---
# source: src/content/solutions/en/hpe/hpe-private-cloud-ai.md
# route:  /en/solutions/hpe/hpe-private-cloud-ai/
title: HPE Private Cloud AI
tags: [ai, cloud]
vendors: [hpe]
summary: A preconfigured AI platform from HPE and NVIDIA, in your own datacentre. Public-cloud ways of working without the data leaving the building.
photoNeed: "HPE hardware in our hands: a ProLiant or Alletra being racked, cabled, or opened, with the model badge readable"
stub: false
draft: false
kind: product
addon: false
vendorName: HPE Private Cloud AI
status: current
---

## What it is

Private Cloud AI is an AI-optimised private cloud that HPE built together with NVIDIA.
Preconfigured hardware: ProLiant servers, NVIDIA GPUs, networking and storage matched to AI
workloads and scalable across several sizes.

On top sits a self-service layer where data scientists take tools and models without
raising a request for every experiment.

## What it is for

Organisations that want to do productive work with AI and whose data cannot leave the
building. That is the actual use case: the public cloud's way of working with a local
estate's control.

## Why preconfigured is an argument

Building an AI platform yourself means reconciling GPU drivers, framework versions, network
throughput and storage latency before a single model runs. That reconciliation is what has
been done here in advance.

## What to watch

GPUs are expensive and want to be busy. Before size and quantity are fixed, it should be
clear how many teams share the platform and how allocation is governed. That is an
operations question rather than a procurement one.