/ legal · sub-processors
Sub-processors
Draft — pending solicitor review
This document is a working draft and not yet legally binding. It will be finalised after professional review.
Infrastructure
| Sub-processor | Purpose | Data Location | Tiers |
|---|---|---|---|
| Microsoft Azure | Compute (AKS), storage (Blob + managed disks), secrets (Key Vault), edge/DNS (Front Door) | West Europe (EU) | Cloud, Enterprise |
| Amazon Web Services (AWS) | Compute and GPU model training | eu-west-2 (London) | Enterprise, Sovereign |
AI Model Providers
AI inference runs on Inferrex's own proprietary models, served within Inferrex's Azure infrastructure (West Europe) — first-party processing, not a sub-processor. The third-party model below is retained solely as an emergency fallback for the Cloud tier (reliability policy) and is not used in Sovereign or Air-Gapped deployments.
| Sub-processor | Purpose | Data Location | Tiers | Consent |
|---|---|---|---|---|
| OpenAI (GPT) | Emergency fallback only | US | Cloud | Explicit |
Sovereign and Air-Gapped deployments use Inferrex's proprietary model serving exclusively — no external AI provider receives any data.
Billing
| Sub-processor | Purpose | Data Location | Tiers |
|---|---|---|---|
| Stripe | Payment processing, subscriptions, overage billing | US/EU | Cloud, Enterprise |
Communications
| Sub-processor | Purpose | Data Location | Tiers |
|---|---|---|---|
| [To be confirmed] | Transactional email | [TBC] | All |
30-day notice to account email before any new sub-processor. 14-day objection window per DPA.