Good partnerships don't always begin at the negotiating table. Sometimes a conversation between two exhibition booths is enough. That's exactly how it happened in May 2026 at the KH-IT Manager Conference in Kassel. That's where basebox and FAST LTA crossed paths. Within a few months, that encounter turned into a joint offering. It's no longer aimed solely at hospitals, but at any organization working with especially sensitive data – whether that's patient records, citizen data in a municipality, or classified information at a security agency.

Two Exhibitors, One Shared Topic

KH-IT, the German federal association of hospital IT directors, hosted its annual conference in May. It offers a clear picture of where IT leaders in hospitals currently stand: caught between the desire to put AI to productive use and the obligation to protect patient data.

A few rows of booths away, FAST LTA from Munich presented its offering. For years, the company has specialized in secure storage "Made in Germany" – WORM archives and ransomware-proof backup systems that run on-premises, entirely without any cloud connection. With its Silent AI product line, FAST LTA has applied this same approach to AI hardware: a local AI appliance, GPU-optimized, Active Directory-integrated, and fully "off-cloud."

In short: FAST LTA builds the machines. basebox delivers the software that runs on them. Two companies, one shared conclusion: digital sovereignty isn't a nice-to-have for regulated organizations – it's a fundamental requirement.

From Booth Conversation to Finished Appliance

Organizations can now purchase a machine directly from FAST LTA. basebox comes pre-installed and ready to use. Specifically, this is the Silent AI Core Appliance – a 4U server with an Intel Xeon Gold CPU and PCIe 5 architecture.

How powerful the appliance needs to be depends on the size of the organization. basebox offers three performance classes for this, each matched with the right hardware by FAST LTA. These classes are based on the number of people who will be using AI within the organization:

• Class 1 is suited for organizations with around 200 people – such as a smaller municipality or a specialized hospital.

• Class 2 supports around 500 people – a typical hospital or a larger cluster of government offices.

• Class 3 is designed for 1,000 to 2,000 people – such as a hospital group or a state authority.

Importantly, these classes aren't based on the number of registered user accounts. Anyone within the organization can log in, at no extra cost. What matters is how many people are making requests simultaneously – and even the smallest class supports a multiple of the expected number of users. If performance later proves insufficient, an upgrade is possible. You're buying an entry point, not a ceiling.

At every class level, basebox delivers the complete platform: chat assistant, document analysis, text recognition, speech recognition, custom apps, and an orchestrator for automated workflows. Everything included, with no feature tiers.

Three paths, one appliance. Organizations without their own server room don't have to go without. The machine can also be hosted and operated by basebox in a German data center. And for those who'd rather start without any hardware of their own, there's a third option: FAST LTA additionally provides basebox with a dedicated machine on which we run our cloud variant. Either way, data processing stays within Germany – regardless of which path an organization chooses.

Why This Model Works Far Beyond Hospitals

This partnership originated in the hospital setting. But the underlying problem isn't unique to hospitals. Wherever organizations work with especially sensitive data, the same question arises: how do you use AI operationally without handing control over that data to a cloud provider?

Legally, basebox isn't a classic cloud service. The machine belongs to the organization itself. That's why the combination of FAST LTA hardware and basebox software can be applied across industries:

• Healthcare – for patient data, from hospitals to public health departments.

• Municipalities and government agencies – for citizen and social welfare data, such as in mayor's offices, job centers, or ministries.

• Security agencies and defense – for classified data, where any external processing is off the table from the outset.

• Financial sector – for especially sensitive customer data at insurance companies and banks.

The common denominator: the data must never leave the building. And yet, no organization wants to sacrifice the functionality of a modern AI platform.

Why a Pre-Configured Solution Pays Off

Less integration risk from day one. Hardware and software arrive already matched and tested. Drivers, network topology, system access – all of it is resolved in advance, not discovered as a problem later.

One point of contact for hardware, one for software. Clear responsibilities are worth their weight in gold in regulated environments. If a component fails, it's immediately clear who's responsible. FAST LTA handles phone support, remote sessions, and on-site repairs through its Care program. basebox takes care of platform and model updates.

Predictable costs, no hidden line items. A fully configured appliance with a defined purchase price, maintenance contract, and annual license can be cleanly budgeted. That sets it apart from cloud services, whose actual costs often only become visible after the fact.

Full data sovereignty from day one. The appliance sits within the organization's own premises – or, optionally, at basebox. In either case, basebox is legally not a classic cloud service. Organizations can process patient, citizen, social welfare, and classified data without restriction.

Built for growth, not capped at a ceiling. If performance later proves insufficient, an upgrade to the next class is possible. If that's still not enough, another machine can be added.

A Partnership Born From a Real Conversation

In the end, what emerges is an offering where hardware and software fit together – not just technically, but culturally. Both companies think in terms of data sovereignty first, not the convenience of a cloud login – regardless of whether that data originates in a hospital, a town hall, or a security agency.