CAP Spending Model

CAP is GPU Render Studio’s safety mechanism for controlling overflow runtime spending while guaranteeing your base execution slot. It allows heavy Blender workloads to run longer when needed — without risk of runaway costs.

What CAP Is

CAP defines the maximum extra amount you are willing to spend beyond the base slot price. It never replaces the slot price and never stops a job during guaranteed base runtime.

How CAP Works

The execution slot always runs first and is fully prepaid.
Overflow runtime begins only after the slot window ends.
Overflow is billed hourly based on real GPU consumption.
CAP applies only to overflow spending.
When CAP is reached, the job pauses automatically.

Example Scenario

Flexible Slot: 8h / $4.00
CAP Extra: $4.50

→ Maximum total billed: $8.50
→ Expected runtime: ~12.5h if capacity allows
→ If finished within 8h → billed $4.00 only

Paused Jobs & Resume

When CAP is reached after slot exhaustion, the job transitions to paused_cap. All rendered frames remain intact and billing is locked.

Resume orchestration is planned for Phase 1.4, allowing users to extend CAP or restart execution without re-uploading assets.

Guarantees & Safety Rules

CAP can never interrupt base slot runtime.
Overflow cannot exceed the configured CAP.
Failed jobs are never billed.
Partial jobs are charged only for completed GPU time.
All terminations include transparent technical reason codes.