NeuronPlant does not publish its own model cards. The names, profiles and system claims below are taken from NVIDIA NIM and NVIDIA DGX / GB200 public documentation. We use them to size node vs rack, fabric, storage, and the inference serving plane when delivering a GPU plant.
NP-MDL / Models drive the factory
Profiles and product claims below are taken from NVIDIA public documentation. NeuronPlant uses them as engineering inputs, not as our own benchmarks.
Enterprise reasoning / chat
NVIDIA NIM publishes BF16, FP8 and NVFP4 profiles at TP 1, 2, 4 and 8, with and without LoRA.
Fits a single NVIDIA DGX B200 or NVIDIA DGX B300 for many FP8/NVFP4 profiles. TP8 is the full 8-GPU node. Size KV cache and concurrency separately.
Larger MoE / reasoning
NVIDIA NIM lists BF16, FP8 and NVFP4 across TP 1-8, base and LoRA.
Needs more GPU memory than a 49B. On Blackwell Ultra (NVIDIA DGX B300) dense FP4/NVFP4 is the practical serving path. Training or distillation belongs on a coupled fabric, not on the inference node.
Frontier inference
NVIDIA describes GB200 NVL72 as a 72-GPU NVLink domain for real-time trillion-parameter LLM inference and MoE.
This is a rack (or many racks), liquid loop, and scale-out fabric problem. Do not size it as nine 8-GPU servers on a 100G ToR.
Sources: NVIDIA NIM for LLMs support matrix; NVIDIA GB200 NVL72.
The compute layer is accelerators. GPU is the common case, not the only one. TPU is a named family buyers already specify. The third class is other accelerators and ASICs built to run a model or a graph fast. We do not maintain a weekly chip catalog. NIM is the cited inference runtime on the AI platform, not a NeuronPlant product. Profile sizing belongs with the plant: local NVMe, planes, and concurrency. See the platform layer.
NVIDIA platforms used to size a GPU plant
Node · 8x NVIDIA Blackwell
Unified develop-to-deploy node. NVIDIA lists 1,440 GB GPU memory and up to 400 Gb/s ConnectX-7 / BlueField-3 ports.
Node · 8x NVIDIA Blackwell Ultra
NVIDIA positions NVIDIA DGX B300 for reasoning / inference density. Docs list 8x B300 SXM, ConnectX-8 up to 800 Gb/s, BlueField-3 for storage and management.
Rack · 72 Blackwell + 36 Grace
Liquid-cooled rack-scale NVLink domain. NVIDIA describes the GB200 NVL72 architecture as a 72-GPU NVLink domain for large-model inference and MoE.
Rack · 72 Blackwell Ultra + 36 Grace
NVIDIA DGX rack-scale system. Published networking includes ConnectX-8 at 800 Gb/s InfiniBand and NVLink Switch System.
NP / Intake
One accountable delivery partner from AI requirement to a delivered factory: site, power, cooling, compute, fabric, structured cabling & passive infrastructure, storage, platform, procurement, and handover. You do not need a vendor list or a hall name first. The hall is a supplier.