NVIDIA · Platform · Compute node

NVIDIA GH200 NVL2 288GB

Exact GH200 NVL2 reference specifications and AI deployment guidance, separated from FusionPointAI commercial inventory.

VRAM288 GB

Memory technologyHBM3e

Form factorDual superchip · Node

InterfaceNVLink

PowerNot listed

Current Reference only

Compare GPU

Technical diagrams

Neutral technical reference illustration for NVIDIA GH200 NVL2 288GB; not a product photograph — Reference illustration; consult the verified specification table for exact hardware identity. This is not a product photograph. FusionPointAI neutral GPU reference illustration

Specifications

Field	Value
Manufacturer	NVIDIA
Exact SKU	GH200 NVL2
Product family	NVIDIA GH200
Architecture	Hopper
Generation	Grace Hopper
Entity type	Compute node
Lifecycle	Current
Launch date	Not listed
VRAM	288 GB
Memory technology	HBM3e
Memory bandwidth	9,800 GB/s
Power	Not listed
Interface	NVLink
Form factor	Dual superchip · Node
Interconnect	NVLink
MIG / partitioning	Supported
Supported numeric formats	FP64, FP32, TF32, BF16, FP16, FP8, INT8
Software stack	cuda

AI workload suitability

LLM inference

High memory capacity supports large checkpoints and multi-GPU inference when runtime and topology are certified.

Fine-tuning and training

Suitable for limited fine-tuning and development when software support, cooling and memory capacity are verified.

Image, video and embeddings

FP16 acceleration supports common generative and vector workloads within the listed VRAM envelope.

Power and cooling

The manufacturer does not publish a directly comparable board-power value for this entity.

Multi-GPU use

NVLink is the recorded interconnect path; topology must still be validated at node level.

Software ecosystem

cuda

Product timeline