Verified answer
Which neoclouds are NVIDIA Cloud Partners?
Of the neoclouds DEPLOY tracks, 88 build on NVIDIA, and 16 carry an explicit NVIDIA partner designation on their record. Here they are, with the tier shown where a source confirms it, and the one clear non-NVIDIA exception named.
NVIDIA-based neoclouds
| AceCloud | Asia | NVIDIA-based (tier unstated) | NVIDIA A30 / L4 / L40S / A100 / H100 / H200; INR billing |
| Alem.Cloud | Asia | NVIDIA-based (tier unstated) | 512 NVIDIA H200 (64 HGX servers, ~2 exaflops) |
| ASI Solutions | Oceania | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 / L40S / RTX PRO 6000; CDC NZ data centers |
| AutoDL | Asia | NVIDIA-based (tier unstated) | NVIDIA A100 / A800 / RTX 3090 / 2080Ti; own multi-node clusters |
| BUZZ HPC | North America | NVIDIA-based (tier unstated) | NVIDIA GB200 NVL72 / HGX B200 / H200 / H100 |
| Capitalonline | Asia | NVIDIA-based (tier unstated) | NVIDIA H20 / L20 (domestic) + overseas training clusters; bare-metal GPU |
| Cassava Technologies | Africa | Africa's first NVIDIA Cloud Partner | NVIDIA GPUs (first ~3,000 in South Africa); ~12,000 more planned across Africa |
| ChronoScale | North America | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 |
| Cirrascale Cloud Services | North America | Google Public Sector collaboration for research-compute acceleration; ecosystem partners include Dell (Platinum), WekaIO, IBM PowerAI, H2O.ai | Nvidia B200, H200, H100, A100, RTX A5000; AMD Instinct MI300X and MI250; Qualcomm Cloud AI; Tenstorrent Galaxy |
| Civo | Europe | NVIDIA-based (tier unstated) | NVIDIA A100 / H100 / H200 / B200; UK/US/EU/India regions |
| Core42 | Middle East | NVIDIA | NVIDIA H100 / H200 (+ AMD, Cerebras WSE-3) |
| Corvex | North America | NVIDIA-based (tier unstated) | NVIDIA H200 / HGX B200 (GB200 NVL72 roadmap); liquid-cooled US Tier III |
| CtrlS | Asia | NVIDIA-based (tier unstated) | Bare-metal NVIDIA GPU clusters (GPU Private Cloud) |
| CUDO Compute | Europe | NVIDIA Preferred Partner | Dedicated NVIDIA GPU clusters; 250+ MW contracted |
| Cyfuture | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 / A100 / L40S / V100; per-second billing |
| Data Vault Pakistan | Asia | Telenor Pakistan | NVIDIA GPUs (H100-class); solar-powered Karachi AI factory |
| Denvr Dataworks | North America | Nvidia Preferred Partner and Intel Partner; ships Intel Gaudi 2 as a listed production SKU (uncommon among neoclouds) | NVIDIA H100 |
| Digi Power X | North America | NVIDIA-based (tier unstated) | NVIDIA B200 / B300 |
| du | Middle East | NVIDIA-based (tier unstated) | NVIDIA GPUaaS; du + NextGenAI supercluster; Tier III UAE DCs |
| E2E Networks | Asia | NVIDIA-based (tier unstated) | NVIDIA B200 / H200 / H100 / A100; 2×1,024-GPU H200 clusters + 1,024-GPU B200 cluster; TIR platform |
| ePLDT | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 / H200; VITRO Sta. Rosa (50 MW AI-ready DC) |
| ESDS | Asia | NVIDIA-based (tier unstated) | NVIDIA B200 / B300 / GB200 / GB300 + AMD Instinct MI300X SuperPODs |
| Fluidstack | North America | Nvidia (primary GPU supplier) and Dell (systems integrator); involved in Anthropic's planned Google TPU deployment. | NVIDIA H100 / H200 / B200 / GB200 |
| FPT Smart Cloud | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 / GB200 (AI Factory) |
| Freedom Holding | Asia | NVIDIA-based (tier unstated) | NVIDIA exascale (Sovereign AI Hub) |
| Gcore | Europe | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 with InfiniBand; bare-metal + VMs |
| GMO GPU Cloud | Asia | NVIDIA-based (tier unstated) | NVIDIA H200 / HGX B300 (Blackwell Ultra); TOP500-ranked cluster |
| GPU Solutions | Europe | NVIDIA-based (tier unstated) | NVIDIA HGX B200 (192 GB HBM3e); InfiniBand NDR; Tier-III Madrid DC |
| GreenNode | Asia | Preferred NVIDIA Cloud Partner | ~1,024 NVIDIA H100; bare-metal across 6 AZs |
| Highreso | Asia | NVIDIA-based (tier unstated) | NVIDIA HGX H200 / B200 (GPUSOROBAN) |
| Highrise AI | North America | NVIDIA-based (tier unstated) | 1,000 NVIDIA H100 (HPE Cray) |
| HUMAIN | Middle East | NVIDIA-based (tier unstated) | NVIDIA GB300 (Grace Blackwell Ultra) |
| Hypertec Cloud | North America | NVIDIA-based (tier unstated) | NVIDIA GPU clusters; own North American DC footprint |
| Impossible Cloud | Europe | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 / B200 / B300; own EU data centers |
| ionstream | North America | NVIDIA-based (tier unstated) | NVIDIA L40S / H200 / B200 + AMD Instinct MI300X; bare metal |
| Jio Platforms | Asia | NVIDIA-based (tier unstated) | NVIDIA GB300 (Blackwell Ultra) |
| KDDI | Asia | NVIDIA-based (tier unstated) | NVIDIA GB200 NVL72; Osaka Sakai AI DC |
| Krutrim | Asia | NVIDIA-based (tier unstated) | NVIDIA GB200 NVL72 / H100 |
| L&T Vyoma | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 / B200 (B300 deploying); own AI DCs |
| Lanyun | Asia | NVIDIA-based (tier unstated) | NVIDIA (RTX 4090, H100/H800) + domestic GPUs; own multi-province clusters |
| Latitude.sh | Latin America | Nvidia: H100, RTX PRO 6000 and HGX B300 configurations listed on its pricing page. | — |
| Lintasarta | Asia | First NVIDIA Cloud Partner in Indonesia | NVIDIA H100 SXM (Deka GPU / GPU Merdeka) |
| Lyceum | Europe | NVIDIA-based (tier unstated) | NVIDIA B200 (B300 imminent); own Berlin/Zurich DCs |
| Massed Compute | North America | Nvidia Preferred Partner; catalog spans A100, L40S, H100, H200, B100, B200, B300, plus RTX Pro Blackwell and RTX 6000 Ada | NVIDIA B200 / B300 / H100 / A100 / A30 |
| Mistral Compute | Europe | NVIDIA-based (tier unstated) | ~18,000 NVIDIA GPUs incl. GB300; own ~44 MW Paris DC |
| Nebius | Europe | Oct 2026: Launch partner offering open-source NVIDIA Run:ai as an app in Nebius AI Cloud Applications on its Managed Kubernetes service. | — |
| Nebul | Europe | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 / B200 / B300; largest Benelux SuperCluster |
| NeevCloud | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 / B200 / B300; HGX with 3,200 Gbps InfiniBand |
| Neutra Compute | Asia | NVIDIA-based (tier unstated) | NVIDIA + AMD accelerators; Cikarage Tier III/IV DC |
| NexGen Cloud | Europe | NVIDIA Elite Cloud Partner | NVIDIA A100 / H100 / H200 + Blackwell GB200 NVL72 / HGX B300; renewable-powered sites |
| Neysa | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 / L40S on-demand; Velocis platform |
| Northern Data | Europe | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 |
| NTT DATA | Asia | NVIDIA-based (tier unstated) | NVIDIA DGX B200 / DGX SuperPOD |
| Nxera | Asia | Announced May 2024 with Nvidia to build a sovereign AI cloud on Nvidia GPUs in Singapore; specific GPU counts not publicly disclosed by Nxera | NVIDIA GPUs (RE:AI) |
| NxtGen | Asia | NVIDIA-based (tier unstated) | NVIDIA A100 / H100 / L40S / T4; bare-metal + vGPU (SpeedCloud AI) |
| Oman Data Park | Middle East | NVIDIA Partner Network Cloud Service Provider | NVIDIA DGX-based GPU/HPC (Nebula AI) |
| Ooredoo | Middle East | NVIDIA Cloud Partner | NVIDIA Hopper (Syntys, Qatar) + H200 (Kuwait); GB300 via Zankore JV |
| Ori | Europe | NVIDIA-based (tier unstated) | NVIDIA H100 / H200; on-demand + private clusters (EdgeConneX, Centersquare) |
| Pi Datacenters | Asia | NVIDIA-based (tier unstated) | NVIDIA A100 / H100; hourly pricing |
| Polarise | Europe | NVIDIA Cloud Partner | NVIDIA HGX B300 (live Oslo); ~10,000-GPU Munich AI Factory |
| ResetData | Oceania | NVIDIA-based (tier unstated) | 1,024 liquid-cooled NVIDIA H200 (AI-F1); on-demand |
| RUTILEA | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 on-demand |
| Sakura Internet | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 / B200 |
| Scaleway | Europe | Nvidia (H100 SXM5 with ConnectX-7 InfiniBand); H100/H200 SKUs listed in PAR-2 pricing catalog | NVIDIA H100 (H200 / GB200 planned) |
| SCCC by stc | Middle East | NVIDIA-based (tier unstated) | NVIDIA L20 / H20 GPU clusters (in-Kingdom) |
| Seeweb | Europe | NVIDIA-based (tier unstated) | NVIDIA H100 / H200 / A100 / L40S + AMD MI300X; own Italian DCs |
| Servernah Cloud | Africa | NVIDIA-based (tier unstated) | NVIDIA GPUs (model not disclosed); iXAfrica NBOX1 DC |
| SF Compute | North America | NVIDIA-based (tier unstated) | NVIDIA H100 clusters; VMs, bare metal, managed Slurm |
| Sharon AI | Oceania | NVIDIA Cloud / Preferred Partner | NVIDIA H200 / H100 NVL + AMD MI300X; scaling to 40,000 GB300 |
| Siam AI | Asia | Thailand's NVIDIA Cloud Partner | NVIDIA H100 (1,024-GPU cluster) / H200 / GB200; on-demand pricing |
| Sify Technologies | Asia | NVIDIA-based (tier unstated) | NVIDIA (H100 / H200 / GB200; deployed count undisclosed) |
| SK Telecom GPUaaS | Asia | Lambda (investor/tech), VAST Data | 1,000+ NVIDIA B200 at Gasan AIDC; Petasus virtualization |
| SoftBank AI Cloud | Asia | NVIDIA-based (tier unstated) | NVIDIA GB200 NVL72; Infrinia AI Cloud OS |
| Sustainable Metal Cloud | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 (GB300 planned); immersion-cooled |
| SwissGPU | Europe | NVIDIA-based (tier unstated) | NVIDIA A100 / H100 + RTX-class; own Swiss data-center servers |
| Taiwan AI Cloud | Asia | NVIDIA-based (tier unstated) | Taiwania + NVIDIA H100 / AFS PODs; pay-as-you-go |
| Tata Communications | Asia | NVIDIA (full stack: NIM, NeMo, AI Enterprise) | NVIDIA H100 / H200 / L40S (Blackwell planned); HGX with InfiniBand |
| TensorWave | North America | AMD-exclusive; runs MI300X, MI325X, MI355X and MI455X Instinct accelerators, with no Nvidia in the fleet. | — |
| Türk Telekom | Middle East | NVIDIA-based (tier unstated) | NVIDIA virtual GPU instances + dedicated GPU servers; own Turkish DCs |
| Ubilink | Asia | NVIDIA-based (tier unstated) | NVIDIA H100 SuperPOD (~1,024 GPUs); 100% renewable |
| Ubitus | Asia | NVIDIA-based (tier unstated) | NVIDIA Blackwell GPU cloud (from H100 DGX) |
| UzCloud | Asia | NVIDIA-based (tier unstated) | NVIDIA DGX B200 (8x B200/node); Run:ai orchestration |
| Vast AI | North America | NVIDIA-based (tier unstated) | Marketplace: NVIDIA H100 / A100 / RTX (third-party supply) |
| Verda (formerly DataCrunch) | Europe | Nvidia Preferred Partner since 2025. | — |
| Viettel | Asia | NVIDIA-based (tier unstated) | NVIDIA H200 + DGX B200 (~1.5 exaFLOPS); vCGPU product |
| Voltage Park | North America | NVIDIA (reference NVIDIA Cloud Partner, Nov 2025), Dell, Penguin Solutions, Vast Data | NVIDIA H100 (B200 / GB300 added) |
| Yotta Data Services | Asia | Nvidia (H100 and L40S GPUs across Shakti Cloud AI Lab, Workspace VM, Bare Metal, and Cluster tiers) | NVIDIA H100 (+ Blackwell Ultra) |
| YTL AI Cloud | Asia | NVIDIA | NVIDIA GB200 NVL72 Grace Blackwell (from H100); ~600 MW solar-supported data-center park |
Partner status and accelerators are drawn from each operator’s sourced record. “NVIDIA-based (tier unstated)” means NVIDIA usage is confirmed but no specific partner tier is on the record yet. The full roster and inclusion test are at /neoclouds.
The non-NVIDIA exception
The neocloud category is overwhelmingly NVIDIA-based, which is why NVIDIA partner tier is a meaningful axis at all. The clearest exception DEPLOY tracks is TensorWave, a rare AMD-Instinct neocloud. As alternative accelerators (AMD, custom silicon) win more deployments, this list will grow, and it is tracked here honestly rather than assumed.
Common questions
What is an NVIDIA Cloud Partner?
It is a cloud provider that NVIDIA validates to deliver reference-architecture GPU infrastructure. NVIDIA runs several partner designations that a neocloud can hold, including Reference Platform partner, and the Elite and Preferred tiers within the NVIDIA Partner Network, plus program badges such as Cloud Partner and Exemplar Cloud. The tier signals how closely the operator's stack matches NVIDIA's validated design.
Which neoclouds hold the highest NVIDIA partner tiers?
Among tracked operators, Denvr Dataworks (Nvidia Preferred Partner and Intel Partner; ships Intel Gaudi 2 as a listed production SKU (uncommon among neoclouds)); GreenNode (Preferred NVIDIA Cloud Partner); Ooredoo (NVIDIA Cloud Partner); Sharon AI (NVIDIA Cloud / Preferred Partner); Voltage Park (NVIDIA (reference NVIDIA Cloud Partner, Nov 2025), Dell, Penguin Solutions, Vast Data); Massed Compute (Nvidia Preferred Partner; catalog spans A100, L40S, H100, H200, B100, B200, B300, plus RTX Pro Blackwell and RTX 6000 Ada) carry an explicit NVIDIA partner designation on their record.
Are there neoclouds that do not use NVIDIA?
Yes, but they are the exception. The category is overwhelmingly NVIDIA-based. The clearest non-NVIDIA operator DEPLOY tracks is TensorWave, which builds on AMD Instinct GPUs. A handful of others experiment with AMD or custom silicon alongside NVIDIA, but no tracked operator runs at NVIDIA’s scale on non-NVIDIA chips today.
How is partner status verified?
Each partner designation is drawn from the operator's sourced registry record, not from a marketing page alone. Where a primary source (the operator's own announcement, NVIDIA, or reputable press) confirms a specific tier, it is shown; where only NVIDIA usage is confirmed, the operator is listed as NVIDIA-based with the tier unstated.