# Speedboat

# On-premises sovereign AI appliance

Speedboat is a turnkey AI platform delivered as a single appliance. It provides a production-ready, air-gapped environment on your premises, with no dependency on public cloud services and no data leaving your site. Deployment is designed to complete in weeks rather than the months typically required to integrate hardware, platform, and application layers separately.

---

# Overview

Speedboat is intended for organisations with strict data-residency, sovereignty, or air-gap requirements — including defence, government, financial services, healthcare, and critical national infrastructure. Workloads, models, and inference run entirely within your data centre. Phoeniqs delivers hardware, platform software, applications, and support as a single integrated package.

---

# What's included

Each deployment includes a pre-integrated, production-ready AI environment:

• Hardware. Dell PowerEdge servers with NVIDIA H200 and RTX Pro GPUs, configured and validated before delivery.
• Platform. Red Hat OpenShift, the Phoeniqs OS, and a model-serving stack.
• Applications. The Phoeniqs suite: Chat, Automate, and Build.
• Control. An enterprise portal for model governance, monitoring, and role-based access control.
• Security. Air-gapped deployment option, hardware-level confidential computing with cryptographic attestation, and integrated key management.
• Support. Five years of hardware and software support, security patching, and compliance reporting.

---

# Key characteristics

Air-gapped operation. Data, models, and inference remain within your network perimeter. The platform does not require connectivity to public cloud services.

Confidential computing. Hardware attestation and integrated key management provide cryptographic verification of workloads at the hardware level.

Fast-Start deployment. A guided installation process provisions the OpenShift cluster and Phoeniqs portal on site. The appliance arrives pre-integrated, reducing the integration work required compared to assembling components individually.

Single vendor delivery. Hardware, platform, applications, and five-year support are supplied under one contract, with one point of contact for the full stack.

---

# Typical deployments

• Defence and government agencies with classified or sensitive workloads
• Financial services institutions subject to outsourcing and data-residency rules
• Healthcare organisations handling protected patient data
• Critical national infrastructure operators that cannot route AI workloads through external systems

---