Supermicro Ramps Full Production of NVIDIA Blackwell Rack-Scale Solutions with NVIDIA HGX B200
Supermicro announces full production availability of its end-to-end AI data center solutions powered by the NVIDIA Blackwell platform. The company introduces new air-cooled and liquid-cooled systems featuring the NVIDIA HGX B200 8-GPU architecture with enhanced cooling capabilities.
The new liquid-cooled 4U system includes improved cold plates and a 250kW cooling distribution unit, doubling the previous generation's cooling capacity. The rack-scale design supports configurations of 42U, 48U, or 52U, enabling up to 96 NVIDIA Blackwell GPUs in a 52U rack.
The air-cooled 10U system features a redesigned chassis supporting eight 1000W TDP Blackwell GPUs, delivering up to 15x inference and 3x training performance. The SuperCluster designs incorporate NVIDIA networking solutions, enabling scalable units of up to 768 GPUs in nine racks.
Supermicro annuncia la disponibilità di produzione completa delle sue soluzioni di data center AI end-to-end alimentate dalla piattaforma NVIDIA Blackwell. L'azienda presenta nuovi sistemi a liquido e ad aria che incorporano l'architettura NVIDIA HGX B200 con 8 GPU e capacità di raffreddamento potenziate.
Il nuovo sistema a liquido 4U include piastre di raffreddamento migliorate e un'unità di distribuzione del raffreddamento da 250kW, raddoppiando la capacità di raffreddamento della generazione precedente. Il design rack-scale supporta configurazioni di 42U, 48U o 52U, consentendo fino a 96 GPU NVIDIA Blackwell in un rack da 52U.
Il sistema ad aria 10U presenta un telaio riprogettato che supporta otto GPU Blackwell da 1000W TDP, offrendo prestazioni di inferenza fino a 15 volte superiori e prestazioni di training 3 volte superiori. I design SuperCluster incorporano soluzioni di networking NVIDIA, consentendo un'unità scalabile fino a 768 GPU in nove rack.
Supermicro anuncia la disponibilidad completa en producción de sus soluciones completas de centro de datos AI impulsadas por la plataforma NVIDIA Blackwell. La empresa presenta nuevos sistemas de refrigeración por aire y líquida que cuentan con la arquitectura NVIDIA HGX B200 de 8 GPU y capacidades de refrigeración mejoradas.
El nuevo sistema de refrigeración líquida de 4U incluye placas frías mejoradas y una unidad de distribución de refrigeración de 250kW, duplicando la capacidad de refrigeración de la generación anterior. El diseño a escala de rack admite configuraciones de 42U, 48U o 52U, permitiendo hasta 96 GPU NVIDIA Blackwell en un rack de 52U.
El sistema refrigerado por aire de 10U cuenta con un chasis rediseñado que soporta ocho GPUs Blackwell de 1000W TDP, ofreciendo hasta 15 veces el rendimiento de inferencia y 3 veces el rendimiento de entrenamiento. Los diseños SuperCluster incorporan soluciones de red de NVIDIA, lo que permite unidades escalables de hasta 768 GPUs en nueve racks.
슈퍼마이크로는 NVIDIA 블랙웰 플랫폼으로 구동되는 종합 AI 데이터 센터 솔루션의 전체 생산 가능성을 발표했습니다. 이 회사는 향상된 냉각 기능을 갖춘 NVIDIA HGX B200 8-GPU 아키텍처를 특징으로 하는 새로운 공냉 및 수냉 시스템을 소개합니다.
새로운 수냉식 4U 시스템은 개선된 냉각판과 250kW의 냉각 분배 장치를 포함하여 이전 세대의 냉각 용량을 두 배로 늘렸습니다. 랙 스케일 디자인은 42U, 48U 또는 52U 구성 지원, 52U 랙에서 최대 96 NVIDIA 블랙웰 GPU를 지원합니다.
공냉식 10U 시스템은 1000W TDP 블랙웰 GPU 8개를 지원하는 재설계된 섀시를 특징으로 하며, 최대 15배의 추론과 3배의 훈련 성능을 제공합니다. 슈퍼클러스터 디자인에는 NVIDIA 네트워킹 솔루션이 포함되어 있어 최대 768 GPU의 확장 가능한 유닛을 아홉 개의 랙에 배치할 수 있습니다.
Supermicro annonce la disponibilité complète en production de ses solutions de centre de données AI de bout en bout alimentées par la plateforme NVIDIA Blackwell. L'entreprise présente de nouveaux systèmes refroidis par air et par liquide dotés de l'architecture NVIDIA HGX B200 à 8 GPU avec des capacités de refroidissement améliorées.
Le nouveau système refroidi par liquide de 4U comprend des plaques froides améliorées et une unité de distribution de refroidissement de 250 kW, doublant la capacité de refroidissement de la génération précédente. La conception à l'échelle du rack prend en charge des configurations de 42U, 48U ou 52U, permettant jusqu'à 96 GPU NVIDIA Blackwell dans un rack de 52U.
Le système refroidi par air de 10U dispose d'un châssis redessiné supportant huit GPU Blackwell de 1000W TDP, offrant jusqu'à 15 fois la performance d'inférence et 3 fois la performance d'entraînement. Les conceptions SuperCluster intègrent des solutions de mise en réseau NVIDIA, permettant des unités évolutives allant jusqu'à 768 GPU dans neuf racks.
Supermicro kündigt die vollständige Produktionsverfügbarkeit seiner durchgehenden AI-Datenzentrumslösungen an, die auf der NVIDIA Blackwell-Plattform basieren. Das Unternehmen stellt neue luft- und flüssigkeitsgekühlte Systeme vor, die die NVIDIA HGX B200 8-GPU-Architektur mit verbesserten Kühlmöglichkeiten bieten.
Das neue 4U-Flüssigkeitskühlsystem umfasst verbesserte Kaltplatten und eine 250-kW-Kühlverteilungseinheit, die die Kühlkapazität der vorherigen Generation verdoppelt. Das Rack-Skalendesign unterstützt Konfigurationen von 42U, 48U oder 52U und ermöglicht bis zu 96 NVIDIA Blackwell GPUs in einem 52U-Rack.
Das luftgekühlte 10U-System verfügt über ein neu gestaltetes Gehäuse, das acht 1000W TDP Blackwell GPUs unterstützt und bis zu 15-fache Inferenz- und 3-fache Trainingsleistung bietet. Die SuperCluster-Designs integrieren NVIDIA-Netzwerklösungen, die skalierbare Einheiten von bis zu 768 GPUs in neun Racks ermöglichen.
- New cooling technology doubles previous generation capacity while maintaining form factor
- Up to 15x inference and 3x training performance improvement
- Scalable architecture supporting up to 768 GPUs in nine racks
- Global manufacturing capabilities ensuring timely delivery
- None.
Insights
Supermicro's rollout of NVIDIA Blackwell-powered solutions marks a pivotal advancement in AI infrastructure, positioning the company at the forefront of the AI computing revolution. The technological leap is substantial, with systems delivering up to 15x inference and 3x training performance improvements over previous generations - a game-changing efficiency boost for AI workloads.
The enhanced cooling capabilities, particularly the new 250kW cooling distribution unit, represent a significant competitive advantage. This innovation effectively doubles cooling capacity while maintaining the same form factor, enabling denser GPU configurations and improved power efficiency. For data center operators, this translates to substantially lower total cost of ownership (TCO) and improved performance per square foot.
Three key strategic advantages emerge from this launch:
- The ability to scale up to 96 NVIDIA Blackwell GPUs in a 52U rack positions Supermicro as a leader in high-density AI computing solutions
- The comprehensive end-to-end deployment service, including global manufacturing capabilities, reduces implementation complexity for enterprise customers
- The native support for NVIDIA AI Enterprise software accelerates time-to-market for AI applications, a important factor in the current AI race
This announcement significantly strengthens Supermicro's market position in the rapidly growing AI infrastructure sector. The company's ability to deliver both liquid and air-cooled solutions, coupled with its global manufacturing scale, positions it favorably to capture market share in the expanding AI data center market.
Supermicro Offers Next-Generation Air-Cooled and Liquid-Cooled Architecture for NVIDIA Blackwell Platform
"In this transformative moment of AI, where scaling laws are pushing the limits of data center capabilities, our latest NVIDIA Blackwell-powered solutions, developed through close collaboration with NVIDIA, deliver outstanding computational power," said Charles Liang, president and CEO of Supermicro. "Supermicro's NVIDIA Blackwell GPU offerings in plug-and-play scalable units with advanced liquid cooling and air cooling are empowering customers to deploy an infrastructure that supports increasingly complex AI workloads while maintaining exceptional efficiency. This reinforces our commitment to providing sustainable, cutting-edge solutions that accelerate AI innovation."
For more information, please visit www.supermicro.com/AI
Supermicro's NVIDIA HGX B200 8-GPU systems utilize next-generation liquid-cooling and air-cooling technology. The newly developed cold plates and the new 250kW coolant distribution unit (CDU) more than double the cooling capacity of the previous generation in the same 4U form factor. Available in 42U, 48U, or 52U configurations, the rack-scale design with the new vertical coolant distribution manifolds (CDM) no longer occupy valuable rack units. This enables 8 systems, comprising 64 NVIDIA Blackwell GPUs in a 42U rack, and all the way up to 12 systems with 96 NVIDIA Blackwell GPUs in a 52U rack.
The new air-cooled 10U NVIDIA HGX B200 system features a redesigned chassis with expanded thermal headroom to accommodate eight 1000W TDP Blackwell GPUs. Up to 4 of the new 10U air-cooled systems can be installed and fully integrated in a rack, the same density as the previous generation, while providing up to 15x inference and 3x training performance.
The new SuperCluster designs incorporate NVIDIA Quantum-2 InfiniBand or NVIDIA Spectrum-X Ethernet networking in a centralized rack, enabling a non-blocking, 256-GPU scalable unit in five racks or an extended 768-GPU scalable unit in nine racks. This architecture — purpose-built for NVIDIA HGX B200 systems with native support for the NVIDIA AI Enterprise software platform for developing and deploying production-grade, end-to-end agentic AI pipelines — combined with Supermicro's expertise in deploying the world's largest liquid-cooled data centers delivers exceptional efficiency and time-to-online for today's most ambitious AI data center projects.
Liquid-cooled or air-cooled: Supermicro NVIDIA HGX B200 Systems
The new liquid-cooled 4U NVIDIA HGX B200 8-GPU system features newly developed cold plates and advanced tubing design that further enhance the efficiency and serviceability of the predecessor that was used for the NVIDIA HGX H100/H200 8-GPU system. Complemented by a new 250kW cooling distribution unit, more than doubling the cooling capacity of the previous generation while maintaining the same 4U form factor, the new rack-scale design with the new vertical coolant distribution manifolds (CDM) enables denser architecture with flexible configuration scenarios used for various data center environments. Supermicro offers 42U, 48U, or 52U rack configurations for liquid-cooled data centers. The 42U or 48U configuration provides 8 systems and 64-GPU in a rack, and 256-GPU scalable unit in five racks. The 52U rack configuration allows 96-GPU in a rack and enables 768-GPU scalable unit in nine racks for the most advanced AI data center deployments. Supermicro also offers an in-row CDU option for large deployments, as well as liquid-to-air cooling rack solution that doesn't require facility water.
Supermicro's NVIDIA HGX B200 systems natively support NVIDIA AI Enterprise software to accelerate time to production AI. NVIDIA NIM microservices allow organizations to access the latest AI models for fast, secure, and reliable deployment on NVIDIA accelerated infrastructure anywhere — whether in data centers, the cloud or workstations.
For traditional data centers, the new 10U air-cooled NVIDIA B200 8-GPU system is also available, with a redesigned modular GPU tray to house the NVIDIA Blackwell GPUs in an air-cooled environment. The air-cooled rack design follows the proven, industry-leading architecture of the previous generation, four systems and 32 GPUs in a 48U rack, while providing NVIDIA Blackwell performance. All Supermicro NVIDIA HGX B200 systems are equipped with a 1:1 GPU-to-NIC ratio supporting NVIDIA BlueField-3 SuperNICs or NVIDIA ConnectX-7 NICs for scaling across a high-performance compute fabric.
Supermicro provides support for systems included in the NVIDIA-Certified Systems program. This program incorporates NVIDIA GPUs, CPUs, and high-speed, secure networking technologies into systems from leading NVIDIA partners, ensuring configurations that are validated for optimal performance, reliability, and scalability. By choosing an NVIDIA-Certified System, enterprises can confidently select hardware solutions to power their accelerated computing workloads. NVIDIA has certified Supermicro systems with NVIDIA H100 and H200 GPUs.
End-to-end Liquid-cooling Solution for NVIDIA GB200 NVL72
Supermicro's SuperCluster solution, based on the NVIDIA GB200 NVL72 system, represents a breakthrough in AI computing infrastructure, combining Supermicro's end-to-end liquid-cooling technology. The system integrates 72 NVIDIA Blackwell GPUs and 36 NVIDIA Grace CPUs in a single rack, delivering exascale computing capabilities through NVIDIA's most extensive NVLink network to date, achieving 130 TB/s of GPU communications.
The 48U solution's versatility supports both liquid-to-air and liquid-to-liquid cooling configurations, accommodating various data center environments. Additionally, Supermicro's SuperCloud Composer software provides management tools for monitoring and optimizing liquid-cooled infrastructure, delivering a complete solution from proof of concept to full-scale deployment.
End-to-end Data Center Solution and Deployment Services for NVIDIA Blackwell
From proof-of-concept (PoC) to full-scale deployment, Supermicro serves as a comprehensive one-stop solution provider with global manufacturing scale, delivering all necessary components, data center-level solution design, liquid-cooling technologies, networking solutions, cabling, management software, testing and validation, and onsite installation services. Its in-house liquid-cooling ecosystem offers a complete, custom-designed thermal management solution, featuring optimized cold plates for GPUs, CPUs, and memory modules, along with versatile coolant distribution unit form factors and capacities, manifolds, hoses, connectors, cooling towers, and sophisticated monitoring and management software. With production facilities across
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-ramps-full-production-of-nvidia-blackwell-rack-scale-solutions-with-nvidia-hgx-b200-302368747.html
SOURCE Super Micro Computer, Inc.
FAQ
What cooling options are available for Supermicro's NVIDIA Blackwell systems (SMCI)?
How many NVIDIA Blackwell GPUs can fit in Supermicro's (SMCI) 52U rack configuration?
What performance improvements does Supermicro's (SMCI) new air-cooled system offer?
What is the cooling capacity of Supermicro's (SMCI) new cooling distribution unit?
