Summary
- NTT DATA opened AI Factory Munich on 7 October inside NTT Global Data Centers' operational Munich 2 facility.
- The environment incorporates NVIDIA Blackwell GPUs, Spectrum-X Ethernet networking, NVIDIA AI Enterprise software and Dell AI infrastructure.
- Enterprise and public-sector customers can arrange testing of their own workloads; no additional facility capacity or verified customer performance results were disclosed.
NTT DATA has opened an AI workload testing environment within NTT Global Data Centers’ operational Munich 2 facility, giving enterprise and public-sector customers a place to evaluate their own applications and data before deciding on wider deployments.
Opened on 7 October, AI Factory Munich combines accelerated computing equipment, storage, networking and AI software with NTT DATA’s consulting, security and managed services. The company says customers can assess not only how applications run, but also the governance, data handling and operational arrangements needed when those applications move into production.
The opening is a new testing service inside an existing data centre. NTT DATA has not announced a new Munich data centre building, an expansion in electrical or cooling capacity, or the number of GPUs dedicated to the environment. Describing it as additional commissioned hyperscale capacity would therefore misrepresent the development.
Infrastructure available for customer testing
The hardware and software partners identified by NTT DATA are NVIDIA and Dell Technologies. NVIDIA provides Blackwell graphics processing units, Spectrum-X Ethernet networking and NVIDIA AI Enterprise software. Dell supplies AI-optimised servers, storage and associated services through its Dell AI Factory with NVIDIA offering, while NTT DATA brings together the infrastructure and customer engagement.
Those named components establish the technology families involved, but they do not specify an installed cluster size or performance envelope. The announcement gives no rack count, accelerator quantity, networking topology, storage capacity or workload benchmark. An operator assessing the site’s electricity demand or heat rejection would need separate equipment and facility measurements.
Rather than presenting demonstrations on a remote platform alone, NTT DATA says clients can bring their own applications and data into a live data centre environment. That permits testing against a particular combination of computing equipment and supporting services. It does not mean every client workload will run with identical performance, or that the results will transfer unchanged to a differently configured production installation.
Customer tests could examine how software uses accelerators and communicates with storage and network services, alongside the controls needed to manage the system. The release does not describe a standard measurement methodology or publish results from completed client trials, so comparisons of throughput, latency or cost between this environment and other facilities are not yet substantiated.
Governance and operating requirements
At the opening event, visitors saw demonstrations of orchestration, sovereignty and data residency controls, with applications drawn from banking, manufacturing and healthcare. These were demonstrations of the environment’s intended use; the announcement does not identify customers that have subsequently commissioned production systems as a result of the tests.
NTT DATA frames ownership and control of applications and information as central to the service. Running a test inside a German data centre establishes where that particular environment is located, but it does not by itself prove compliance with every jurisdictional or sector-specific rule. Such an assessment would depend on the customer’s data flows, access permissions, contracts and operational procedures.
The distinction also applies to the software and managed services supporting the hardware. Customers may need to establish who can administer systems, how applications are updated, what information is retained, and which processes apply during incidents. NTT DATA says these requirements can be considered during testing, although the release does not publish an independent security audit or certification for a particular customer’s implementation.
The company positions the facility between an initial application pilot and a decision to operate AI services more broadly. The physical data centre setting gives clients a place to assess their intended workload alongside the infrastructure and management arrangements required for sustained operation. It remains a pre-deployment evaluation service, rather than evidence that any proposed customer architecture has already entered production.
Opening and next steps
Access is available by appointment for enterprise and public-sector organisations, arranged through an NTT DATA account team or its website. The company describes AI Factory Munich as one part of a wider network of resources used to explore, develop and deploy technology, but the 7 October announcement centres on this particular location and its opening.
NTT DATA has not disclosed capital expenditure for the installation, the number of bookings, commercial contract values or independently verified outcomes from visiting customers. Nor has it identified a timetable for increasing the hardware allocated to tests. The presence of Blackwell systems and an operational hosting site should not be interpreted as a commitment to a particular future capacity expansion.
Munich 2 now hosts a client-accessible environment for evaluating AI infrastructure and operational requirements. Whether those tests lead to additional data centre demand, dedicated deployments or measurable efficiency changes will depend on decisions and projects that have not been announced as part of the opening.
