Question 1

How long does setup take?

Accepted Answer

Our team provisions and hardens the server by hand, typically within 6–8 hours and no later than 24 hours. You receive working credentials — there is nothing to configure yourself.

Question 2

What uptime do you commit to?

Accepted Answer

99.9% uptime, backed by pro-rata service credits. It is a binding term, not a marketing figure.

Question 3

What is the DGX Spark good for?

Accepted Answer

128 GB of unified memory at 273 GB/s. It suits fine-tuning, prototyping and batch or agent inference — Mixture-of-Experts models such as gpt-oss-120B run at roughly 45–60 tokens per second. It is bandwidth-bound, so it is not suited to single-session low-latency chat on dense models.

Question 4

How are backups handled?

Accepted Answer

Backups are managed by our team as part of the service.

Question 5

How do I get support?

Accepted Answer

Through the ticket form in your dashboard. Every ticket is picked up by the team that provisioned your server.

Sign in to your account