Geodd now has EU serverless inference infrastructure available for customers who need AI workloads served from within the European region.
Our EU infrastructure is currently hosted in a Norway data center and is designed for teams that care about data residency, regional traffic control, production reliability, and operational support when deploying large language models.
This is an important step for Geodd as we continue expanding our inference platform for global customers with regional infrastructure requirements.
Why EU Inference Matters
For many teams, moving from AI experimentation to production changes the infrastructure conversation.
It is no longer only about whether a model can run.
The real questions become:
- Where is the workload being served?
- Where does the traffic go?
- Can the platform keep performance stable under load?
- Can costs stay predictable?
- Can the provider support the system when something breaks?
For EU customers, regional control is becoming increasingly important. Teams want to know that their inference workloads, traffic, and related customer data are handled within the region they selected.
Geodd’s EU region is built around that requirement.
Available Models in the EU Region
The first models available through Geodd’s EU serverless inference region include:
| Model | Model ID | Precision | Context Length |
|---|---|---|---|
| Kimi K2.5 | moonshotai/Kimi-K2.5 | FP8 | 262,144 |
| GLM 5 | zai-org/GLM-5 | FP8 | 200,000 |
| OpenAI GPT-OSS 120B | openai/gpt-oss-120b | BF16 | 131,100 |
| OpenAI GPT-OSS 20B | openai/gpt-oss-20b | BF16 | 131,100 |
| NVIDIA Nemotron 3 Nano 30B A3B | nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | BF16 | 262,144 |
These models are available through Geodd’s inference platform for serverless usage, giving teams a faster way to test, deploy, and scale production AI workloads without managing their own GPU infrastructure.
Regional Traffic and Data Handling
For customers using the EU region, inference traffic is served from the selected EU region.
Customer-related data and traffic for EU workloads are handled in line with Geodd’s Data Policy and Privacy Policy. We take this seriously, especially for teams building production systems where infrastructure location and data handling are part of the buying decision.
Our goal is to give customers more control over where their workloads are served, while keeping the deployment experience simple.
Region Selection From the Geodd Console
Customers will be able to select the region they want to use directly from the Geodd console.
When a region is selected, inference traffic is routed to infrastructure in that region. This allows teams to separate workloads by region and align deployments with their internal compliance, latency, and data residency requirements.
GDPR Implementation
Geodd is currently implementing GDPR support for its EU offering.
Our EU infrastructure expansion is part of that broader process. We are working toward giving customers clearer regional controls, stronger data handling practices, and infrastructure options that better support European production workloads.
Dedicated GPU Deployments for Enterprise Customers
Serverless inference is the fastest path for most teams that want to start running models without managing infrastructure.
However, some enterprise customers need a higher level of control. This may include dedicated capacity, private deployments, custom networking, workload isolation, or specific infrastructure requirements.
For those customers, Geodd is planning to introduce dedicated GPU deployments in the coming quarter.
This will give enterprise teams more control over their AI infrastructure while still receiving support from Geodd’s engineering team.
Built for Production AI Workloads
Geodd is not just providing access to GPUs.
We are building an inference platform for teams that need models running reliably in production.
That includes:
- Deployment
- Monitoring
- Scaling
- Runtime optimization
- Regional infrastructure
- Engineering support
As AI usage grows inside companies, infrastructure becomes a serious operating layer. Performance, reliability, cost, region control, and support all start to matter.
Geodd is built around that operating reality.
EU Serverless Inference Is Now Available
Geodd EU serverless inference is now live, starting with infrastructure in Norway and support for several high-demand open models.
Teams can now explore EU-ready models through Geodd and begin deploying inference workloads with regional serving support.
Explore available models:
https://geodd.io/models