Resource network
The electricity a workload actually consumes, measured on the node, with its source and regional carbon intensity attached.
Metered in kWh, with a carbon score
The same resource, the same meter, three answers to “whose machine is this?” You choose per workload, and you can move a workload inward when the rules tighten.
On your device
Available
Inside your gram
Available
On a stranger’s machine — not available
Not applicable for this resource
Each node benchmarks itself and signs the result with its own key. Where a benchmark does not exist yet, the node says so and emits nothing in its place. So do we.
Signed metric
energy.joules_per_token
Unit
joules per output token, with how each turn’s figure was obtained
Every signed turn the node serves carries the energy it cost, and the node publishes the joules it spent over the output tokens it spent them on. Each turn’s figure is labelled: read from the processor’s energy counter (Intel RAPL, Linux), read from the card’s own draw with the CPU term modelled (NVIDIA, any platform), or modelled from vCPU-seconds and hardware class. A separate energy.joules_per_hash carries the RAPL benchmark under its true name.
The token figure is real work, not a synthetic load, and it is absent until a metered turn has run. “Estimated” turns are order-of-magnitude, and the label travels with the number so a reader can never mistake a model for a sensor. The GPU reading is the card’s draw at the turn’s end, not an integral; the CPU counter reads the processor package, not the wall socket.
Compute-only networks cannot tell a regulator, a CFO or an ESG committee what a workload cost the grid. We measure it per node, from the silicon’s own counter. This is the resource that makes the rest of the network legible to people who must answer for it.
We surface estimates. We do not bill you for your own electricity.
There is no 24/7 operations centre, because we do not employ one. Instead a node that cannot prove it is healthy is evicted from the ring rather than quietly serving your work. Fail-closed, not fail-silent.
The measurement is yours. Export it, audit it, put it in your disclosure.
Every resource above reaches your workload through the same substrate, whichever ring it was drawn from.
Each node runs k3s. One orchestration layer schedules all seven resources, so a workload moves between rings without being rewritten.
Workloads that cannot be containerised run as virtual machines on the same cluster, through the open-source KubeVirt project. Same scheduler, same meter.
Standard containers, standard VMs, content-addressed objects, exportable receipts. The cost of leaving is the reason to trust the platform.
Energy is metered in kWh, with a carbon score. Each unit of work produces a receipt naming the node that performed it and the price it was charged at, on the one ledger the whole platform shares. We publish no hourly rate table on this page, because a price is meaningless without the signed unit it is counting.
See how we price compute →Not a specification, not a vendor sheet. The best verified result any node on this network has signed for energy — and, below it, everything else that node put inside the same signature.
energy.joules_per_token
125 joules/token
A benchmark reports one number and signs several. These travelled inside the same signature, so they are as verifiable as the figure above — and they are usually the ones that decide whether the machine suits your work.
estimated_turns
76
estimated_turns
Signed by the node. We have not written an explanation for this field yet, and would rather show it unexplained than invent one.
gpu_measured_turns
0
gpu_measured_turns
Signed by the node. We have not written an explanation for this field yet, and would rather show it unexplained than invent one.
joules
1,483,151
joules
Signed by the node. We have not written an explanation for this field yet, and would rather show it unexplained than invent one.
measured_turns
0
measured_turns
Signed by the node. We have not written an explanation for this field yet, and would rather show it unexplained than invent one.
Method
estimated
method
What was actually run, in the words of the node that ran it.
source
signed turns
source
Signed by the node. We have not written an explanation for this field yet, and would rather show it unexplained than invent one.
tokens
11,822
tokens
Signed by the node. We have not written an explanation for this field yet, and would rather show it unexplained than invent one.
turns
76
turns
Signed by the node. We have not written an explanation for this field yet, and would rather show it unexplained than invent one.
energy.joules_per_hashrecorded 2026-07-09The RAPL benchmark runs only on Intel processors with RAPL, on Linux. The energy resource itself is energy.joules_per_token, published from signed turns on every platform once one has run.