Three purpose-built tiers — from single-tenant entry nodes to multi-rack dedicated clusters — plus bandwidth, storage, and managed-service add-ons priced by the gigabit, the terabyte, and the hour.
Every plan ships on dedicated, fully-provisioned bare-metal nodes in our Tier-III facilities. No noisy neighbors, no shared kernel — just your workload on metal, billed by the month or by the hour.
Single-region entry node for lean workloads.
or $0.11 / hour · cancel anytime
Specifications
Production-grade dual-socket for databases & APIs.
or $0.31 / hour · monthly or annual terms
Specifications
Multi-rack dedicated infrastructure for HPC & GPU.
starting price · custom configurations available
Specifications
Every plan includes a generous baseline allotment. Burst above it at transparent per-GB rates, or commit to a monthly pool to drop your unit cost by up to 70%.
Symmetric 10 Gbps uplink. 95th-percentile billing, no overage surprises. Ideal for steady-state traffic with modest spikes.
25 Gbps committed with elastic bursting to 100 Gbps. Suited for media, gaming, and live-event workloads with sharp peaks.
Flat-fee symmetric 100 Gbps with no caps, no shaping, no 95th-percentile math. Designed for CDN edges and large-data egress.
Direct cross-connects to AWS, Azure, GCP, Oracle, and IBM Cloud via Megaport & Equinix Fabric. Free intra-region traffic.
Commit to a monthly bandwidth pool and share it across every node in your account. Pool credits roll over for 90 days.
Always-on L3/L4 mitigation at the edge, with optional L7 application-layer rules and BGP-sourced blackholing on request.
Block storage, off-site backup, snapshot schedules, and a managed-engineering tier that turns your bare-metal fleet into a hands-off platform.
Hot, low-latency NVMe volumes attachable to any server in the same region. Snapshots are instant and free to keep for 7 days.
11-nines durability, immutability, and lifecycle rules. Ideal for cold archives, media repositories, and disaster-recovery targets.
Encrypted, geographically redundant backups with point-in-time recovery up to 30 days. Daily and weekly schedules included.
A dedicated engineering pod handles patching, monitoring, incident response, and capacity planning. 24/7 on-call coverage.
Prometheus, Grafana, Loki, and OpenTelemetry collectors pre-deployed. 90-day metrics retention with optional long-term archive.
Spin up H100, L40S, or A100 nodes on demand. Pre-installed NVIDIA drivers, CUDA, and your choice of ML frameworks.
A complete look at what's bundled with each plan — and what you can opt into à la carte. No asterisks, no upgrade walls.
| Feature | Edge | Core | Cluster |
|---|---|---|---|
| Compute & Memory | |||
| CPU cores (physical) | 24 | 32 | 192 |
| RAM | 64 GB | 256 GB | 1 TB |
| Local NVMe storage | 2 TB | 1.9 TB | Custom |
| GPU accelerators | — | Optional add-on | H100 / L40S / A100 |
| Network | |||
| Private fabric | 10 Gbps | 25 Gbps | 100 Gbps |
| Outbound bandwidth | 2 TB / mo | 10 TB / mo | 50 TB / mo |
| Cross-region peering | — | ||
| Security & Compliance | |||
| Hardware firewall | — | ||
| DDoS Shield Pro | Optional | Included | Included |
| SOC 2 / ISO 27001 reports | |||
| Support & SLA | |||
| Uptime SLA | 99.9% | 99.99% | 99.99% + credits |
| 24/7 managed support | — | — | |
| Provisioning SLA | 15 min | 6 min | 4 min |
| Dedicated TAM | — | Add-on | |
Multi-rack deployments, private cages, GPU pods, and bare-metal build-outs with custom silicon — engineered to your spec sheet, under your timeline. We carry the rack, you carry the workload.
Everything we get asked most about contracts, billing, migrations, and what happens when something goes sideways at 3 a.m.
No. Every plan is provisioned with zero setup fee, zero deposit, and no minimum spend on hourly plans. Monthly plans require one month minimum; annual plans unlock 15% off and rollover credits for unused bandwidth.
Hourly plans meter to the second with a 1-hour minimum. You can stop, restart, resize, or destroy a node from the API or control plane at any time. Local NVMe data is wiped within 24 hours of destruction, certified to NIST 800-88.
Overage is billed at the listed per-GB rate, capped at 3× the allotment per cycle. We send a notification at 50%, 80%, and 100% so there are no surprises. You can also pre-purchase a bandwidth pool to drop overage to as low as $0.012/GB.
Yes. Our migration engineers will replicate your data, validate cutover in a parallel environment, and execute a coordinated switchover. For qualifying workloads, the first migration is free and includes a 30-day parallel-run credit.
We do. Verified non-profits, accredited universities, and pre-Series-A startups receive 25% off monthly plans plus waived setup on managed services. Apply through the contact form with your documentation and we'll respond within two business days.
USD, EUR, GBP, JPY, SGD, and AUD via Visa, Mastercard, Amex, ACH/SEPA, wire transfer, and major cryptocurrencies. Enterprise customers can request NET-30 invoicing with a credit check.
Pick a plan, point your domain, push your code. The first month on any Edge or Core node is prorated to the day — and we'll credit you back if we miss the provisioning SLA.