Neocloud Rankings

Neocloud GPU fleets and hardware

28 neoclouds disclose GPU counts or hardware models; 9 have a verified count. Core42 (76,000), Nscale (50,000), Lightning AI (36,000), GMI Cloud (30,000) state the most. Fleet scope varies: a deployed cluster, a company fleet and contracted GPUs are labelled in each note. Partner supply and maximum cluster sizes are excluded from counts. GPUs carry 10% of the score.

28 of 51 neocloudsChecked

GPU modelsOwnership
Core4276,0001—104Participation Ribbon—subsidiary2026-11-11
Nscale50,0002—46Unavailable1,3701private2026-11-11
Lightning AI36,0001—143Unavailable—private2026-11-11
GMI Cloud30,0004NVIDIA H100, H200 and Blackwell (Blackwell shown as pre-order on the home page).86Silver—private2026-11-11
Taiga Cloud24,0001—402——subsidiary2026-11-11
Vast.ai17,0001—233Participation Ribbon—private2026-11-11
TensorWave8,1923—55Silver2,0003private2026-11-11
Shakti Cloud8,0001—383Unavailable—subsidiary2026-11-11
Scaleway5,0001—303Unavailable—subsidiary2026-11-11
Bitdeer AI—NVIDIA GB200 NVL72 and GB300 NVL72.323Participation Ribbon——2026-11-11
Boost Run—NVIDIA B300, B200, H200, H100 and A100 (listed as available now).463Unavailable—public2026-11-11
BUZZ HPC—NVIDIA GB200 NVL72, B200, H200 and H100.323Participation Ribbon——2026-11-11
Cirrascale—NVIDIA GB300, B300, B200, RTX PRO 6000; AMD MI355X and MI350; Qualcomm AI 100 Ultra464Unavailable——2026-11-11
CUDO Compute—NVIDIA H100 and H200 deployed in the company case study; newer architectures advertised separately.463Unavailable—private2026-11-11
Denvr Dataworks—NVIDIA H200, H100, A100; Intel Gaudi2463Unavailable——2026-11-11
Gcore—NVIDIA L40S, A100, H100, H200, B300 and GB300 (B300/GB300 reservable).173Bronze——2026-11-11
GMO GPU Cloud—NVIDIA HGX B300 and H200.173Bronze——2026-11-11
Hot Aisle—AMD Instinct MI300X (eight GPUs per Dell PowerEdge XE9680 server).464———2026-11-11
Hydra Host—NVIDIA H200 and GH200 listed in the product catalog.413Underperforming——2026-11-11
Mithril—NVIDIA H100 and H200.413Underperforming——2026-11-11
Moonlite—H100 tested in ClusterMAX 3.0; report also describes initial B300 and GB300 handovers.173Bronze——2026-11-11
OVHcloud—NVIDIA H200, H100, A100, L40S and L4 (Public Cloud catalog).254Underperforming—public2026-11-11
Prime Intellect—NVIDIA H100, H200 and B300 listed for compute.173Bronze——2026-11-11
QumulusAI—NVIDIA B300, B200, RTX PRO 6000, H200, H100 SXM and H100 NVL.463Unavailable——2026-11-11
SHARON AI—NVIDIA H200, H100, L40S and A40; H200/B200 supercluster described separately.413Underperforming——2026-11-11
STN—B200, B300 and GB300 listed; Vera Rubin also advertised.283Participation Ribbon——2026-11-11
TensorDock—NVIDIA H100 SXM5, A100 SXM4 and RTX 4090 (among the models listed).452———2026-11-11
VESSL—NVIDIA B300, B200, GB300, H200, H100, A100 and L40S (newer models reservable).323Participation Ribbon——2026-11-11

Every neocloud on this list

Full ranking
  • #4 Nscale 50,000 GPUs · score 53.1
  • #5 TensorWave 8,192 GPUs · score 50.7
  • #8 GMI Cloud 30,000 GPUs · score 45.6
  • #10 Core42 76,000 GPUs · score 37.7
  • #14 Lightning AI 36,000 GPUs · score 20.4
  • #17 Gcore NVIDIA L40S, A100, H100, H200, B300 and GB300 (B300/GB300 reservable). · score 16.5
  • #17 GMO GPU Cloud NVIDIA HGX B300 and H200. · score 16.5
  • #17 Moonlite H100 tested in ClusterMAX 3.0; report also describes initial B300 and GB300 handovers. · score 16.5
  • #17 Prime Intellect NVIDIA H100, H200 and B300 listed for compute. · score 16.5
  • #23 Vast.ai 17,000 GPUs · score 16.1
  • #25 OVHcloud NVIDIA H200, H100, A100, L40S and L4 (Public Cloud catalog). · score 12.5
  • #28 STN B200, B300 and GB300 listed; Vera Rubin also advertised. · score 11.5
  • #30 Scaleway 5,000 GPUs · score 11
  • #32 Bitdeer AI NVIDIA GB200 NVL72 and GB300 NVL72. · score 10.5
  • #32 BUZZ HPC NVIDIA GB200 NVL72, B200, H200 and H100. · score 10.5
  • #32 VESSL NVIDIA B300, B200, GB300, H200, H100, A100 and L40S (newer models reservable). · score 10.5
  • #38 Shakti Cloud 8,000 GPUs · score 7.4
  • #40 Taiga Cloud 24,000 GPUs · score 6.6
  • #41 Hydra Host NVIDIA H200 and GH200 listed in the product catalog. · score 4.5
  • #41 Mithril NVIDIA H100 and H200. · score 4.5
  • #41 SHARON AI NVIDIA H200, H100, L40S and A40; H200/B200 supercluster described separately. · score 4.5
  • #45 TensorDock NVIDIA H100 SXM5, A100 SXM4 and RTX 4090 (among the models listed). · score 3
  • #46 Boost Run NVIDIA B300, B200, H200, H100 and A100 (listed as available now). · score 0
  • #46 Cirrascale NVIDIA GB300, B300, B200, RTX PRO 6000; AMD MI355X and MI350; Qualcomm AI 100 Ultra · score 0
  • #46 CUDO Compute NVIDIA H100 and H200 deployed in the company case study; newer architectures advertised separately. · score 0
  • #46 Denvr Dataworks NVIDIA H200, H100, A100; Intel Gaudi2 · score 0
  • #46 Hot Aisle AMD Instinct MI300X (eight GPUs per Dell PowerEdge XE9680 server). · score 0
  • #46 QumulusAI NVIDIA B300, B200, RTX PRO 6000, H200, H100 SXM and H100 NVL. · score 0

How to read the columns

Methodology
GPUs
GPUs deployed, owned or under contract as the company states it.
GPU models
Hardware advertised or tested, with reservations and future availability explicitly noted. Context only; not a fleet count or scoring input.
Rank
Position by composite score, computed at build time from the signals below with the weights on /methodology/. Ties share the higher position. Never set by hand and never for sale.
ClusterMAX
Tier in SemiAnalysis ClusterMAX 3.0 (23 Sep 2026), the public hands-on rating of GPU clouds. Medal tiers run Platinum, Gold, Silver, Bronze; Participation Ribbon sits between Bronze and Underperforming. Underperforming and Unavailable are the two groups SemiAnalysis lists under "Not Recommended": Unavailable covers providers it could not test or judged too small to rate. Empty when the provider is not listed.
Contracted power MW
Power the company states as contracted, secured or committed in its portfolio, in MW. Speculative pipeline and planned campuses are excluded; the power note quotes the company's own word for it.
Ownership
Publicly listed, privately held, or a subsidiary of another company.