LogoLogo
Foundry Documentation
Foundry Documentation
  • Welcome to Foundry
  • Quickstart guide
  • Compute & Storage
    • Compute overview
    • Instance types & specifications
    • Reserving compute
    • Spot bids
      • Spot auction mechanics
    • Startup scripts
    • Access & manage instances
      • Statuses
    • Compute quotas
    • Managing open ports
    • Persistent storage
      • File shares
      • Block storage
    • Ephemeral storage
  • Foundry API
    • API overview and quickstart
    • API reference
      • Projects
      • Instance types
      • SSH Keys
      • Volumes
      • Instances
      • Spot
        • Bids
        • Availability
      • API Keys
      • Profile
    • Specification
  • Access Management
    • Access Management Overview
    • SSH keys
  • Account and Billing
    • Billing overview
    • Foundry Referral Program
  • Security & trust
    • Foundry's approach to security
    • Reporting security concerns
Powered by GitBook
LogoLogo

© 2025 Foundry Technologies, Inc.

On this page
  • Base image
  • Hardware specifications
  1. Compute & Storage

Instance types & specifications

PreviousCompute overviewNextReserving compute

Last updated 16 days ago

Foundry offers several instances of various NVIDIA GPU types fit for your workloads. We will continue to add instance types over time.

Depending on your instance type, Ethernet or InfiniBand interconnect may be available.

Base image

All Foundry instances are provisioned with the following by default:

  • Ubuntu 22.04

  • NVIDIA driver version: 535

  • Cuda version: 12.2

You can use to customize your environment during the startup process of your instances.

Hardware specifications

Each instance comes with its own ephemeral storage (outlined below), usually with NVMe SSD. This storage will exist until the instance has been terminated or pre-empted. There is no cost to use the ephemeral storage. Ephemeral storage is automatically mounted for you (and re-mounted after preemptions) at /mnt/local.

Instance name
GPU type
GPU memory (GB)
GPU quantity
# CPU cores
RAM (GB)
Ephemeral Storage (GB)

1x A100 80GB SXM

A100 SXM

80

1

24

250

1,750

2x A100 80GB SXM

A100 SXM

80

2

48

500

3,500

4x A100 80GB SXM

A100 SXM

80

4

96

1,000

7,000

8x A100 80GB SXM

A100 SXM

80

8

192

2,000

14,000

8x H100 80GB SXM

H100 SXM

80

8

168

2,048

12,500

1x A40 48GB PCIe

A40 PCIe

48

1

12

62

862

2x A40 48GB PCIe

A40 PCIe

48

2

24

124

1,725

4x A40 48GB PCIe

A40 PCIe

48

4

48

248

3,450

8x A40 48GB PCIe

A40 PCIe

48

8

96

496

6,900

1x A5000 24GB PCIe

A5000 PCIe

24

1

12

62

862

2x A5000 24GB PCIe

A5000 PCIe

24

2

24

124

1,725

4x A5000 24GB PCIe

A5000 PCIe

24

4

48

248

3,450

8x A5000 24GB PCIe

A5000 PCIe

24

8

96

496

6,900

startup scripts