The hardware.
An application-specific miner does one calculation, very fast, for as long as it stays economic to run it. Everything else about this business follows from what that machine costs to buy and what it costs to keep switched on.
Talk through hardware choice 30 minutes. No cost. I will tell you if it does not fit.
Where it runs matters
as much as which one you buy.
The same machine in two different facilities has two different lifespans. Heat, repair turnaround and power stability decide how much of its service life you actually get. This is the part a buyer has no realistic way to assess alone, and it is most of what I am for.
Thermal management
Machines held near the top of their thermal envelope degrade faster and fail more often. A facility that runs hot is quietly shortening the life of an asset you paid for.
Repair turnaround
Fans and power supplies are consumables. What matters is whether spares are held on site and swapped in days, or shipped in and swapped in weeks. Every day down is production you do not get back.
Who you actually diversify against
Spreading machines across several sites owned by one operator is not diversification. That operator can fail, or make one bad decision, and every site goes with it. Real diversification means separate hosting companies, and where it can be arranged, separate electrical grids, so that one grid event does not take your whole position offline at once.
The rate is a bundle, not a meter
Hosting is a bundled service fee quoted in dollars per kilowatt-hour as a billing convenience. It is not a resale of electricity, there is no underlying bill to audit it against, and comparing facilities on that number alone compares bundles rather than power. What is inside the bundle is the thing to compare.
The contract is where the risk sits
Term length, the mid-term rate change provision and its short election window, and the security interest the facility takes over your machines and the Bitcoin they produce. I negotiate these and I will walk you through what you are signing. Read the full disclosure.
Watts per terahash is the whole argument.
Hashrate tells you how much work a machine does. Efficiency tells you what that work costs. Two machines quoted at similar hashrate can have completely different futures, and the cheaper one to buy is very often the one that stops being worth running first.
| Generation | How it tends to get sold | Efficiency | Unit |
|---|---|---|---|
| Current hydro | Best efficiency available, highest capital cost, needs a facility that supports it | 11 | J/TH |
| Current air | The mainstream choice for most deployments | 15 | J/TH |
| Previous generation | Offered as a bargain on capital cost | 21 | J/TH |
| Two generations back | Frequently resold to people new to this | 29 | J/TH |
A machine at 29 J/TH burns nearly three times the power of one at 11 J/TH for the same work. At a fixed hosting rate that lands entirely on your bill. It also decides the order in which machines get switched off when difficulty climbs or the price falls, because the least efficient machine on the network is always the first one that stops making sense.
How I source it.
At cost
The vendor invoices your business directly, at the price I negotiated. I take no spread and I hold no stock, so I have nothing to clear.
In your name
Title passes to you, never to me. That is what makes it your asset, your depreciation schedule and your machine to sell.
Serial recorded
Every unit is recorded by serial against your deployment, which is what your accountant needs and what keeps your machines identifiable.
No stock pressure
I am not trying to move inventory. If the right answer is to wait a quarter for the next generation, that is what I will tell you.
Find out whether this fits.
Thirty minutes on your capital, your timeline and your entity. If the numbers do not work I will say so on the call.
Book a discovery call No cost. No obligation. No follow-up sequence.