Deutsche Telekom's AI Factory 2.0: sovereign AI cloud for German businesses

Today, 9 November 2026, Deutsche Telekom presents its AI Factory 2.0 at a press conference in Munich, under the title AI Factory 2.0: Greater Digital Sovereignty for Customers in Germany and Europe. The announcement matters for German businesses because the core of it is already running: Telekom's Industrial AI Cloud is in production, and it gives a company a private AI cloud environment within hours, with the data staying in Germany.

What Telekom is presenting today

The press conference in Munich starts at 10:00 and carries digital sovereignty as its theme. It follows a strong third quarter: Telekom reported organic revenue growth of 5.5 percent to 31.1 billion euro, raised its full year outlook, and plans a dividend of 1.10 euro per share. At the Digital Summit in Berlin the week before, T-Systems signed a memorandum of understanding with Sparxity, a Dresden based company, to bring energy efficient AI computing into Telekom data centres. The next chip generation from that cooperation is targeted for rollout in 2028.

Chief executive Timotheus Hoettges has named AI as one of the group's biggest challenges, and has said more than 500 AI projects are already running in Europe. The AI Factory is the infrastructure answer to that: German compute, under German and European rules, for German customers.

What is already running: the Industrial AI Cloud

The most concrete part of the story is not the press conference, it is the Industrial AI Cloud. Deutsche Telekom has put NVIDIA DSX OS into production in its Industrial AI Cloud, the first large scale industrial AI cloud to do so. The system is built so that customers get a private AI cloud environment within hours, instead of waiting days while network, platform and infrastructure teams set it up by hand.

It is aimed at organisations that want to keep sensitive industrial data within Germany, and it is designed to meet Germany's C5 cloud security criteria. Telekom plans to scale from 500 NVIDIA GPUs on the platform today to more than 10,000 NVIDIA Blackwell and Rubin GPUs. For a manufacturing company or a supplier, that means the compute for training and running AI on production data can sit in a German cloud, not on a server abroad.

Why sovereignty matters for a German business

Sovereignty sounds abstract until a customer asks where their data is processed. Three questions come up again and again. First, where does the data physically sit? Second, which law applies to the operator? Third, who can access it, and under what conditions? A German cloud run by a German operator answers all three in a way a US hyperscaler cannot, even with a European region.

That matters most in regulated work: machine data from a factory, patient data, financial records, and anything covered by a customer contract that names Germany or the EU as the place of processing. It also matters for trust. When a customer asks how you handle their data, an answer that names a German data centre is short, clear and easy to check.

Three steps to take this week

First, write down which of your business data must stay in Germany. Customer data, production data, contracts and health related data usually belong on that list. Marketing material and public website content usually do not.

Second, ask your IT provider or your cloud provider three questions: where is our data stored today, who operates the infrastructure, and what happens to our data when an AI feature processes it? If the answers are vague, that is the gap to close.

Third, run one small pilot on sovereign infrastructure. Pick one workflow with real business value, for example checking supplier documents, summarising service reports, or drafting replies to standard customer questions. A pilot shows the real cost and the real benefit before you commit to a larger move.

The takeaway

AI Factory 2.0 is a press conference, but the Industrial AI Cloud behind it is already in production. German businesses no longer have to choose between using modern AI and keeping their data in Germany. The practical work is the same as always: know which data must stay, ask where it sits today, and test one workflow on infrastructure you can explain to a customer. Read our practical guide to sovereign AI for SMEs, and see how workflow automation and digital systems can carry the pilot into daily work.