Compute
Storage
For consumers
Workloads
Why Hivenet
Sign in to Hivenet
Access your files and manage your storage account.
Store
Access your files and manage your storage account.
Compute
Launch instances and manage your compute resources.
Store and Compute still use separate accounts today. We’re working on a unified sign-in.
Just need to send a file?
Use Send directlyNew to Hivenet? Get Started
Get started with Hivenet
Choose what you want to use first.
Store
Back up photos and files across your devices.
Compute
Launch self-serve GPU compute for demanding work.
Store and Compute still use separate accounts today. We’re working on a unified account experience.
Need Hivenet for business?
Talk to salesNeed to send a file?
Open SendAlready have an account? Sign in
Hivenet turns distributed infrastructure into enterprise-grade cloud services.
Hivenet connects Policloud-backed infrastructure, cloud software, and standard interfaces so teams can run compute, inference, storage, and file movement with reliable performance, cost visibility, and practical sovereignty.
Policloud-backed infrastructure
Compute
Inference API
S3 storage
Store
Send
Standard tools
Regional deployment paths
Practical sovereignty
Full-stack infrastructure below. Cloud services above.
Hivenet turns distributed infrastructure into cloud services people and teams can use: GPU and CPU compute, managed inference, S3-compatible storage, personal file storage, and large file transfer.
The architecture matters because performance, pricing, placement, and control are connected. A cloud service is easier to trust when the infrastructure path underneath it is visible.
Cloud platform layer
Hivenet products
Compute
Inference API
S3 storage
Store
Send
Private AI
Platform software
Hivenet workload routing, console, APIs, storage logic, billing, access patterns.
Infrastructure layer
Policloud-backed capacity and enterprise-grade infrastructure paths.
Wider stack
Antimatter full-stack infrastructure system.
Hivenet is part of a full-stack infrastructure system.
Hivenet is part of the Antimatter infrastructure stack. That gives Hivenet a different foundation from providers that only repackage capacity from another cloud.
Site and energy layer
Data Factory supports the site and energy layer behind Antimatter infrastructure.
Modular infrastructure layer
Policloud provides modular infrastructure that can be deployed closer to energy, capacity, and regional demand.
Cloud platform layer
Hivenet turns that infrastructure into cloud services for compute, inference, storage, file storage, and file transfer.
One platform. Product-specific architecture.
Hivenet products share the same direction: usable cloud services on infrastructure Hivenet can explain. They do not all work in the same way. Compute, Inference API, S3 storage, Store, and Send each use the architecture that fits the job.
Compute
Teams launch GPU or CPU instances, choose templates or operating systems, connect with SSH, and run their own workloads. Compute is the instance path.
Inference API
Teams use managed OpenAI-compatible endpoints for open-source and foundational models. Inference API is the managed endpoint path.
S3-compatible storage
Teams use standard S3 tools for datasets, backups, media, archives, application data, and AI pipeline files.
Store with Hivenet
Individuals and small teams use cloud storage and photo backup for everyday files.
Send with Hivenet
Users send large files by secure transfer link, without requiring recipients to create an account.
Compute gives teams direct control over the instance.
Compute with Hivenet is for workloads where the team wants GPU or CPU infrastructure and control over the environment. Use VMs or containers, add SSH keys, choose templates or operating systems, and run the stack your workload needs.
GPU and CPU instances
Choose RTX 4090, RTX 5090, or vCPU configurations based on workload fit.
VM and container paths
Use VM-level control when the operating system matters, or container workflows when the job is more portable.
Public Compute API
Automate instance lifecycle, SSH keys, billing, organization workflows, and quota requests through versioned API paths.
Team organizations
Use organizations, role-based access, and shared billing when infrastructure is operated by more than one person.
Inference gives teams a managed endpoint path.
Hivenet Inference API is for teams that want endpoints for open-source and foundational models without running the serving layer themselves. Your team integrates with an OpenAI-compatible endpoint while Hivenet operates the managed endpoint path.
OpenAI-compatible requests
Keep familiar client patterns and update the endpoint configuration.
Dedicated endpoint model
Use per-replica endpoint capacity for steady production workloads.
Region-pinned deployment paths
Choose available deployment locations for workloads that need clearer jurisdiction and placement.
Model catalog
Start from a managed catalog of model families where supported.
Storage architecture depends on the storage job.
Business object storage, everyday file storage, and file transfer are different jobs. Hivenet keeps those product paths distinct so users do not have to treat every storage problem as the same technical system.
S3-compatible storage
Business object storage for datasets, backups, media, archives, application data, and AI pipeline files, using familiar S3-compatible tools.
Store with Hivenet
Cloud storage and photo backup for personal files, everyday folders, online-only files, and cross-device access.
Send with Hivenet
Secure file transfer for one-off delivery, client handoffs, and large files shared by link.
Sharding and encryption where applicable
For storage products that use Hivenet's distributed storage model, files are encrypted, split into fragments, and distributed so no single node holds a complete usable copy.
Infrastructure you can trust for enterprise-grade workloads.
Policloud gives Hivenet a modular infrastructure layer for workloads that need reliable performance, cost visibility, and regional deployment. The value is not hardware ownership as a claim. The value is a trusted infrastructure path built to support serious workloads.
Modular capacity
Policloud-backed infrastructure gives Hivenet a practical way to place capacity closer to energy, region, and workload demand.
Enterprise-grade reliability
Hivenet is built for workloads where predictable performance, stable access, and operational transparency matter.
Distributed computing benefits
The architecture helps improve capacity use, support regional deployment, and reduce dependence on centralized hyperscaler defaults.
Standard interfaces
Use familiar tools such as SSH, S3-compatible APIs, boto3, aws-cli, rclone, OpenAI-compatible endpoints, and the Public Compute API.
Sovereignty means control you can use.
Hivenet treats sovereignty as practical control over location, infrastructure path, access model, operational interface, and exit route. The value is not the word itself. The value is knowing what you are choosing.
Location
Choose available regional deployment paths across France, the UAE, and the US depending on the product and workload.
Infrastructure path
Run suitable workloads on Policloud-backed infrastructure instead of routing everything through default hyperscaler choices.
Access model
Use the access pattern that fits the product: SSH for Compute, managed endpoints for Inference, standard APIs for S3 storage, and product-level controls for Store and Send.
Operational interface
Use familiar tools and workflows such as S3-compatible APIs, boto3, aws-cli, rclone, SSH, OpenAI-compatible APIs, and the Public Compute API.
Exit route
Standard tools and interfaces make it easier to move data and workloads when your needs change.
Use familiar tools where they matter.
A different infrastructure path should not force teams to rebuild every workflow. Hivenet uses standard interfaces, which makes adoption easier.
S3-compatible tools
Use boto3, aws-cli, rclone, aws-sdk, and S3-compatible APIs for object storage workflows.
SSH and templates
Use SSH, templates, OS images, Docker, and your own runtime choices for Compute workloads.
OpenAI-compatible endpoints
Use familiar request patterns for managed inference where supported.
Public Compute API
Automate Compute workflows with versioned API paths and a documented API surface.
Architecture is only useful when it can be tested.
Hivenet's architecture is tied to measured results, research depth, and clear methodology.
Inria-backed research
Hivenet's distributed cloud work is supported by long-running research with Inria.
GPU VM vs bare metal benchmark
Compute with Hivenet measured single-host multi-GPU NCCL AllReduce performance inside a VM and compared it with bare metal.
Cost and pricing transparency
Published pricing and per-second billing help teams understand cost before they run suitable workloads.
Sustainability methodology
Hivenet's efficiency claims are strongest when the methodology and boundaries are visible.
FAQ
Common architecture questions
Understand the infrastructure before you move the workload.
Explore the trust page, read the benchmarks, or talk to Hivenet about the right architecture for your AI, compute, or storage workload.
