GPU clouds / Scaleway GPU Instances alternative

A Scaleway GPU alternative for the Linux agent around GPU work

Scaleway GPU Instances combine accelerators, cloud networking and selectable storage. A Block Storage boot volume can already retain the operating system and installed software. GPU shape, availability zone and the exact power operation determine how that retained state and its bill behave.

Start buildingCompare capabilities ↓

Reviewed 2026-10-19 · Published by the Sandbox0 team

When Scaleway GPU Instances fits

Scaleway fits workloads needing GPU instances and European cloud services, networking and storage. Its durable boot volumes provide real whole-system disk retention; keep accelerator work and cloud dependencies where required.

When to consider Sandbox0

Consider Sandbox0 for compatible Linux agents around GPU workloads, with a complete durable writable RootFS, automatic online growth and supported paused-file operations. GPU execution and required cloud services remain separate.

Capabilities

Scaleway GPU Instances and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeScaleway GPU InstancesNot established in this reviewSandbox0ttl and hard_ttl default to 0 (disabled). Configure idle pause and deletion policies for the same persistent sandbox identity; an idle running sandbox still accrues compute charges.Read documentation →
Persistent filesScaleway GPU InstancesBlock Storage boot/data volumes retain data independently of powered-off compute; scratch requires its own lifecycle check.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSScaleway GPU InstancesThe OS can boot from retained Block Storage, preserving installed software and system configuration outside separate runtime/scratch mounts.View evidence ↗Sandbox0Persistence extends beyond /workspace to paths such as /usr/local, /etc, and /home. Runtime tmpfs and virtual filesystems are excluded.Read documentation →
Release compute, keep stateScaleway GPU InstancesPower off/on provides disk continuity with released host capacity and continuing retained-resource charges; Standby is a different billed state.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateScaleway GPU InstancesThe power-off guide does not establish durable GPU, RAM or running-process checkpoint restoration.View evidence ↗Sandbox0Experimental memory checkpoints can preserve compatible process state. Use memory: true on both pause and resume on a compatible deployment. A filesystem checkpoint alone does not preserve running processes.Read documentation →

Online RootFS growth

On growth-enabled images, Sandbox0 automatically expands the writable RootFS as free space runs low, within a configured sparse logical ceiling up to 64 TiB. Agent processes keep running. Growth is enabled in the qualified Cloud region; confirm activation for other regions. Existing images and snapshots retain their captured storage geometry. The ceiling is logical capacity, not preallocated host disk.

Scaleway already offers explicit Block Storage enlargement and instance resizing. These operations differ from automatically extending an enabled writable agent RootFS as free space runs low; check boot-volume and filesystem steps for the selected GPU shape.

Files while compute is paused

Sandbox0's Files API reads persistent files without resuming compute. Filesystem-only pauses also permit edits; retained-memory pauses are read-only and reject mutations with 409. SSH, SFTP, file watching, and runtime-only mounts require running compute.

See paused file operations →

Retained boot storage and scratch have separate roles

Scaleway supports booting the operating system from Block Storage, while selected GPU shapes also list Scratch NVMe. Preserve indispensable state on storage with the required lifecycle and verify scratch behavior for the chosen instance. Do not equate a listed scratch device with the durable boot volume.

Source 1 ↗Source 2 ↗Source 3 ↗

Power off and Standby have different charges

Power off releases the physical host; the guide archives applicable local volumes and retains independent Block Storage. Compute charges stop, while allocated storage and IP resources continue billing. Standby keeps the instance allocated and is charged like a running instance. Disk continuity alone does not establish a saved GPU or process-memory resume.

Source 1 ↗Source 2 ↗Source 3 ↗

An availability-zone selector changes the GPU catalog

The selected PAR-2 price table includes L4, L40S, H100 and B300 shapes. PAR-1 can show some families unavailable, which is not global product unavailability. Match the accelerator, vCPU/RAM shape and zone before comparing hourly references.

Source 1 ↗

Pricing

Compare billing models, then price your workload

CPU allocation, memory, billing minimums, disk, plan fees, and supported workload differ. These meters are not equal-performance machine quotes. Confirm region, resource units, and included allowances before estimating a total.

Scaleway GPU Instances

Pricing reviewed 2026-10-19

PAR-2 L4-1-24G
€0.79 per instance-hour, excluding VAT
1 L4, 8 vCPUs and 48 GB RAM.
Rate source ↗
PAR-2 L40S-1-48G
€1.47 per instance-hour, excluding VAT
1 L40S, 8 vCPUs and 96 GB RAM.
Rate source ↗
PAR-2 H100-1-80G
€2.87 per instance-hour, excluding VAT
1 H100, 24 vCPUs and 240 GB RAM.
Rate source ↗
PAR-2 H100-SXM-2-80G
€6.62 per instance-hour, excluding VAT
2 GPUs, 32 vCPUs and 240 GB RAM; not a per-GPU quote.
Rate source ↗

GPU uptime is billed per minute including startup and Standby; RENDER-S is an hourly exception. Retained storage, public IPs and applicable network services cost extra. Rates use the selected PAR-2 configuration and EUR before VAT; regional availability and capacity can change.

Check current Scaleway GPU Instances pricing ↗

Sandbox0

$0.015/GiB-hour compute (billed by the second at approximately $0.000004167/GiB-second), plus $0.02/GiB-month of retained incremental RootFS and memory checkpoint data. Shared base images and platform recovery retention are excluded; paused sandboxes accrue storage charges only.

Compute / GiB-hour
$0.015
Retained data / GiB-month
$0.02
Paused compute
$0.00

A 4 GiB agent, 100 running hours

$6.00 compute

4 GiB × 100 hours, billed by the second. Keeping 20 GiB of billable incremental data for a full billing month adds $0.40. CPU is derived from memory by the platform; choose a tier that fits your workload.

Understand storage billing →

Before migrating from Scaleway GPU Instances

  1. Record GPU shape, zone, Block Storage boot/data volumes, scratch paths, IPs and lifecycle billing.
  2. Back up required state with the chosen storage policy and keep GPU/cloud services that remain necessary.
  3. Port compatible Linux agents and test system retention, online growth, supported paused files and complete retained-resource costs.

Common questions

Is Sandbox0 a drop-in Scaleway GPU Instances replacement?

Evaluate API compatibility and product scope. These products serve different core workflows, so a change may require a different architecture.

Does keeping files also keep a running agent?

Filesystem state and volatile process state are separate. Sandbox0 requires explicit compatible memory checkpoints to preserve running processes; its default filesystem resume starts fresh processes.

Sources and review scope

The Boat team's Battleships directory is the discovery source. References below describe the product and plan discussed on this page. Check current availability, regional terms and billing units before migrating. Where the cited material does not establish a capability or price, the comparison identifies that uncertainty.

Battleships research record ↗www.scaleway.com/en/docs/instances/faq/ ↗www.scaleway.com/en/docs/gpu/quickstart/ ↗www.scaleway.com/en/pricing/gpu/ ↗www.scaleway.com/en/docs/instances/how-to/power-off-instance/ ↗www.scaleway.com/en/docs/instances/reference-content/understanding-instance-pricing/ ↗Sandbox0 RootFS capacity →

Related comparisons

Azure GPU VMs (NC_A100_v4 reference) vs Sandbox0 →Cerebrium Serverless Compute vs Sandbox0 →Crusoe vs Sandbox0 →DigitalOcean GPU Droplets vs Sandbox0 →Fluidstack vs Sandbox0 →Google Compute Engine GPU VMs vs Sandbox0 →