Cirrascale Cloud Services, a leader in cloud and managed solutions for AI and high-performance computing (HPC), announced the upcoming availability of AMD Instinct™ MI350 Series GPUs and platforms in its AI Innovation Cloud. These new GPU options empower customers with enhanced capabilities for training large AI models, accelerating inference, and running demanding HPC simulations with greater efficiency and scalability.
AMD Instinct MI350: Cutting-Edge Architecture for AI and HPC
Built on AMD’s fourth-generation CDNA™ architecture, the Instinct MI350 Series sets new performance benchmarks for cloud-based AI and HPC workloads. With up to 288GB of high-bandwidth HBM3E memory, the MI350 delivers exceptional speed and capacity for massive AI model training, high-throughput inference, and complex scientific computations.
Expanded support for advanced datatypes including FP16, FP8, and upcoming FP6 and FP4 formats enables maximum throughput and energy efficiency—critical for GenAI and HPC innovations.
Enhanced Flexibility Through Collaboration
Cirrascale’s collaboration with AMD has focused on delivering customized cloud environments optimized for the unique architectures and performance demands of leading AI models. The addition of the MI350 Series expands these capabilities, offering customers finely tuned configurations and greater flexibility for next-generation AI and HPC workloads at scale.
This builds on the recent inclusion of AMD Instinct MI325X GPUs in Cirrascale’s AI Innovation Cloud—already a trusted platform for large-scale deployments of models such as Llama 405B and GPT variants.
Industry-Leading Memory and Scalability
Mike LaPan, Vice President of Marketing at Cirrascale Cloud Services, highlighted the significance of the MI350’s large memory capacity and bandwidth, expanded datatype support, and robust security features. These attributes enable innovators to accelerate AI model development and deployment with confidence and efficiency.
The new platforms support a broad range of applications—from scientific research simulations to financial modeling—offering seamless scalability and future-ready networking for diverse enterprise and research environments
By integrating AMD Instinct MI350 GPUs into its AI Innovation Cloud, Cirrascale strengthens its position as a top choice for scalable, secure, and high-performance AI and HPC infrastructure. This addition empowers organizations to push the boundaries of AI innovation with optimized environments tailored to meet the evolving demands of next-generation models and workloads.
