Supermicro Introduces a New Petascale All-Flash Storage Server Using NVIDIA Grace CPU Superchip for High Performance Software-Defined AI Storage Workloads
Supermicro (NASDAQ: SMCI) has unveiled a new Petascale All-Flash Storage Server featuring the NVIDIA Grace CPU Superchip. This innovative 1U storage system (ARS-121L-NE316R) is designed for high-performance software-defined storage workloads in AI and ML applications.
The server incorporates 144 Arm Neoverse V2 cores and supports 16 hot-swap EDSFF PCIe Gen5 E3.S NVMe drives, enabling 983TB raw capacity using 61.44TB SSDs. A rack containing 40 systems can provide 39.3PB of raw storage capacity.
The system features integrated 960GB of LPDDR5X on the CPU platform and can support two NVIDIA BlueField-3 or ConnectX-8 SuperNICs. Supermicro has collaborated with WEKA to optimize the solution for enterprise AI and ML training, analytics, and inferencing workloads, promising improved performance scaling and power efficiency compared to x86-based systems.
Supermicro (NASDAQ: SMCI) ha presentato un nuovo server di archiviazione All-Flash Petascale che utilizza il NVIDIA Grace CPU Superchip. Questo innovativo sistema di archiviazione 1U (ARS-121L-NE316R) è progettato per carichi di lavoro di archiviazione software-defined ad alte prestazioni nelle applicazioni di AI e ML.
Il server incorpora 144 core Arm Neoverse V2 e supporta 16 unità EDSFF PCIe Gen5 E3.S NVMe hot-swap, consentendo una capacità grezza di 983TB utilizzando SSD da 61.44TB. Un rack contenente 40 sistemi può fornire una capacità di archiviazione grezza di 39.3PB.
Il sistema presenta 960GB di LPDDR5X integrato sulla piattaforma CPU e può supportare due NVIDIA BlueField-3 o ConnectX-8 SuperNICs. Supermicro ha collaborato con WEKA per ottimizzare la soluzione per la formazione, l'analisi e l'inferenza di AI e ML in ambito enterprise, promettendo un miglioramento delle prestazioni e dell'efficienza energetica rispetto ai sistemi basati su x86.
Supermicro (NASDAQ: SMCI) ha presentado un nuevo servidor de almacenamiento All-Flash Petascale que cuenta con el NVIDIA Grace CPU Superchip. Este innovador sistema de almacenamiento 1U (ARS-121L-NE316R) está diseñado para cargas de trabajo de almacenamiento definido por software de alto rendimiento en aplicaciones de IA y ML.
El servidor incorpora 144 núcleos Arm Neoverse V2 y admite 16 unidades EDSFF PCIe Gen5 E3.S NVMe hot-swap, permitiendo una capacidad bruta de 983TB utilizando SSD de 61.44TB. Un rack que contenga 40 sistemas puede proporcionar una capacidad de almacenamiento bruto de 39.3PB.
El sistema cuenta con 960GB de LPDDR5X integrado en la plataforma de la CPU y puede soportar dos NVIDIA BlueField-3 o ConnectX-8 SuperNICs. Supermicro ha colaborado con WEKA para optimizar la solución para la formación, análisis e inferencia de IA y ML en entornos empresariales, prometiendo una mejora en la escalabilidad del rendimiento y la eficiencia energética en comparación con los sistemas basados en x86.
슈퍼마이크로 (NASDAQ: SMCI)는 NVIDIA Grace CPU Superchip을 특징으로 하는 새로운 페타스케일 올플래시 스토리지 서버를 공개했습니다. 이 혁신적인 1U 스토리지 시스템 (ARS-121L-NE316R)은 AI 및 ML 애플리케이션의 고성능 소프트웨어 정의 스토리지 작업을 위해 설계되었습니다.
서버는 144 Arm Neoverse V2 코어를 통합하고 16개의 핫스왑 EDSFF PCIe Gen5 E3.S NVMe 드라이브를 지원하여 983TB의 원시 용량을 제공합니다. 40개의 시스템이 포함된 랙은 39.3PB의 원시 스토리지 용량을 제공할 수 있습니다.
시스템은 CPU 플랫폼에 통합된 960GB의 LPDDR5X를 특징으로 하며 두 개의 NVIDIA BlueField-3 또는 ConnectX-8 SuperNICs를 지원할 수 있습니다. 슈퍼마이크로는 WEKA와 협력하여 기업 AI 및 ML 교육, 분석 및 추론 작업을 최적화하여 x86 기반 시스템에 비해 성능 확장성과 전력 효율성을 개선할 것을 약속합니다.
Supermicro (NASDAQ: SMCI) a dévoilé un nouveau serveur de stockage All-Flash Petascale doté du NVIDIA Grace CPU Superchip. Ce système de stockage 1U innovant (ARS-121L-NE316R) est conçu pour des charges de travail de stockage défini par logiciel à haute performance dans des applications d'IA et de ML.
Le serveur intègre 144 cœurs Arm Neoverse V2 et prend en charge 16 disques EDSFF PCIe Gen5 E3.S NVMe hot-swap, permettant une capacité brute de 983 To avec des SSD de 61,44 To. Un rack contenant 40 systèmes peut fournir une capacité de stockage brute de 39,3 Po.
Le système dispose de 960 Go de LPDDR5X intégré sur la plateforme CPU et peut prendre en charge deux NVIDIA BlueField-3 ou ConnectX-8 SuperNICs. Supermicro a collaboré avec WEKA pour optimiser la solution pour la formation, l'analyse et l'inférence d'IA et de ML en entreprise, promettant une amélioration de l'évolutivité des performances et de l'efficacité énergétique par rapport aux systèmes basés sur x86.
Supermicro (NASDAQ: SMCI) hat einen neuen Petascale All-Flash-Speicherserver mit dem NVIDIA Grace CPU Superchip vorgestellt. Dieses innovative 1U-Speichersystem (ARS-121L-NE316R) ist für hochleistungsfähige softwaredefinierte Speicherarbeitslasten in KI- und ML-Anwendungen konzipiert.
Der Server integriert 144 Arm Neoverse V2-Kerne und unterstützt 16 hot-swap EDSFF PCIe Gen5 E3.S NVMe-Laufwerke, die eine rohkapazität von 983TB mit 61.44TB SSDs ermöglichen. Ein Rack mit 40 Systemen kann eine rohkapazität von 39.3PB bereitstellen.
Das System verfügt über 960GB LPDDR5X, das in die CPU-Plattform integriert ist, und kann zwei NVIDIA BlueField-3 oder ConnectX-8 SuperNICs unterstützen. Supermicro hat mit WEKA zusammengearbeitet, um die Lösung für Unternehmens-KI- und ML-Training, Analytik und Inferenzarbeitslasten zu optimieren und verspricht eine verbesserte Leistungsskalierung und Energieeffizienz im Vergleich zu x86-basierten Systemen.
- First-of-its-class storage server from a major Tier 1 provider
- Significant storage capacity: 983TB raw capacity per system
- High-density scaling capability: 39.3PB per rack
- Enhanced power efficiency compared to x86-based systems
- Linear scalability with PCIe Gen5 performance
- None.
Insights
Supermicro's new Petascale storage server represents a significant technological advancement in the AI infrastructure space. By leveraging the NVIDIA Grace CPU Superchip with 144 Arm Neoverse V2 cores, Supermicro has created a specialized high-performance storage solution that breaks away from traditional x86 architectures. The system's specifications are impressive - supporting 16 hot-swap EDSFF PCIe Gen5 E3.S NVMe drives with up to
What makes this announcement particularly notable is Supermicro's strategic pivot to address the specific storage bottlenecks in AI workloads. Storage I/O has become a critical constraint in AI training and inference pipelines, and this purpose-built system demonstrates Supermicro's understanding of evolving market requirements. The collaboration with both NVIDIA for hardware and WEKA for storage software creates a well-rounded ecosystem approach that should enhance adoption.
The emphasis on power efficiency compared to x86-based systems aligns perfectly with the growing focus on sustainable computing and reduced operational costs in data centers supporting AI workloads. This could provide Supermicro with a meaningful competitive advantage as energy costs and environmental considerations become increasingly important customer decision factors.
This storage server introduction addresses a critical and often overlooked component of the AI infrastructure stack. As AI models continue to grow exponentially in size and complexity, the storage layer has become a significant bottleneck in training and inference workflows. Supermicro's collaboration with NVIDIA and WEKA demonstrates a sophisticated understanding of the entire AI pipeline requirements.
The use of NVIDIA's Grace CPU represents an important strategic direction. While NVIDIA is known primarily for GPUs, their expansion into specialized AI CPUs creates new possibilities for system design. The 144 Arm cores paired with 960GB of integrated LPDDR5X memory enables a zero-copy architecture that significantly reduces data movement overhead - a important factor in AI workload efficiency.
WEKA's involvement is equally significant, as their high-performance file system is specifically optimized for AI workloads. Initial testing showing "significant performance scaling and power efficiency" suggests real-world benefits rather than just theoretical improvements. For organizations building AI factories or large language model training infrastructure, this combination of hardware and software optimization could deliver meaningful operational advantages.
The timing of this announcement coinciding with GTC 2025 maximizes visibility within the AI community and positions Supermicro as an innovator in the increasingly competitive AI infrastructure market. This product differentiates Supermicro from generalist server providers and strengthens their position in the high-growth AI sector.
New System Combines NVIDIA Power Efficient CPU with Supermicro's Petascale Architecture
"The Supermicro Petascale server with the NVIDIA Grace CPU Superchip highlights Supermicro's focus on innovation and customer choice," said Charles Liang, president and CEO of Supermicro. "We have developed the Petascale storage server using the NVIDIA Grace CPU with 144 Arm Neoverse V2 cores enabling high-performance I/O for software-defined storage workloads. We have demonstrated that the system can fully unleash the system's PCIe Gen5 performance SSD bandwidth with linear scalability. Supermicro continues to bring to market the most advanced and optimized storage solutions available."
For more information about the new Grace Storage server, please visit www.supermicro.com/en/products/petascale-grace-storage.
Utilizing Supermicro's Building Block Solutions® system design strategy, the ARS-121L-NE316R 1U storage system incorporates a symmetrical architecture perfect for scaling AI workloads. The Supermicro Petascale storage server using the energy efficient NVIDIA Grace CPU Superchip, supports 16 hot-swap EDSFF PCIe Gen5 E3.S NVMe drives enabling 983TB raw capacity using 61.44TB SSDs and a rack containing 40 systems provides 39.3PB of raw storage capacity.
"We worked closely together on the development and testing of Supermicro's storage server with NVIDIA Grace, giving storage software providers a range of hardware choices to meet different application requirements," said Rob Davis, vice president of storage networking technology at NVIDIA. "This new storage server allows customers to benefit from NVIDIA's innovations in both CPUs and DPUs to accelerate networking protocols, such as GPUDirect Storage, while using less power than comparable x86 servers."
Supermicro is collaborating with WEKA, a leader in high-performance storage software for AI, HPC and enterprise data management workloads. The combination of the 144 Arm cores, integrated 960GB of LPDDR5X on the CPU platform, and ability to support two NVIDIA BlueField-3 or ConnectX-8 SuperNICs will unleash the full performance of the WEKA® Data Platform's zero-copy architecture, providing the highest performance with lower power utilization compared with x86-based systems.
"WEKA is proud to be collaborating with Supermicro and NVIDIA in creating a high-performance solution optimized for enterprise AI and ML training, analytics, and inferencing," said Nilesh Patel, Chief Product Officer at WEKA. "Initial testing of WEKA Data Platform software running on Supermicro's petascale storage server with the NVIDIA Grace CPU-based pre-production cluster shows significant performance scaling and power efficiency."
For more information on the Supermicro storage solutions for AI, please visit: https://www.supermicro.com/en/products/storage.
Supermicro at GTC 2025
Supermicro and NVIDIA will present the new Petascale storage server using the Grace CPU in more detail at NVIDIA GTC in
About Super Micro Computer, Inc.
Supermicro (NASDAQ: SMCI) is a global leader in Application-Optimized Total IT Solutions. Founded and operating in
Supermicro, Server Building Block Solutions, and We Keep IT Green are trademarks and/or registered trademarks of Super Micro Computer, Inc.
All other brands, names, and trademarks are the property of their respective owners.
View original content to download multimedia:https://www.prnewswire.com/news-releases/supermicro-introduces-a-new-petascale-all-flash-storage-server-using-nvidia-grace-cpu-superchip-for-high-performance-software-defined-ai-storage-workloads-302405305.html
SOURCE Super Micro Computer, Inc.
FAQ
What is the storage capacity of Supermicro's new Petascale Storage Server (SMCI)?
What CPU technology does the new Supermicro storage server (SMCI) use?
How does Supermicro's new storage server (SMCI) benefit AI workloads?
What networking capabilities does the new Supermicro storage system (SMCI) offer?
