Boot a VM in seconds.
Pay per second after that.
Basic, general-purpose, memory-optimized, and CPU-optimized instances. Per-second billing, free intra-region traffic, the same image library in every region. No commitments to get the listed price.
Configure the VM. See the bill, line by line.
Compute is billed per second. Block storage, snapshots, and the reserved IP show up as separate lines — no bundle, no surprise add-ons after launch. Pricing examples below are Virginia (US-East). Use the region picker for Los Angeles, Frankfurt (+8%), or Singapore (+18%).
Four families, one price list.
Linux. Per-second billing. Virginia (us-east-1) prices in USD.
Shared cores. Ideal for dev, side projects, low-traffic apps.
| vCPU | RAM | Bandwidth | On-demand | 1-year | 3-year |
|---|---|---|---|---|---|
| 1 | 2 GB | 50 Mbps | $0.010/hr | $4.90/mo | $3.50/mo |
| 2 | 4 GB | 50 Mbps | $0.021/hr | $9.80/mo | $7.00/mo |
| 4 | 8 GB | 100 Mbps | $0.042/hr | $19.60/mo | $14.00/mo |
| 8 | 16 GB | 100 Mbps | $0.083/hr | $39.20/mo | $28.00/mo |
| 16 | 32 GB | 100 Mbps | $0.167/hr | $78.40/mo | $56.00/mo |
Monthly price assumes 672 hours (a 28-day cap — sustaining a VM all month never bills more than that, even in March). Basic uses shared cores; general / memory / CPU-optimized run on dedicated threads. Prices shown reflect US East (Ashburn).
Stock images, app stacks, or your own ISO.
Ubuntu, Debian, Rocky, AlmaLinux, Fedora, CentOS Stream, openSUSE, FreeBSD, Windows Server. Patched on the vendor’s schedule.
Pre-built images for n8n, Plausible, Mattermost, GitLab, NocoDB, OpenWebUI, and more. SSH keys injected at boot.
Upload raw / qcow2 / vmdk / VHD via the dashboard or `nimbus image push`. Or build with Packer. Images are private by default.
Operations without ticking a box.
NVMe SSD volumes at $0.11/GB·mo, 10K IOPS, resize live. Attach multiple per VM.
Incremental at $0.04/GB·mo. Cross-region copy. Retain 1–365 days.
S3-compatible at $0.013/GB·mo. Free transfer to your VM in the same region. 11 nines of durability.
Archive tier at $0.004/GB·mo. Retrieval $0.02/GB, restore in under 12 hours. 90-day minimum.
Static IPv4 at $2.50/mo, transferable between VMs. IPv6 is free, always.
L4 and L7. From $9/mo per LB plus traffic. 5 Gbps capacity per LB.
IPsec at $22/mo per gateway. 100 GB of traffic included.
Compute ↔ storage in the same region is always free. First 1 TB of internet egress per VM is free; $0.009/GB after.
$0.005/GB in both directions. The same flat rate whether you cross continents or stay on the same coast.
Postgres, MySQL, Redis, Kafka. Billed as the underlying compute plus a 25–35% managed surcharge. HA optional.
Control plane is free. You only pay for the worker-node VMs. CNI-compatible, CSI for block and object.
$0.05/GB stored. Free pulls in-region. Image vulnerability scans on push.
What teams build on NimbusNexus VMs.
Rails / Django / Next.js behind a load balancer in US-East. The default for B2B SaaS, internal tools, anything serving a North-American customer base.
Spin up cpu-* for ~6 hours during a deploy window, tear down at night. Per-second billing means you pay $0 the other 18 hours. Co-located with your prod artifact registry.
mem-* nodes give 8 GB RAM per vCPU — exactly what an in-memory store wants. Deployed in Frankfurt for GDPR data-residency; replica per AZ. Frankfurt prices include the +8% EU multiplier.
cpu-* high clock for 60 fps. Dedicated threads mean no neighbor stealing CPU during the boss fight or live stream. LA keeps West-Coast players & creators inside ~30 ms RTT.
Freight tracking, order routing, and customs APIs that need low latency to ports across SEA. Singapore is in the coming-soon tier — pricing already shows the +18% APAC multiplier so what you see is what you pay at launch.
4 vCPU, 16 GB, 100 GB disk. Three invoices.
One always-on production web server. Virginia (US-East). Source: published price lists, May 2026.
| Line item | NimbusNexus gp-4-16 | AWS m7a.xlarge | GCP n2d-std-4 | Azure D4as v5 |
|---|---|---|---|---|
| 4 vCPU + 16 GB | $72.00 | $143.81 | $132.86 | $140.55 |
| 100 GB SSD block | $11.00 | $17.20 | $17.00 | $19.20 |
| Static IPv4 | $2.50 | $3.65 | $2.92 | $2.92 |
| 1 TB outbound traffic | $0.00 | $88.20 | $85.00 | $82.00 |
| Daily snapshots (30 GB) | $1.20 | $1.50 | $1.30 | $1.50 |
| Basic monitoring | $0.00 | $2.10 | $2.50 | $2.30 |
| Total / month | $86.70 | $256.46 | $241.58 | $248.47 |
The line that does the work is outbound traffic. AWS / GCP / Azure all bill $0.08–$0.09 per GB beyond a tiny free tier. NimbusNexus includes 1 TB per VM, then $0.009/GB.
Click, command, or commit it.
$ nimbus vm create my-api \--size gp-4-16 \--image ubuntu-24.04 \--region us-east-1 \--block 100 \--snapshots daily:7→ Provisioning...✓ 203.0.113.45 ssh -i ~/.ssh/id_ed25519
resource "nimbus_vm" "api" {name = "my-api"size = "gp-4-16"image = "ubuntu-24.04"region = "us-east-1"block_volume {size_gb = 100}}
import * as nimbus from "@nimbus/pulumi";new nimbus.Vm("api", {size: "gp-4-16",image: "ubuntu-24.04",region: "us-east-1",block: { sizeGb: 100 },snapshots: { schedule: "daily", retainDays: 7 },});
Virtual machine questions.
See it built end-to-end.
Hypothetical workloads, concrete sizes, the failure modes worth knowing about, and the cost math at every layer. No fictional companies — just the architecture we would recommend if you walked in and asked.
What basic, general-purpose, memory-optimized, and CPU-optimized instances are actually for, the symptoms that mean you picked wrong, and a 5-minute resize plan when you did.
Read the playbook →PLAYBOOK · LOS ANGELES60 fps under sustained load: dedicated threads on cpu-* sizes, p99 RTT budgets per region, the matchmaking layer in Redis, and the launch-weekend spend pattern that scales linearly with concurrent players.
Read the playbook →PLAYBOOK · FRANKFURTValidator and full-node fleets on memory-optimized VMs with big block volumes, the snapshot rhythm that survives a chain reorg, and the GDPR / BaFin posture that keeps a treasury desk inside the EU regulatory perimeter.
Read the playbook →Move a running fleet in a weekend. No exit fees on the way in.
Bring AMIs, qcow2s, or live VMs via WireGuard sync. Our tool maps EC2 / GCE / Azure instance types to the closest NimbusNexus size, calculates the new bill, and cuts over with minimal downtime per VM.