Deutsche Telekom will expand its AI factory in Munich’s Tucherpark, CEO Tim Höttges announced on 9 November, according to the company’s press release. Telekom says it is adding Nvidia Blackwell Ultra B300 GPUs to its B200 systems, which doubles the number of chips of that class, and extending the site to 20 megawatts.
Telekom says the first stage, launched in February, was “already fully booked seven months later”, in our translation. It describes the new spending as just under a further quarter of a billion; Bloomberg put it at about €250 million.
Business Pill · Sovereign AI Explained
An evergreen Business Engineer explainer on sovereign AI. It teaches the general idea only and does not discuss Deutsche Telekom.
The key insight: Deutsche Telekom is adding Nvidia Blackwell Ultra B300 GPUs to its Munich AI factory, doubling the number of chips in that class, and extending the site to 20 megawatts. It says the first stage was fully booked seven months after its February launch and that it is investing just under a further quarter of a billion; Bloomberg puts the figure at about €250 million.
What Telekom Announced
The release lists three changes: further Nvidia Blackwell Ultra systems (B300) alongside the existing B200 systems, a doubling of the number of chips in that class, and capacity of 20 megawatts after the expansion. Telekom says computing power rises by more than the chip count because the new chips are more powerful, and that Nvidia RTX PRO 6000 Blackwell Server Edition GPUs stay in use.
When the factory opened on 4 February, Telekom’s launch release said it ran on around 10,000 Nvidia Blackwell GPUs, including DGX B200 systems, with computing power of up to 0.5 exaFLOPS. The expansion was announced 278 days later, by our arithmetic.

Who Uses It
Telekom lists Currenta, Siemens, Synopsis, Agile Robots, Wandelbots, PhysicsX, Noxtua, EDAG and Soofi among the users. T-Systems CEO Ferri Abolhassan said, in our translation: “The first expansion stage is fully utilised, well-known industrial companies such as Siemens or Daimler Trucks already use the infrastructure and a well-filled project pipeline shows further demand”.
Telekom says the first stage raised available AI computing capacity in Germany by 50 percent, and that it began planning further stages in May. It also says it is examining, with partners, a larger engagement in Bavaria, that talks are in an intensive phase, and that no final investment decision has been taken.
Connecting the Coverage
Germany’s Nvidia agreement is in Germany Signs Nvidia AI Pact Reportedly Backed by €100M+, and Europe’s AI funding this year is in Europe’s AI Funding Hit $27.6B in 9 Months; London Took 54%.
One Korean chipmaker’s NVLink Fusion tie-up with Nvidia is in Korea’s Rebellions to Link Its NPUs to Nvidia NVLink Fusion. Norway’s grid operator’s pause on new data-centre grid reservations is in Norway Pauses Data Centre Grid Bookings; Pipeline Tops Peak.
The Structural Read
As we read the release, Telekom presents the expansion both as following through on its February announcement and as a response to demand: it says the first stage was fully booked and that it began planning further stages in May.
As we read the customer list, the users Telekom names are mostly industrial companies and institutes, such as Siemens, Currenta and EDAG, alongside start-ups such as Agile Robots and Wandelbots.
Ferri Abolhassan, Telekom board member and T-Systems CEO, 9 November 2026 (our translation)
“The first expansion stage is fully utilised, well-known industrial companies such as Siemens or Daimler Trucks already use the infrastructure and a well-filled project pipeline shows further demand”
Three Implications
FOR INDUSTRIAL USERS Telekom says the expansion adds Nvidia B300 GPUs and doubles the number of chips in that class.
FOR CAPACITY Telekom says the site will be extended to 20 megawatts.
OPEN QUESTION Telekom says it is examining a larger engagement in Bavaria with partners and has not taken a final investment decision.
The Business Engineer Lens
This story maps onto the Business Engineer framework The State of AI Data Centers.
The framework says: “But infrastructure reality is catching up with AI ambitions”.
As we read it, Telekom’s release is one example of that: the company says its first stage sold out in seven months and that the next step is measured in megawatts, 20 of them.
What Is Not Established
The release we read does not give the number of B300 GPUs, an exact investment figure or the site’s capacity before the expansion. The €250 million figure is Bloomberg’s; we did not read Bloomberg’s article. The Bavarian expansion is under review, with no investment decision. Quotations are our translation from German.
The Bottom Line
Deutsche Telekom is adding Nvidia B300 GPUs to its Munich AI factory and extending it to 20 megawatts, after the first stage was fully booked seven months after its February launch, according to the company. It says it is investing just under a further quarter of a billion, and Bloomberg puts the figure at about €250 million.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Figures and quotations are from Deutsche Telekom’s releases of 4 February and 9 November 2026, published in German; the English quotations are our translation. The €250 million figure is Bloomberg’s, from its post; we did not read the Bloomberg article. The day count is our arithmetic.
Sources: Deutsche Telekom, KI-Nachfrage steigt: Telekom erweitert KI-Fabrik in München (9 Nov 2026) · Deutsche Telekom, launch release (4 Feb 2026)









