Lenovo TruScale IaaS meets the growing demand for sovereign AI
When organisations set out to embed artificial intelligence into their day‑to‑day processes, raw compute power is only half the equation. Data‑sensitive sectors such as public administration, healthcare and finance also need to keep workloads under strict jurisdictional control, meet EU‑wide compliance rules and avoid surprise cost spikes as usage expands.
Indasys IT Systemhaus AG, a Baden‑Württemberg‑based IT services firm, saw these challenges as a chance to create a new revenue stream. By turning its own AI experiments into a commercial, locally‑hosted cloud service, the company could offer German businesses a secure, sovereign alternative to the big public clouds.
Building the platform – hardware and consumption model
Indasys worked closely with Lenovo and NVIDIA to assemble a private‑cloud AI stack that could be delivered as a service. The core hardware consists of:
- Lenovo ThinkSystem servers equipped with Intel Xeon 6 processors – providing the CPU horsepower for data‑intensive preprocessing.
- NVIDIA L4 Tensor Core GPUs – optimised for inference workloads and low‑latency model serving.
- Lenovo enterprise storage – ensuring that large training datasets stay on‑premises.
What makes the deployment distinct is the Lenovo TruScale Infrastructure‑as‑a‑Service (IaaS) model. Rather than buying the rack‑level equipment outright, indasys pays a monthly fee that scales with usage, allowing the firm to add or retire capacity as demand fluctuates. This pay‑as‑you‑go approach eliminates the need for large upfront capital expenditure, a key selling point for mid‑size firms wary of over‑investing in AI hardware.
From internal efficiency to marketable services
Indasys first used the platform to streamline its own operations. AI‑driven analytics now sift through public‑sector tender documents, flagging opportunities and recommending bid strategies. The result? A six‑fold increase in the number of tenders the company can handle, with more than 150 AI‑assisted submissions and a 32 % win rate.
Support teams also benefit. Real‑time AI scheduling tools map engineer availability, automatically allocating tasks and reducing idle time. These internal gains gave indasys concrete case studies to showcase to prospective clients.
Extending the offering through BaWü Cloud and Agentum GmbH
With a proven internal model, indasys is now commercialising the platform via two channels:
- BaWü Cloud – a regional cloud marketplace aimed at small and medium‑sized enterprises (SMEs) in Baden‑Württemberg. The service promises full data sovereignty, a crucial advantage as the EU AI Act tightens rules around where AI models and training data may reside.
- Agentum GmbH – a newly created AI consultancy that packages modular, private‑cloud solutions for customers who need on‑premises or locally‑hosted deployments. According to CEO Tobias Drescher, “our modular AI private‑cloud offering built with Lenovo TruScale allows our customers to effortlessly adopt AI while protecting critical intellectual property.”
Both avenues rely on the same underlying Lenovo‑powered infrastructure, giving indasys a repeatable, scalable product that can be tailored to the security and compliance requirements of sectors such as public administration and healthcare.
Why UK businesses should take note
Although the story is centred on a German firm, the underlying model has relevance for UK organisations that are increasingly concerned about data residency post‑Brexit. The combination of:
- Local control – workloads stay within national borders, easing GDPR and upcoming UK‑specific AI regulations.
- Flexible consumption – TruScale’s monthly pricing mirrors the pay‑as‑you‑go models popular with UK cloud adopters.
- Accelerated hardware – NVIDIA Tensor Core GPUs deliver the performance needed for inference‑heavy workloads common in edge AI for Android devices.
UK‑based system integrators could replicate a similar approach, pairing Lenovo’s TruScale IaaS with local data‑centres to offer sovereign AI services to finance, health and public‑sector clients.
The broader partnership impact
Lenovo’s involvement goes beyond supplying hardware. The company’s Professional Services team helped indasys design, deploy and operationalise the platform, establishing best‑practice guidelines from day one. Lenovo Vice President Linda Yao highlighted that “AI adoption depends on more than powerful infrastructure – organisations need secure, scalable environments that keep data‑intensive workloads close while helping them move quickly from pilots to outcomes.”
The partnership has also elevated indasys to Lenovo 360 Platinum Partner status, signalling a deep technical alignment that could open doors to further joint solutions across Europe.
Looking ahead
Indasys’s journey illustrates how a mid‑size IT services firm can leverage a flexible, sovereign AI infrastructure to both improve its own efficiency and spin‑off a marketable service. For UK firms watching the AI regulatory landscape, the model offers a blueprint: combine locally‑hosted compute, a consumption‑based pricing model and strong vendor support to deliver AI that respects data‑sovereignty while remaining financially predictable.
As AI moves from experimental pilots to production workloads, the demand for such private‑cloud, as‑a‑service platforms is likely to rise. Lenovo’s TruScale IaaS, paired with NVIDIA’s accelerated GPUs, may become a reference architecture for organisations that need the performance of a hyperscale cloud without surrendering control over their data.
If you’re evaluating AI infrastructure for your business, consider whether a sovereign, pay‑per‑use model could give you the scalability you need without the capital risk of building a dedicated data‑centre from scratch.