Compare all 6 plans
Every plan runs the identical NVMe-backed platform in both regions - the only thing that changes size to size is how much of it you get. Surface is the base plan every comparison below is measured against.
Pick up to 4 plans to line up side by side - use the dropdowns in the table header. Choosing a plan that's already shown swaps it with the column you picked from.
| Base plan | Most deployed | Largest | ||
|---|---|---|---|---|
| Price | $6 /mo | $12 /mo | $24 /mo | $192 /mo |
| vCPUs | 1 | 2 | 4 | 32 |
| Memory | 2 GB | 4 GB | 8 GB | 64 GB |
| NVMe storage | 40 GB | 80 GB | 160 GB | 1280 GB |
| Transfer / mo | 2 TB | 4 TB | 6 TB | 12 TB |
| Vs. Surface | - | 2× vCPU / 2× RAM | 4× vCPU / 4× RAM | 32× vCPU / 32× RAM |
Which plan actually fits
A bigger plan is never strictly better - it's more capacity you're paying for whether you use it or not. Here's what each step up actually buys you, and what it costs you.
Surface
$6/moPersonal projects, static sites, dev/test sandboxes.
- Benefit
- The cheapest way onto the same platform as every other plan - full root access, no feature gate.
- Trade-off
- One vCPU can bottleneck under concurrent load, and 40GB won't hold much beyond an OS and a small app.
Cruise
$12/moSmall production apps and side projects with real traffic.
- Benefit
- Double the compute and storage of Surface for double the price - the most balanced option on the lineup.
- Trade-off
- Still tight if you're running more than one meaningful service on the box at once.
Stratosphere
$24/moSmall teams running a few services on one box.
- Benefit
- Four vCPUs handle concurrent requests and background jobs without fighting each other for CPU time.
- Trade-off
- Paying for headroom you may not touch if traffic is genuinely light.
Mesosphere
$48/moDatabases, CI runners, and multi-container stacks.
- Benefit
- 16GB of RAM keeps a working dataset or a handful of containers in memory instead of swapping.
- Trade-off
- Overkill if what you actually need is a single small service - Stratosphere covers that for half the price.
Thermosphere
$96/moHigher-traffic APIs and larger Postgres or Redis instances.
- Benefit
- 16 vCPUs and 32GB of RAM handle real concurrency - the kind of load that would need a second server on smaller plans.
- Trade-off
- At this size, a managed multi-node setup can be more resilient than a single large VM - worth comparing before you commit.
Exosphere
$192/moBatch processing, ML inference, and monorepo builds.
- Benefit
- The full lineup's ceiling - top-of-range specs without leaving the simplicity of a single VM.
- Trade-off
- The highest blast radius on the lineup: it's one machine, so losing it affects the most work.
Same 6 plans, either region
Pricing and specs are identical in San Francisco and Frankfurt - pick whichever sits closer to your users. Toggle below for a rough sense of network latency from each region.
Both regions sit on the same private backbone and run the same 6 plans at the same prices - there's no regional markup anywhere on the lineup. The only real difference is which part of the world you're closest to.
Most people start with Cruise.
You can resize up or down later without rebuilding the instance.