DATACENTER

Consumer GPUs, hydro power, cold air that does the cooling for free most of the year. The cards take jobs on the GPU marketplaces around the clock. Net income market-buys DATACENTER and burns it. Creator fees buy the next card. The timeline shows where that ends up.

DATACENTER
6GPU
6GPUs
net income / month
bought & burned / month
creator fees → GPU #7read from the chain
$0.000.000 SOL
of
0% · to go
every fee the coin generates lands here. when the bar fills, a card gets ordered.
scroll · the datacenter

The datacenter

Six identical boxes: an RTX 4090 with 24 GB, an i9-14900K, 64 GB DDR5, 2 TB NVMe, a 1000 W Platinum supply.

Six RTX 4090 cards on two shelves of a black steel rack against a white workshop wall
rack 01 · top shelf 01 to 03, bottom shelf 04 to 06finnmark · NO4
GPUs in the rack
working
net per GPU / month
net for the rack / month
marketplaces
expected uptime

the loop

two flows. both point at the coin.

The cards earn money whether anyone holds the coin or not. That money buys the coin and burns it. The coin's creator fees do not pay anyone a salary. They buy the next card, which earns more, which burns more.

treasury, live on-chain
towards GPU #7
DATACENTER burned
buybacks so far
01 · marketplace income

100% of net income buys and burns.

The cards are listed on the GPU marketplaces people already use. Jobs arrive on their own: inference, image generation, fine-tunes. Payouts land in USD. Power is paid first, because the cards do not run on optimism. The rest is converted to SOL, DATACENTER is bought on the open market and sent to the burn address. Every buyback is listed below with its transaction.

vast.ai / runpodUSD− powerSOLbuy + burn
02 · creator fees

Fees buy hardware.

Trading fees from the coin accrue to the treasury wallet. When the balance covers a new host and card, one gets ordered, built on the workshop bench and racked as the next node. Then it starts earning, and the first loop gets bigger. The treasury balance you see here is read from the chain, not typed in.

datenet incomeSOL spentburnedtx

how the loop runs

five steps, then again.

01

the cards work

Six 4090s listed on the marketplaces, online around the clock. Jobs come to them. Nobody here picks the work.

02

payouts land

The marketplaces pay out in USD when a job ends, normally weekly in practice.

03

power gets paid

Hydro power in NO4 is cheap, but it is not free. The bill comes off the top, always.

04

buy and burn

What is left becomes SOL, buys DATACENTER on the open market, and goes to the burn address. Transaction in the ledger.

05

fees buy the next card

Creator fees accumulate in the treasury. Enough for a host and a card means GPU #7, and step one gets bigger.

the farm

Economics

Home hosting pays modestly. A 4090 that is busy most of the month covers its power many times over and slowly pays itself off. None of that was ever going to be a salary, which is why all of it can go to buying and burning the coin without anyone missing it.

Why north

  • Cheap hydro. NO4 is usually the cheapest price area in Norway, and household power in Finnmark has no VAT and no consumption tax.
  • Free cooling. Ambient air is below 10°C most of the year. The cards run cooler than they would in any apartment.
  • Waste heat. Six loaded cards are a 2 kW heater. In winter the workshop needs that anyway.
Finnmark, Norway · exact address not published

Hardware, as facts

  • 6× NVIDIA GeForce RTX 4090 24 GB GDDR6X
  • Intel Core i9-14900K, 24C/32T, per box
  • 64 GB DDR5-6000 · 2 TB NVMe Gen4 · 1000 W Platinum
  • Fiber, typically 500 to 1000 Mbps down, strong upload
  • Arctic ambient air + case fans / AIO. No water plant.

FAQ