Submit
Choose a workload and provide its supported inputs, prompt, parameters or files.
For people and businesses that need compute
Bring the workload.
We find the capacity.
Submit supported AI, GPU and CPU workloads to SolarCompute. The network validates the request, finds a compatible available machine and returns the completed result.
One request. One coordinated network.
Use distributed computing capacity without managing the machines behind it.
You describe the work and choose a supported workload. SolarCompute handles validation, queuing, compatible-machine allocation, execution status and result delivery.
Customers do not connect directly to Host computers. Workloads run through controlled network services using approved runtime configurations.
How it works
From workload request to completed result.
Validate
SolarCompute checks safety limits, runtime requirements and required compute resources.
Queue
The accepted workload enters the network queue with its GPU, CPU and runtime requirements.
Allocate
The network selects compatible available capacity based on reported machine capabilities.
Run
The approved workload executes while status, runtime and billable resource use are recorded.
Receive
Review the completed response, metrics, logs or downloadable result supported by that workload.
Supported workload paths
Choose the type of compute your work needs.
Each launcher applies workload-specific validation, limits and result handling.
01 · GPU
AI inference
Prompt-based AI tasks, questions, summaries and evaluation.
Open AI inference → 02 · GPUAI training
Supported training sessions with progress and completed metrics.
Open AI training → 03 · GPUImage rendering
GPU-backed generation and rendering with downloadable results.
Open rendering → 04 · CPUCPU jobs
Controlled computational tasks using allocated vCPU capacity.
Open CPU jobs → 05 · CPU/GPUSimulation
Supported modelling and simulation workloads with public-safe results.
Open simulation → 06 · CPU/GPUCustom containers
Controlled container validation and execution on compatible infrastructure.
Open containers →Understanding compute rates
Pay for the resources allocated while your workload runs.
GPU-hour means one GPU allocated for one hour. Two GPUs for 30 minutes also equal one GPU-hour.
vCPU-hour means one virtual CPU allocated for one hour. Four vCPUs for 15 minutes equal one vCPU-hour.
Current default rates come directly from SolarCompute’s billing policy. A particular GPU model or workload may use an applicable override. The confirmed price should be shown before paid work is submitted.
Resource count×Runtime×Applicable rate=Estimated compute cost
Controlled execution
The workload moves through SolarCompute—not directly into a Host’s home.
Validated inputs
Requests are checked against supported workload rules and public or account-level limits.
Capability matching
Allocation uses reported hardware capabilities and the approved recipe for the workload.
Status visibility
Submitted, queued, allocated, running and completed states make the execution journey visible.
Result delivery
Results are returned through the supported response, metric or downloadable-artifact flow.
Common questions
Before you submit work.
Are the displayed GPU and CPU prices live?
Yes. The page reads the current default rates from SolarCompute’s public pricing endpoint. Administrators can update network billing policy as capacity and demand change.
Does every GPU cost the default GPU rate?
Not necessarily. The default rate applies unless an eligible GPU-model or workload-specific override is used. Confirm the applicable price shown by the service before submitting paid work.
What happens if compatible capacity is not immediately available?
An accepted workload may remain queued while SolarCompute looks for capacity that meets its requirements. Workload status should show where the request is in the process.
Can I access the Host machine running my work?
No. SolarCompute coordinates the workload through controlled network services. Customers do not receive direct access to Host computers or private Host infrastructure.
Where do I begin?
Use the Early Access page to begin the workload journey, or create an account for dashboard tools and private platform participation.
Run on SolarCompute
Bring useful work to renewable-powered capacity.
Choose a supported workload and follow it from request to result.