InstantGPU
Effective date: [September, 2026]
Last updated: [September, 2026]
Uptime Commitment
Tier |
Monthly Uptime Commitment |
Service Credit if Missed |
| Standard | [99.0%] | [5%] of monthly fees per [1%] below target |
| Dedicated | [99.9%] | [10%] of monthly fees per [0.5%] below target |
| Secured / Enterprise | [99.95%] | Per custom agreement |
Third-Party Infrastructure Disclaimer
This section is central to your multi-provider model — get it reviewed carefully. It should clearly define what counts as “downtime” when the root cause sits with an upstream provider, and cap your exposure accordingly.
- InstantGPU sources GPU capacity from third-party infrastructure providers (see Subprocessors). Our uptime commitment is measured at the InstantGPU platform/API level, not at the level of any single underlying provider.
- Where a GPU becomes unavailable due to an upstream provider issue, InstantGPU will attempt automatic migration to an equivalent GPU on another connected provider. Downtime is measured from the point InstantGPU is unable to provide equivalent replacement capacity within [X minutes], not from the point the original provider’s resource became unavailable.
- InstantGPU is not liable for outages caused by factors outside its reasonable control, including upstream provider-wide outages affecting all connected providers simultaneously, force majeure events, or customer misconfiguration.
Exclusions
- Scheduled maintenance with [X hours/days] advance notice
- Issues caused by Customer Content, misconfiguration, or violation of the Acceptable Use Policy
- Serverless cold-start latency and Public Endpoint response times (covered separately, if at all)
- Force majeure events
Claim Process
- To claim a service credit, submit a request to [[email protected]] within [30 days] of the incident, including the affected instance ID(s) and timestamps. Credits are applied to future invoices and are not paid in cash.
Definitions
- “Downtime” means the GPU Cloud API is unreachable or an active instance is unavailable and cannot be restored via automatic migration within [X minutes].
- “Monthly Uptime Percentage” is calculated as (total minutes in month − downtime minutes) / total minutes in month × 100.
Qmoniqs Software Private Limited
C 2518, Sushant Lok 1, Sector 43, Gurgaon 122002 (Haryana), India
Email: [email protected]
Website: https://instantgpu.ai
InstantGPU personnel do not access Customer workload content in the ordinary operation of the Services. Customer workload content is not used by InstantGPU to train foundation models or general-purpose AI models.
