Voltage Park’s current site announces its merger with Lightning AI. Confirm the applicable product, account and order terms; the site continues to advertise on-demand and reserved infrastructure. Read the source ↗

GPU clouds / Voltage Park GPU Infrastructure alternative

A Voltage Park alternative for the agent around GPU infrastructure

Voltage Park currently announces its merger with Lightning AI and continues to present on-demand and reserved GPU infrastructure. Its On Demand lifecycle separates ordinary shutdown from VM-only compute deallocation. Retained whole-system data and the action that actually ends compute charges need separate checks.

Start buildingCompare capabilities ↓

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

When Voltage Park GPU Infrastructure fits

Voltage Park fits teams needing GPU VMs or bare metal, fast interconnects and managed cluster/storage options. Existing instances retain data on shutdown, and eligible virtualized instances can release compute while keeping storage.

When to consider Sandbox0

Consider Sandbox0 for compatible Linux agents around GPU work that need a complete durable writable RootFS, automatic online growth and supported files while guest compute stays paused. Keep required accelerator capacity separately.

Capabilities

Voltage Park GPU Infrastructure and Sandbox0: what persists?

Feature comparison with documentation and verification status
Environment and runtime lifetimeVoltage Park GPU InfrastructureNot 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 filesVoltage Park GPU InfrastructureShutdown retains data; eligible VM deallocation retains disk while releasing compute. Independent shared storage has its own lifecycle.View evidence ↗Sandbox0The writable RootFS survives successful checkpoints, including software, system configuration, dependencies, and user data.Read documentation →
Whole RootFSVoltage Park GPU InfrastructureThe retained VM/instance data includes its OS environment and files; termination and separate mounts have different preservation policies.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 stateVoltage Park GPU InfrastructureShutdown/start retains disk at full rate; virtualized-instance deallocation provides storage-only billing and can relocate the VM.View evidence ↗Sandbox0Pause compute between tasks; resume on available capacity. Default resume starts fresh processes from the saved filesystem.Read documentation →
Memory and process stateVoltage Park GPU InfrastructureReleasing GPU, CPU and RAM with retained storage does not establish saved process or GPU-memory continuation.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.

Voltage Park offers independent storage and VM/bare-metal resource choices. The reviewed lifecycle does not establish automatic online growth of the complete writable agent RootFS.

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 →

Ordinary shutdown retains data and the full running rate

The instance guide says Shut Down preserves instance data but continues charging the full running rate. Terminate removes the instance and associated data. A shutdown action therefore cannot be used as evidence of compute-cost savings, even though it preserves the operating system and files.

Source 1 ↗

VM-only deallocation releases compute with retained storage

The separate Deallocate compute resources action releases GPU, CPU and RAM and leaves storage-only billing, exclusively for virtualized instances. Restart can relocate the saved VM to another host and change its external address or ports. This differs from suspending process/GPU memory in place.

Source 1 ↗

Current pricing combines an entry reference with configured quotes

The pricing FAQ advertises one H100 from $1.99 per hour, while Ethernet, InfiniBand and reserve cards request quotes. The homepage describes dedicated reserve terms of 12+ months while the pricing card says 6+ months; confirm the effective order rather than assuming one universal minimum.

Source 1 ↗Source 2 ↗

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.

Voltage Park GPU Infrastructure

Pricing reviewed 2026-10-19

Advertised H100 entry reference
From $1.99 per GPU-hour
Confirm the selected VM/bare-metal configuration, interconnect and availability.
Rate source ↗
Configured Ethernet or InfiniBand capacity
Current configuration-specific quote
Rate source ↗
Ordinary stopped instance
Full running billing rate continues
Rate source ↗
Eligible deallocated VM
Storage-only billing
Compute deallocation is exclusive to virtualized instances.
Rate source ↗

The public entry rate is not a universal multi-GPU or interconnect quote. Confirm storage, reservation duration, payment terms and the current merged-company agreement. Deallocation and shutdown have different charges; bare-metal eligibility must not be inferred from VM-only functionality.

Check current Voltage Park GPU Infrastructure 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 Voltage Park GPU Infrastructure

  1. Record VM versus bare-metal deployment, interconnect, storage, current agreement and shutdown versus deallocation behavior.
  2. Back up required state before termination and retain suitable GPU/shared-storage services.
  3. Port compatible Linux agent execution and test complete-system retention, growth, paused files and storage-only restart behavior.

Common questions

Is Sandbox0 a drop-in Voltage Park GPU Infrastructure 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 ↗support.voltagepark.com/article/my-instances ↗voltagepark.com/pricing ↗voltagepark.com/ ↗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 →