- Deployment
- vpc
- Hq Country
- Germany
- Hq City
- Cologne
- Ownership
- private
- Founded
- 1999
- Training Default
- never
- Residency Options
- Germany (4 pluscloud data centres, georedundant)
- Gov Access Exposure
- none-eu
- Eu Ai Act Role
- deployer
- Certs
- BSI C5, ISO 27001, ISO 9001, PCI DSS
- DPA available
- yes
- Retention Default
- The platform is a dedicated GPU instance running a customer-chosen LLM — "a dedicated GPU with sufficient resources to run an LLM on the inference servers we provide" — so there is no shared inference plane to retain prompts in. No inference retention window is published.
- Retention Exceptions
- plusserver states that with pluscloud open "your data is stored in our pluscloud data centers in Germany and is not subject to the CLOUD Act". The AI platform is built on that same open-source cloud.
- Subprocessors
- plusserver GmbH — four first-party data centres in Germany
- Subprocessor Count
- 1
- Transfer Mechanism
- intra-eu
- Eu Procurement Ready
- pass
- Eu Procurement Reason
- The only row in the registry sold as a private instance rather than a shared endpoint: a dedicated L4, L40 or H100 with an LLM of the customer's choosing, on BSI C5-tested German infrastructure, priced monthly rather than per token. Being a founding member of Gaia-X and the Sovereign Cloud Stack is not a certification and earns nothing here, but the underlying OpenStack cloud is C5-tested and ISO 27001 certified, and "not subject to the CLOUD Act" is an unusually direct claim for a vendor to put in writing. What is missing is anything AI-specific: no published statement about prompts, no retention window, no zero-retention commitment. The architecture makes those questions less pressing — a dedicated instance has no shared plane to leak into — but the buyer is inferring that rather than reading it.
- Training Claim Basis
- inferred
- Training Confidence
- medium
- Residency Claim Basis
- stated
- Residency Confidence
- high
- Retention Claim Basis
- inferred
- Retention Confidence
- medium
- Subprocessors Claim Basis
- inferred
- Subprocessors Confidence
- medium
- Residency Class
- eu-only
- Portability / Exit Path
- serves-open-models
- Subprocessor Jurisdiction
- eu-only
- Underlying model providers
- customer-selected open-weight LLMs (quantised 14B / 32B / 70B tiers)
