Baremetal for projects that need guaranteed resources, fast storage, stable networking, and hardware-level access.
Entry baremetal for stable websites, panels, and APIs.
Server for databases, high-load apps, and fast-storage workloads.
GPU server for AI, rendering, video, and heavy compute.
Availability and provisioning time are confirmed before payment.
Large-storage server for archives, backup, media, and internal storage.
Availability and provisioning time are confirmed before payment.
Availability and provisioning time are confirmed before payment.
Availability and provisioning time are confirmed before payment.
Availability and provisioning time are confirmed before payment.
Availability and provisioning time are confirmed before payment.
Availability and provisioning time are confirmed before payment.
Availability and provisioning time are confirmed before payment.
Availability and provisioning time are confirmed before payment.
| Item | What we confirm | Why it matters |
|---|---|---|
| Availability and delivery | configuration and delivery date | GPU and rare configurations are checked for availability first |
| Network | port, traffic, included IP, DDoS profile | traffic limits and abuse rules are clear before launch |
| Hardware | IPMI/KVM, RAID, drives, component replacement | long-term hosting needs a clear recovery process |
| Billing | monthly payment | the project can renew month by month |
Monthly payment is available. Multi-month prepayment is optional, not required.
Before moving, we check the stack, DNS TTL, database, backup, cutover window, and rollback path.
For baremetal, disk replacement, RAM/CPU diagnostics, IPMI access, and remote-hands flow are agreed in advance.
Infrastructure, OS, application, backups, and managed work are separated clearly so incidents have no hidden scope.
For a location, request a test IP, speedtest file, and route check to your region before payment.
Request testBefore ordering, clarify the traffic profile: TCP/UDP, game ports, VPN/proxy, L3/L4, and application needs.
Network termsFor IPv4/BGP tasks, we verify origin ASN, ROA/RPKI, route objects, LOA, and abuse contacts.
IPv4 subnetsPlans show port and traffic terms clearly: unmetered VPS traffic or a stated server traffic package.
View serversProjects with predictable traffic, CPU-bound tasks, and strict performance requirements.
NVMe, RAM, RAID, and dedicated networking help keep latency under control.
Low latency, stable CPU, and network attack protection for game services.
GPU servers for CUDA, inference, rendering, video processing, and batch workloads.
Exact GPU model, VRAM, number of cards, and configuration availability are confirmed before payment.
Multi-GPU workloads depend on CPU, PCIe generation, memory, scratch disk, and network port.
Driver, CUDA, OS, and image versions are agreed before launch, especially for ML and rendering.
Long-running GPU jobs require checking power, cooling, and sustained-load constraints.
Request an invoice, company details, offer terms, and renewal rules before ordering if accounting needs them.
Billing termsOrders, invoices, tickets, and service history are managed through the client area; commercial questions go to sales.
ContactsFor torrents, VPN/proxy, blockchain nodes, mail, BGP, and IP pools, abuse conditions and documents are agreed in advance.
Abuse policyCritical incidents are handled through tickets with priority, service linkage, and clear response time.
SLAChoose dedicated hardware when you need guaranteed CPU, stable I/O, large RAM, RAID/IPMI, GPU, or sustained load without neighbors. If a VPS hits disk or CPU limits, baremetal is often more predictable and cost-effective.
Databases need cores, RAM, and NVMe; game servers often need higher single-core frequency; high-load sizing depends on whether the workload is CPU-bound, I/O-bound, or network-heavy.
GPU model, VRAM, CUDA/drivers, CPU, RAM, NVMe scratch storage, network port, traffic, and load duration. Multi-GPU servers also require checking power, cooling, and configuration availability.
IPs, disks, and network ports depend on the location and chassis. Specify requirements before ordering: IPv4 count, RAID/ZFS, NVMe/SATA, 1/10/25/40/50/100 Gbit/s, and monthly traffic.
We first document the current stack, data volume, and switching window. Then the server is provisioned, data is copied, services are tested, changes are synced, and DNS or IP routing is switched.
The billing period depends on the plan: most services renew monthly, while selected plans have a minimum term. A 6–12 month prepayment, discount, or price lock applies only after confirmation in an invoice or ticket.
A ticket is opened, IPMI/SMART/logs are checked, then disk, RAM, node replacement, or migration is coordinated. Critical projects should use RAID, backup, and managed support from the start.
VPS upgrades are usually available by moving to a larger plan. Baremetal upgrades depend on chassis, slots, drives, and available hardware, so growth should be discussed before ordering.
By default, the customer is responsible for application data. Controlled backup/restore is available as a separate option or managed service.
Yes. Request a test IP, speedtest, and route check for the target location. DDoS-sensitive workloads should describe protocols and traffic profile in advance.