the chain's tip, which sits on Cardano where anybody can read it — or are fixed by the protocol. The fourth is a market price, which is the one input that is nowhere on the chain. Until a market price exists there is no figure to publish, and the page says so rather than standing an estimate in its place.

whitepaper §9

Four factors, one of them empty

WHAT — computed in the open

R × N × P  /  Ĥ  →  ADA / TH/s / day
R — reward
Not published yet — 3 January 2027 The reward in force at the current height, which the tip itself records, set by the halving schedule.
N — rate
144 blocks a day, nominally — what a ten-minute target comes to. §6 is explicit that the rule controls the block rate rather than the block count and keeps no memory of cumulative error, so this is the rate the rule aims at and not a daily count you can bank.
P — price
Until there is a market price Market price — the only factor that has to come from outside. Until there is one, this factor stays empty and the quotient is not computed. Nothing goes in here and there is no way to put anything in.
Ĥ — hashrate
Not published yet — 3 January 2027 Estimated from each block's committed target and its Cardano-attested time, both of them in the tip data the block writes. The estimator and its two windows are named.
WHAT hashprice
Not computed Because P is empty. An incomplete calculation shown as incomplete.
Bitcoin hashprice
Not published yet External market data, with its source named. It is here because the testable claim is a comparison. Not derived from WHAT's blocks and not checkable against them.
Fig. 6 — Hashprice, computed in the open. Not a calculator: the factors are named slots, three from WHAT's own blocks or fixed by the protocol, one visibly empty. Bitcoin's figure sits beside it as external market data, marked wherever it appears. No price field, no slider, no axis in any currency.

whitepaper §9

Two thresholds, because a rig's alternative is not the same

The figure above is per terahash, and it is the same number for every machine — hashprice does not vary by hardware. What varies is the threshold a rig compares it against, and what a rig gives up by coming here decides which of the two applies.

A rig that is switched off
Gives up nothing by coming here, so the only thing it has to beat is its own power cost per TH. It arrives when WHAT's hashprice exceeds that.
A rig that is running on Bitcoin
Gives up Bitcoin's revenue, so that is what it has to beat. It arrives when WHAT's hashprice exceeds Bitcoin's.

The testable claim is that once a threshold is crossed and stays crossed, hashpower on that side of it arrives within 1,008 blocks — a week at the ten-minute target. A crossing that holds a week with nothing arriving counts against the claim. Until either crosses, there is nothing to report.

The obsolete end is the near one, and it is the end a launch would be expected to reach first — but for a specific reason worth naming. Hardware whose power is already paid for or near free, curtailed or stranded or behind generation that would otherwise be spilled, clears a threshold of almost nothing and can arrive at a hashprice far below Bitcoin's. A rig that is switched off because its power costs more than Bitcoin pays it is the exception: its threshold sits above Bitcoin's, not below, and it is the last to arrive rather than the first.

Half of that calculation needs no trust at all — the hashrate — and it is worth seeing exactly how little it depends on us.