GPU CardsComing soon

MetaX C550 64GB

MetaX C-550 is a flagship OAM module on GPGPU/XCORE architecture, engineered on MetaX's proprietary GPU IP and compliant with OAM 1.5/2.0 standards.

Its key specification is 896 GB/s bandwidth with 8-GPU full interconnect — twice the C-500's 384 GB/s — delivering significantly lower latency in distributed training.

3D Mesh topology support via the C550 3D Mesh Supernode provides low-latency connectivity for up to 8 servers and 64 cards in a supernode.

Compatible with MXMACA and major ML frameworks. SR-IOV vGPU virtualization supports multi-tenant cloud deployments.

Application scenarios

Intelligent computing and GPGPU in high-density data centers.
Training and inference with reduced latency (8-GPU full interconnect). MetaX multi-tenant cloud platforms with vGPU.
HPC clusters with 3D Mesh topology.

Product overview

Purpose-built accelerator for enterprise AI workloads

Designed for dense inference, model adaptation, and private AI infrastructure where predictable availability, localized supply, and software compatibility are critical.

Architecture

GPGPU/XCORE

Memory & Performance

High-bandwidth memory profile for AI inference and training.

VRAM capacity64

Memory typeHBM2e

Memory bandwidth440

Interconnect typeMetaXLink

Interconnect speed896

Architecture

Compute architecture and software execution model.

ArchitectureGPGPU/XCORE

Compute units-

Power & Thermal

Data center integration requirements.

Thermal design power450

CoolingПассивное

Form factorOAM

Pixel Rate-

Texture Rate-

Benchmarks

Peak theoretical compute for common AI precisions.

FP64

FP32

FP16

240

TF32

BF16 Tensor

FP8 Tensor

INT8 Tensor

Compatibility

Interfaces, frameworks, and deployment environment.

PCIe interfacePCIe 5.0x16

Video encoding-

Video decoding-

Physical Dimensions

Card dimensions for server platforms.

Slots-

Length- mm

Height- mm

Width- mm

Pricing

On request

Volume pricing available for cluster deployments and pilot batches.

Documentation