Partner Offerings
An overview of the cloud services offered by Sweden AI Factory's partners. All information on this page is based on the partners' public information.
Compare the partners
| Service | Berget AI | Safespring | Stormgrid |
|---|---|---|---|
| Inference API Fully managed LLM endpoints with no infrastructure to run. | Available | – | – |
| Embeddings, rerank & moderation Companion endpoints for retrieval, ranking and content moderation. | Available | – | – |
| AI coding assistant AI assistant for software development. | Available | – | – |
| GPU compute GPU instances and reserved capacity for AI workloads. | Coming soon | Available | Available |
| AI platform & MLOps Platforms and pipelines for the full model lifecycle. | Coming soon | – | Available |
| Managed Kubernetes Managed container platforms for cloud-native workloads. | – | Available | Available |
| VMs & bare metal Virtual machines and bare metal instances. | – | Available | Available |
| Managed databases Managed database services. | – | Available | – |
| Block storage Block storage volumes for VMs and databases. | – | Available | – |
| Object storage (S3) S3-compatible object storage. | – | Available | Available |
| Archive & retention Long-term archive with compliance retention. | – | – | Available |
| Backup & disaster recovery Managed backup and DR services. | – | Available | – |
| Dedicated connectivity Private network connectivity to the platform. | – | – | Available |
| Sensitive & classified environments Environments for sensitive or classified workloads. | – | – | Available |
| Edge deployment Deployment of AI workloads to edge locations. | – | – | Available |
| On-prem GPU appliance Portable GPU units for on-site use. | – | – | Available |
Berget AI
Swedish sovereign AI inference
Serverless, OpenAI-compatible LLM API running open-weight models in Swedish data centres.
Offerings (5)
-
Inference API Available
OpenAI-compatible serverless API with open-weight models (Gemma, GLM, Kimi, Mistral, Llama, Qwen, Whisper and more).
-
Embeddings, rerank & moderation Available
Companion endpoints for retrieval, ranking and content moderation.
-
AI coding assistant Available
AI coding assistant built on the sovereign platform.
-
Dedicated inference Coming soon
Reserved inference capacity for larger workloads.
-
Managed AI platform Coming soon
Managed platform covering the full AI lifecycle beyond inference.
Safespring
Swedish & Norwegian sovereign IaaS
OpenStack-based infrastructure: VMs, GPU flavors, S3 storage, managed Kubernetes and backup.
Offerings (7)
-
GPU compute Available
NVIDIA A2 and H100 NVL GPU instances for AI workloads.
-
Managed Kubernetes Available
Managed Kubernetes built on Talos Linux and Cilium.
-
VMs & bare metal Available
OpenStack-based virtual machines and bare metal instances.
-
Managed databases Available
Managed databases operated via Severalnines.
-
Block storage Available
Ceph-backed block storage volumes.
-
Object storage (S3) Available
S3-compatible object storage in two tiers.
-
Backup & disaster recovery Available
Managed backup based on Spectrum Protect.
Stormgrid
Sovereign AI cloud, storage & data platform services
GPU infrastructure tiered by data classification, co-located with AI Sweden, plus S3 object storage and archive.
Offerings (10)
-
GPU cloud (open tier) Available
Pay-as-you-go GPU cloud for non-sensitive data.
-
MLOps pipelines Available
Pipelines and tooling for the full model lifecycle.
-
Managed Kubernetes with GPU pooling Available
Managed Kubernetes with GPU pooling for training workloads.
-
On-prem VMs & bare metal Available
On-premises AI infrastructure with virtual machines and bare metal.
-
Object storage (S3) Available
S3-compatible object storage across three Stockholm sites.
-
Archive storage Available
Long-term archive tier with 17-nines durability and object lock / WORM.
-
Dedicated connectivity Available
Private fiber connectivity add-on for high-throughput ingest.
-
Sensitive data Available
Tiered environments, from enhanced controls for sensitive data to an air-gapped tier for classified workloads.
-
Edge deployment Available
Deployment of AI workloads to edge locations.
-
On-site GPU appliance Available
Portable NVIDIA DGX Spark units for on-site experimentation.
