AMD Instinct™ MI325X GPU

AMD Instinct™ MI325X GPUs deliver incredible performance in HPC, AI training, and AI deployment workloads.

With 256 GB of HBME3 memory and 6 TB/s memory bandwidth, AMD Instinct™ MI325X GPUs optimize performance with inference-optimized acceleration and efficiency.

AMD Instinct™ MI300X GPU

AMD Instinct™ MI300X GPUs provide powerful and efficient HPC and AI training and deployments.

With 192 GB of HBM3 memory, 5.3 TB/s memory bandwidth, and the AMD CDNA™ 3 architecture, AMD Instinct™ MI300X GPUs are built to excel efficiently with compute-intensive workloads.

AMD Instinct™ MI325X GPU
Coming soon...
AMD Instinct™ MI300X GPU
Starting at
$1.841 / Per hour

Enterprise-ready at any scale and any location

Clusters at any size

Vultr's enterprise-ready infrastructure seamlessly supports any cluster size of AMD Instinct™ GPUs. Whether you require a small cluster or a massive deployment, Vultr ensures reliable, high-performance computing to meet your specific needs.

Globally available, locally accessible

Large clusters of AMD Instinct™ GPUs are available where you need them, thanks to Vultr's extensive infrastructure. With 32 cloud data center regions across six continents, we guarantee low latency and high availability, enabling your enterprise to achieve optimal performance worldwide.

Enterprise-grade compliance and security

Vultr ensures our platform, products, and services meet diverse global compliance, privacy, and security needs, covering areas such as server availability, data protection, and privacy. Our commitment to industry-wide privacy and security frameworks demonstrates our dedication to protecting our customers' data.

Designed to be inference-optimized from the start, while excelling at HPC and AI training workloads

Deploy large models and process more data in-memory for greater efficiency and high performance.

No information is required for download

AMD Instinct™ MI325X GPU
Specifications
Form factor OAM module
GPU compute units 304
Stream processors 1216
Peak engine clock 19,456
Memory capacity Up to 256 GB HBM3E
Memory bandwith 6.0 TB/s max. peak theoretical
Memory interface 8192 bits
AMD Infinity Cache™ (last level) 256 MB
Memory clock Up to 6.0 GT/s
AI peak theoretical performance
With sparsity
TF32 (TFLOPs) 653.7 1307.4
FP16 (TFLOPs) 1307.4 2614.9
BFLOAT16 (TFLOPs) 1307.4 2614.9
INT8 (TOPs) 2614.9 5229.8
FP8 (TFLOPs) 2614.9 5229.8
HPC peak theoretical performance (TFLOPS)
FP64 vector 81.7
FP32 vector 163.4
FP64 matrix 163.4
FP32 matrix 163.4
AMD Instinct™ MI300X GPU
Specifications
Form factor OAM module
GPU compute units 304
Stream processors 1216
Peak engine clock 19,456
Memory capacity Up to 192 GB HBM3
Memory bandwith 5.3 TB/s max. peak theoretical
Memory interface 8192 bits
AMD Infinity Cache™ (last level) 256 MB
Memory clock Up to 5.2 GT/s
AI peak theoretical performance
With sparsity
TF32 (TFLOPs) 653.7 1307.4
FP16 (TFLOPs) 1307.4 2614.9
BFLOAT16 (TFLOPs) 1307.4 2614.9
INT8 (TOPs) 2614.9 5229.8
FP8 (TFLOPs) 2614.9 5229.8
HPC peak theoretical performance (TFLOPS)
FP64 vector 81.7
FP32 vector 163.4
FP64 matrix 163.4
FP32 matrix 163.4

Customer success stories and use cases

Explore how leading organizations in the manufacturing and energy industry, equipped with the right tools, achieve security, connectivity, and efficiency using Vultr’s cloud solutions.

Hybrid cloud connectivity for IoT image processing
Connected drones capture real-time data from oil fields, processed locally by Gravio for anomaly detection. Data is securely transmitted via Console Connect to Vultr’s high-performance cloud, enabling AI workflows for insights, trend detection, and alerts.
Asset management in the oil and gas sector
Harness Vultr Cloud, AMD GPUs, digital twins, and drones to optimize asset management in oil and gas. Enable real-time monitoring, predictive maintenance, and operational efficiency to reduce costs and boost performance.

"At ConfidentialMind, we specialize in generative AI solutions, including RAG applications and AI agents. With Vultr Cloud GPUs, including the AMD Instinct™ MI300X GPU, we ensure unparalleled computational speed and efficiency. Using the Vultr Kubernetes Engine, we simplify deployments across on-prem, private cloud, and edge environments. Vultr’s infrastructure delivers the performance, scalability, and robust data security required by our customers to meet their growing need for AI features without compromising data sovereignty or operational efficiency."

Severi Tikkala
CTO at ConfidentialMind

Reserve the AMD Instinct™
MI325X and MI300X now

Get ready to build, test, and deploy on The Everywhere Cloud.