Is it winning?
Yes
Assessed capability across ten dimensions · 73.7
Loading this page
Assembling this page. Live sources are being read now.
Assembling this page. Live sources are being read now.
Wednesday, 7 October 2026
The dominant AI accelerator company; GPUs (Hopper/Blackwell) and CUDA are the de-facto training and inference substrate, with DGX Cloud as a direct compute offering and a widening strategic-investment portfolio.
Yes
Assessed capability across ten dimensions · 73.7
Unknown
No reputation data published
Yes
What the company discloses about its finances · 85.0
Analyst interpretationThe single highest-concentration dependency in enterprise AI. Treat NVIDIA exposure as a systemic supply risk to plan around, not a vendor to simply select.
| Capability | maturityScore | status | evidenceGrade | lastVerified |
|---|---|---|---|---|
| agents | 84 | tested | E3 | 2026-05-07 |
| deployment_options | 84 | tested | E3 | 2026-05-07 |
| portability | 73.6 | verified | E4 | 2026-07-14 |
| enterprise_assistant | 45 | verified | E2 | 2026-06-21 |
| governance | 61.8 | verified | E4 | 2026-06-21 |
| security | 94 | verified | E4 | 2026-05-07 |
| rag | 79 | tested | E3 | 2026-05-07 |
| models | 94 | verified | E4 | 2026-05-07 |
| cost_controls | 54 | verified | E3 | 2026-07-14 |
| integrations | 67.3 | verified | E4 | 2026-07-14 |
Not covered in the capability matrix
The AI Enterprise capability matrix does not include NVIDIA. No capability scores are shown rather than invented.
OpenAI is among NVIDIA's largest GPU customers; multi-year compute supply relationship.
Nemotron is NVIDIA's first-party model family (Nano / Super / Speech / Safety variants).
Meta is one of NVIDIA's largest customers, public guidance of 350K+ H100 GPUs by end of 2024 for Llama training and Reality Labs.
NVIDIA participated in xAI's Series B; Colossus training cluster runs on H100/H200.
Anthropic trains on NVIDIA accelerators; NVIDIA participated in funding rounds.
DeepSeek trained V3 / R1 on NVIDIA H800 + H20 GPUs (the China-export-compliant variants), exposure is real but subject to ongoing US export-control tightening.
Alibaba Cloud's Qwen training fleet relies on NVIDIA H800 / H20 silicon alongside Alibaba's own T-Head accelerators.
TII's Falcon training capacity spans NVIDIA DGX H100 (UAE installations) alongside the Cerebras Condor Galaxy partnership.
Aleph Alpha's training stack uses NVIDIA H100 alongside its AMD MI300 sovereign-compute pivot.
NVIDIA NVentures participated in Perplexity Series B / B1.
Z.ai's GLM family trained on NVIDIA H800 capacity inside China; sovereign-silicon transition underway.
Tencent Hunyuan training relies on NVIDIA H800 capacity inside Tencent Cloud.
Baidu ERNIE trained on a mix of NVIDIA H800 and Baidu's own Kunlun P800 chips, vertically integrated stack with sovereign hedge.
LightOn enterprise LLMs trained on NVIDIA GPUs hosted in OVHcloud sovereign-EU infrastructure.
Lightweight NIM hosting / inference partnership; no public investment disclosed. Marked seed pending stronger evidence.
| Model | modelFamily | modelCategory | availabilityStage | ownershipType | dataStatus | Source |
|---|---|---|---|---|---|---|
| Nemotron 3 Nano Omni | Nemotron | multimodal | ga | first_party | seed · E2 | developer.nvidia.com |
| Nemotron 3 Super | Nemotron | llm_text | ga | first_party | seed · E2 | developer.nvidia.com |
| Nemotron Speech | Nemotron | speech_audio | ga | first_party | seed · E2 | developer.nvidia.com |
| Nemotron / NIM (verification required) | Nemotron | unknown | unknown | unknown | seed · E2 | developer.nvidia.com |
Not covered in the reputation seed
The AI Enterprise reputation dataset does not cover NVIDIA.
No curated sources listed
The sourcing manifest holds no curated evidence URLs for NVIDIA.