On-demand
Instant capacity
Launch compute in seconds with per-second billing.
- ✓Per-second billing with 1s granularity
- ✓Auto placement across failure domains
- ✓Full lifecycle API & Terraform modules
Global compute fabric
Bare-metal class Elastic VMs in 80+ facilities
Dedicated bandwidth, NVMe storage and programmable networking close to your users.
- 18Global regions
- 52Availability zones
- <35msEdge-to-core RTT
- per‑secondElastic billing
Elastic billing profiles
Mix on-demand instances with reserved pools and edge bursts.
Reserved
Committed pools
Lock bandwidth and hosts for steady workloads with lower rates.
- ✓Blend dedicated and shared pools
- ✓Plan capacity windows ahead of events
- ✓Named engineer with 30 min SLA
Edge burst
POP expansion
Schedule seasonal capacity near campaigns or regional launches.
- ✓Combine edge POP and core DC compute
- ✓Anycast IP announcement per VM
- ✓Always-on L3-L7 protection included
Global facilities overview
Tier-3+ data centers interconnected with AS211051 backbone.
Asia Pacific
- Singapore · Equinix SG3
- Tokyo · Koto-ku
- Seoul · Pangyo
- Sydney · Alexandria
- Mumbai · Navi Mumbai
Europe / Middle East / Africa
- Frankfurt · Campus Kleyer
- London · Docklands
- Paris · La Courneuve
- Dubai · Silicon Oasis
- Johannesburg · Isando
Americas
- Seattle · Westin Building
- Silicon Valley · Santa Clara
- Dallas · Infomart
- São Paulo · Hortolândia
- Mexico City · Querétaro
Platform highlights
Engineered for low-latency, regulated workloads and burst campaigns.
Unified fabric
x86 and ARM nodes with SR-IOV networking, NUMA pinning and dedicated GPUs on request.
Programmable networking
Per-VM Anycast, dual-stack IPv4/IPv6 with managed NAT translation, BYOIP, private L3 transit and global load balancing.
Built-in security
Confidential compute options, TPM attestation and managed patches.
Deep observability
Second-level metrics, streaming logs and OpenTelemetry exporters.
Deploy anywhere, only pay for runtime
Blend dedicated and burst capacity without rearchitecting.