InstantGPU
Effective date: [September, 2026]
Last updated: [September, 2026]
Security on InstantGPU is a shared responsibility between InstantGPU and our customers. This page describes what we secure and what remains your responsibility.
Shared Responsibility Model
Area | InstantGPU’s Responsibility | Your Responsibility |
| Underlying infrastructure | Sourcing capacity from vetted providers; platform-level access controls | N/A |
| GPU isolation between customers | Ensuring instance-level isolation and secure wipe/reset between rentals | N/A |
| Platform & API security | Securing the InstantGPU control plane, authentication, and API | Safeguarding your API keys and credentials |
| Data in your workloads | Providing storage/network security features | Encrypting sensitive data, managing access within your instance |
| Models, code & outputs | N/A | Ensuring your models, training data, and outputs comply with law and our AUP |
| Account access | Providing MFA/SSO options where available | Enabling and enforcing them for your team |
GPU Isolation & Data Wiping
Because GPU Cloud instances are provisioned from a shared pool of underlying hardware, InstantGPU requires that all connected infrastructure providers perform [secure disk/VRAM wipe procedures — confirm actual technical measures] between customer sessions to prevent data remanence.
Infrastructure Security Measures
- [Encryption in transit (TLS) for all API and console traffic]
- [Encryption at rest for persistent storage volumes, where enabled]
- [Network isolation/VPC segmentation between customer workloads]
- [Access logging and audit trails for account and billing actions]
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.
