Last updated Sep 2026. We collect the minimum to run GPU clouds and AI systems well.
Contact details you send, usage metrics (tokens, latency, error rates) for operated clusters, and standard web logs. We never train shared models on your private data without written consent.
Customer data stays in the selected region/VPC/cage. Backups encrypted at rest (AES-256), TLS 1.2+ in transit. SSO, audit logs, and DPAs available.
Contact details available on request — deletion requests honored within 30 days.