NEW Bare Metal Servers with 20G Dedicated Unmetered Bandwidth 20G Dedicated Unmetered Servers Read more STATUS
4.6Trustpilot Contact Sales Get Started

Platform-as-a-Service

Programmable infrastructure for the platforms you ship

There are many tools that offer Managed Kubernetes, application platforms, runtime services, DBaaS and even function platforms. All of these tools support applications and need to be able to deploy them quickly, run them optimally and scale the underlying infrastructure as needed. We provide the underlying infrastructure for all of these tools.

Browse plans Talk to engineering

REST API · Terraform · Instant deploy · GPU-ready

30 Configurations
30 plans · live stock

  • In stock now24
  • Deploy ETA EU< 4 min

< 6 min

Deploy ETA US

Why PaaS providers pick Netrouting

Infrastructure that gets out of your way

The 'iron' underneath your platform (i.e. managed Kubernetes service, app platform, database as a service, function as a service, etc) determines whether your customers get instant burst capacity or instant tickets to deal with it.

  • Backup2 Instant provisioning Cloud VMs in seconds, bare metal in minutes. Predictable monthly pricing on the cloud side so your platform's autoscaler doesn't pay for idle.
  • Firewall REST API + Terraform Every resource is API-addressable. Official Terraform provider, Ansible modules, Pulumi support. Wire it into your platform's control plane like you would AWS or GCP, without the egress bill.
  • Network Free private fabric Zero-cost private networking between your nodes, control plane traffic, etcd, replication, all on the private VLAN. No public-internet exposure, no egress fees.
  • DDoS protection Bare metal + cloud, mixed Run the data plane on bare metal for predictable performance, and the control plane on cloud VMs for elasticity. Same network, same dashboard, same bill.
  • Security Dedicated CPUs Optimized Cloud plans run on dedicated cores, not vCPUs. Your tenants don't fight for cycles, your p99 latency stays where you set it.
  • Apiform GPU-ready A100, A40, A10 and Quadro inventory available alongside the same network and API. Build a managed-GPU platform without integrating a separate vendor for the AI tier.
  • Publicnetwork Predictable billing No per-GB egress, no inter-AZ data charges, no surprise spikes. Unmetered bandwidth at the listed uplink, transparent flat rates on storage and IPs.
  • Custom isos Custom images & ISOs Bring your own golden image, Talos, Flatcar, custom Debian, hardened Alpine, whatever your platform deploys. Full PXE / iPXE / cloud-init support.

Managed Kubernetes & data plane

The substrate behind your control plane

PaaS providers build their differentiation in the control plane, the API, the auto-scaler, the policy engine, the developer experience. The data plane just needs to run, fast, every time. That's where we sit. Bare-metal Kubernetes nodes with predictable IOPS and 100G fabric, cloud VMs that scale monthly for control-plane components, and a private network knitting them together at no extra cost. No egress charges to bleed your unit economics.

Bare metal for K8s data planes

Netrouting bare-metal server hardware, performance-optimized platforms

Global footprint

Deployed in 10+ datacenters across three continents

Select a location closest to your end users or other facilities. All locations will be provisioned with same hardware and network configuration; all will be managed from a single Network Operations Center (NOC).

Automation

Built for platform engineers

Each action within the dashboard triggers an API call. Terraform allows you to manage resources such as servers, IP addresses, VLANs and DNS as code within resources. Webhooks are triggered for all of the lifecycle events for resources managed by Terraform, meaning your control plane is always in sync with the underlying metal - no polling and therefore no drift.

  • REST API with API reference
  • Official Terraform provider on the registry
  • Webhooks for create / reboot / decommission events
  • Per-team API keys with scoped permissions
  • Audit log of every API action

Observability

Visibility, not vendor lock-in

We collect telemetry data from servers via their API or a Prometheus exporter for all relevant network counters as well as BMC sensors and IPMI events (e.g. power status) and BGP sessions. All data is very easily queryable and can be directly fed into your existing observability stack, so you don't have to run yet another monitoring dashboard from another vendor.

  • Prometheus exporter on every host
  • BMC / IPMI sensor data over API
  • BGP looking glass and live RTT data
  • Status page with webhook subscription
  • Open metrics, no proprietary agents required

Platform archetypes

PaaS substrates we run today

  • Kubernetes / OpenShift platforms Bare-metal worker pools where pod density and node-CPU isolation matter, alongside cloud workers for burst. Multi-NIC bonding, BGP CNI peering and hostpath-friendly storage all supported.
  • Internal developer platforms Backstage portals, Argo workflows, Pulumi-driven IDPs that need a predictable underlying compute surface so platform engineers can promise SLOs without weasel-words.
  • Function / serverless layers Knative, OpenFaaS, Fission and home-grown event runners that need warm-pool capacity without paying hyperscaler-tier per-invocation overhead.
  • Database-as-a-service control planes Per-tenant Postgres, MongoDB, ClickHouse, Cassandra fleets where IO predictability and low jitter make or break the offering. NVMe-backed storage at bare-metal speed, no noisy-neighbor surprises.
  • CI/CD & artifact platforms Self-hosted GitLab/Forgejo runners, JFrog/Harbor registries, Buildkite agents on dedicated cores. Cache-locality matters here, keep the registry inside the same fabric as your build fleet.
  • ML / inference platforms Ray, KServe, vLLM, Triton and your own bespoke inference router. GPU pools (A100/A40/A10) sit on the same fabric as your model registry, so weights load fast and outputs ship cheap.

Built for production

Why teams stay with Netrouting

We connect you to the Internet using network engineers (and not order takers) and hardware and infrastructure that is built to last, so we can pick up where you left off when you need us.

  • Expert-Level Support Our staff is available 24 hours a day, 7 days a week to handle network administration and systems management issues as they occur.
  • Scalable Solutions Build whatever depth or breadth your infrastructure needs and then scale as required.
  • Enhanced Security Enable 2-factor authentication and also limit by IP address from the control panel to secure your account.
  • Cost-Efficient Infrastructure You will always receive the best value from your investment as you will be optimized for budget without any compromise on Quality.

Ready to deploy?

200+ teams rely on Netrouting for bare metal, cloud, GPU servers and colocation. Easy to order servers within minutes. We also love to design custom servers for you and your business, with our experienced engineers.

Browse plans Talk to engineering