Best NeocloudsAn independent guide to GPU clouds

Compare GPU clouds / DigitalOcean

DigitalOcean

General-purpose cloud with NVIDIA and AMD GPU Droplets (NYSE: DOCN); includes Paperspace

Mid-size cloudAMD GPUsInference

DigitalOcean sells self-serve NVIDIA and AMD Instinct GPU Droplets billed per second, plus 8-GPU bare-metal servers through sales; Paperspace is now part of DigitalOcean. Its pricing (effective 2026-08-01) includes H100 $4.41, H200 $4.47, MI300X $2.59 and MI325X $3.80 per GPU-hour on demand, with 12-month rates such as MI300X $1.91 and H100 $3.26. It is rated Bronze in SemiAnalysis ClusterMAX 3.0.

Best for: Self-serve AMD MI300X and MI325X, and teams already on DigitalOcean.

ClusterMAX 3.0: Bronze

  • ✓ Publishes prices
  • ✓ Serverless inference
  • ✓ AMD GPUs
  • – European data centers
Visit DigitalOcean

Facts checked November 9, 2026 against the sources listed below. DigitalOcean does not pay to appear here.

1 Legal nameDigitalOcean
2 HeadquartersBroomfield, Colorado, USA
3 Founded2012
4 Funding or listingPublic (NYSE: DOCN). Up to $1.0B at-the-market equity program (Nov 2026); MUFG equipment facility of up to $1.025B (Sep 2026).
5 How you rentOn-demand, Reserved clusters, Bare metal, Serverless inference
6 Minimum commitmentNone for on-demand (per-second billing, 5-minute minimum); reserved prices need a 12-month contract
7 InterconnectNot disclosed
8 OrchestrationKubernetes (DOKS)
9 GPUs offeredB300, H200, H100 SXM, MI355X, MI350X, MI325X, MI300X, L40S, RTX 6000 Ada, RTX 4000 Ada
10 RegionsUS (New York NYC1 and NYC2, Atlanta ATL1, Richmond RIC1), Canada (Toronto TOR1), Netherlands (Amsterdam AMS3)
11 CertificationsSOC 2, SOC 3, HIPAA-eligible, Global PRP
12 Customers and partnersFeatures Supermaven as a Bare Metal GPU customer story.

Published prices

On-demand list prices per GPU-hour, as published on November 9, 2026. Reserved and negotiated prices are usually lower.
GPUPriceBasisSource
RTX 4000 Ada$0.76on-demand GPU Droplet (1x), billed per secondPricing page
L40S$1.57on-demand GPU Droplet (1x), billed per secondPricing page
RTX 6000 Ada$1.57on-demand GPU Droplet (1x), billed per secondPricing page
MI300X$2.59on-demand GPU Droplet (1x or 8x), billed per secondPricing page
MI325X$3.80on-demand GPU Droplet (1x or 8x), billed per secondPricing page
H100 SXM$4.41on-demand GPU Droplet (1x or 8x), billed per secondPricing page
H200$4.47on-demand GPU Droplet (1x or 8x), billed per secondPricing page

Where we recommend them

Worth knowing before you sign

  • A general-purpose cloud rather than GPU-first; SemiAnalysis found its Kubernetes solid but immature, and its first XID error injection went undetected.
  • No interconnect is stated for GPU Droplets. B300 is spot or 12-month reserved only ($7.94 reserved), and bare-metal GPUs have no published price.
  • Powered-off Droplets are still billed; spot Droplets can be reclaimed with about 2 hours’ notice or less.
  • Paperspace’s own site now shows only Gradient plans and points GPU users to DigitalOcean.

Go to DigitalOcean's site

Sources

  1. DigitalOcean GPU Droplet pricing
  2. DigitalOcean GPU Droplets
  3. DigitalOcean Bare Metal GPUs
  4. DigitalOcean regional availability
  5. DigitalOcean trust
  6. Paperspace (now part of DigitalOcean)
  7. SEC: $1.0B ATM prospectus
  8. SEC: MUFG equipment facility 8-K
  9. SemiAnalysis ClusterMAX 3.0 ratings
  10. SemiAnalysis ClusterMAX 3.0 report

Work at DigitalOcean? Send corrections with a source to editors@bestneoclouds.com. We fix verified errors within five business days.

Compare with

  • CoreWeave GPU cloud built around large InfiniBand training clusters (NASDAQ: CRWV)
  • Nebius Full-stack AI cloud with published on-demand GPU prices (NASDAQ: NBIS)
  • Crusoe AI cloud with self-serve NVIDIA and AMD GPUs on its own data center campuses
  • Nscale European AI cloud selling GB300 and B200 capacity (S-1 filed for NYSE: NSCL)
  • Together AI AI cloud for GPU clusters and model inference