# On-prem Deployments FAQ

Speedboat is a turnkey, on-premises AI appliance delivered as a single package. It combines pre-configured Dell PowerEdge servers with NVIDIA H200 and RTX Pro GPUs, Red Hat OpenShift, the Phoeniqs OS, a model-serving stack, and an enterprise portal for model management, monitoring, and role-based access control. It also includes the Phoeniqs application suite (Chat, Automate, Build). The system is designed to take an organisation from no AI capability to a running, self-contained platform within weeks.

No. Speedboat is built to run entirely inside your own data centre with an air-gapped deployment option, so there is no reliance on public-cloud services. It supports confidential-computing attestation and integrated key management, meaning workloads can be cryptographically verified and data remains under your control at the hardware level. This is the core reason the product exists: to serve organisations that are barred from routing AI workloads through external infrastructure.

Once the appliance is shipped and on site, deployment is intended to be fast. The installation uses a guided "Fast-Start" process that brings up the OpenShift cluster and makes the Phoeniqs portal reachable, with the goal of a working platform in a very short window rather than the months a from-scratch build would take. Exact timelines depend on your site readiness, particularly network isolation requirements.

You bring your own models. Speedboat delivers the full platform to serve, govern, monitor, and version models, but custom model development is client-supplied and sits outside the standard package. Ongoing data-science consulting beyond initial handover and integration with third-party SaaS AI services are also out of scope. If you need GPU-as-a-Service scaling later, that is on the roadmap as an add-on.

Speedboat includes five years of combined hardware and software support, covering the OS subscription and security patching. The support model includes timely application of critical security advisories and compliance reporting, so your platform stays current and auditable over the contract term without you having to manage the underlying stack yourself.