How SONALAB Scales Audio AI on EU-Sovereign Compute and Grows with Lyceum

AT A GLANCE
Client	SONALAB
Industry	Media Post-Production / Voice AI
Use Case	Early-stage workloads on Lyceum infrastructure, with GPU scale-up on the roadmap
Lyceum Products	Lyceum Infrastructure (current + upcoming)
Key Result	Cost-efficient EU compute with a provider that grows with them

Client Overview

SONALAB is a European audio AI company building tools and infrastructure leveraging voice AI for media post-production. SONALAB develops AI-powered audio production solutions which require reliable, cost-efficient compute infrastructure to support development and production workloads.

The Challenge

SONALAB needed straightforward, affordable compute to run their early workloads and support data pipelines and model development. But with GPU-intensive workloads on the roadmap, they needed more than a point solution. They needed a provider they could grow with.

The core requirement was cheap, sovereign EU compute from a provider capable of covering the full compute stack and scaling with them over time. This ruled out US-headquartered hyperscalers where EU data residency and pricing models create friction for European AI teams. For SONALAB, guaranteed EU data residency was non-negotiable.

The Solution

Getting started on Lyceum infrastructure

SONALAB runs their current development and processing workloads on Lyceum infrastructure — cost-efficient, EU-hosted, and ready to scale — reducing the overhead of managing compute infrastructure across multiple vendors. As their customer base grows, they plan to expand their Lyceum footprint while maintaining a streamlined provider relationship, unified billing, and verified EU data residency throughout. When model training requirements eventually come into play, the infrastructure will already be in place.

The advantage with Lyceum — growing with the customer

Lyceum's full-stack compute offering means SONALAB never needs to re-evaluate providers as they scale. As their needs grow, their entire compute stack can scale on one platform. Fewer vendors, less overhead. They have what they need today, and more is ready when they need it.

The Results

By running their core AI workloads on Lyceum infrastructure, SONALAB has reduced the vendor complexity that typically slows down early-stage AI teams and positioned themselves to scale into GPU workloads without switching providers or renegotiating contracts.

"EU data residency is a core pillar of our business strategy. This alone makes Lyceum the best partner for us. But what sold us is how simple it is to work with them. Need something? A single email or Slack message and it's sorted. We have the hosting we need right now, and when our workloads grow, we don't have to think twice about where to go next."

— Sören Hübner, CTO at SONALAB