SpiLLI Host GPU ROI

ROI by use case

Enterprise access Sign in to unlock Enterprise Use.

Public on-demand tokens are served through the community network and hosts are paid for supplied tokens.

Measurement Campaigns

Enterprise-only private measurements can be inspected and selected to refine ROI estimates for this account.

Measured results Finished campaigns with uploaded measurements appear here.

Observed	Model	Hardware	Output t/s	TTFT	Latency	Workload	Runtime
No uploaded measurements yet.

Campaign list

Capacity cost model

Recover purchase price over time, plus local electricity cost.

GPU or machine Hardware profile used for model fit, throughput, power, and payback math. Model Compatible models are ranked by expected ROI for the selected hardware. Model type Multimodal estimates include extra memory and throughput overhead for vision/audio/video encoders. Target payback, months Desired hardware recovery window; token price is computed toward this target. Utilization, percent of month Share of the month spent serving paid token generation. Electricity, dollars per kWh Power price used to estimate monthly electricity cost. Active user target, tokens/sec Reserved output speed per active user for concurrency estimates.

Monthly revenue $0

Net monthly income $0

Payback period n/a

Annual ROI 0%

Serving estimate

Aggregate throughput: 0 tok/s
Host token price: $0 / 1M
Market price range: $0 / 1M
Payback target: n/a
Parallel users: 0
Cap + over-cap utilization: n/a
Clients that fit at entered usage: n/a
Subscriber/session ratio: n/a
Over-cap usage value: n/a
SpiLLI fee per client: n/a
Model VRAM fit: n/a
Capacity cost: $0/mo

Income scenarios

Utilization	Gross	Cost	Net at computed rate	Recovery

How these estimates are computed

Every value below updates from the selected hardware, model, computed price, utilization, and electricity assumptions.

Compatible model catalog

GPU choices

Prices and output are estimates; refresh the dataset before quoting a host payout.