Xference is building an European infrastructure dedicated to AI inference, designed around one non-negotiable principle: your data never leaves your perimeter. We run open-source models in private environments, with data governance and localization guaranteed by design, not as an option, but as an architectural foundation.
We are the European Inference Provider that brings an infrastructure built to give companies and organizations the power of generative AI, without compromising on data sovereignty, regulatory compliance or energy efficiency. Our infrastructure is hosted via colocation at the Aruba data center and benefits from all the certifications held by the data center located in Ponte San Pietro, Bergamo.
Our infrastructure goes live inside Aruba's IT3 data center in Ponte San Pietro (Bergamo), the first Italian node of the Xference network. Here our servers operate in colocation and Hardware-as-a-Service (HaaS) mode — a combination that lets us scale rapidly and linearly, rack by rack, turning capital investment into flexible operating costs.
This is a high-density infrastructure: each rack exceeds 20 kW, well above the industry average, powered by renewable energy sources. Reliability, security and operational continuity are backed by the standards of one of Europe's leading data center operators.
Ponte San Pietro is the starting point. It is the first in a series of nodes that will progressively come online to serve other European regions.
Xference's vision goes beyond a single data center. The first Italian node is the first step toward a distributed infrastructure across Europe, with dedicated locations covering the continent's different regions.
Our long-term goal is a distributed edge architecture that brings compute nodes closer to end users: lower latency, reduced energy consumption, and a genuine European digital autonomy, built step by step in a sustainable way.
Data Sovereignty
Processing in private environments, with data always under the customer's control and localised in Europe.
Compliance by Design
Security, governance and compliance aren't add-ons; they are an integral part of the architecture from the ground up.
Efficiency and Sustainability
High power density managed with efficiency in mind, renewable energy sourcing, and an edge roadmap that further reduces consumption.
Real Scalability
The colocation + HaaS model lets us expand capacity linearly, following the growth of workloads monitored in real time.
Our infrastructure is now live in an invite-only beta phase, involving the first 100 selected early adopters. They are testing the service in real-world scenarios, helping us optimize it ahead of a gradual, controlled rollout.
Alongside direct access to the platform, our API service is also in testing and controlled rollout, designed for those who want to integrate Xference's private AI inference directly into their own applications and workflows. The goal is to deliver the same data sovereignty and the same security and compliance standards at the integration level too: data stays within the customer's perimeter, while processing runs on a European infrastructure.
AI Private Inference Technology made in Europe
