Supermicro expands manufacturing footprint to 5,000 well-tested AI, high-speed computing and liquid-cooled server solutions per month

Supermicro expands global manufacturing footprint, increasing cabinet- level manufacturing capacity to 5,000 fully tested artificial intelligence, high-speed computing and liquid cooling server solutions globally per month

Rack-level manufacturing capacity has been increased in many countries and regions in Europe, the United States and Asia, and the delivery of new AI and HPC technologies has been accelerated. The power density can reach 100 kilowatts per rack.

[ November 15 , 2023 , San Jose and Denver, California, USA ] Supermicro, Inc., as a comprehensive IT solution manufacturer in the fields of AI, cloud, storage and 5G/edge, is expanding its artificial intelligence (AI) and high-end Performance computing (HPC) availability and advanced liquid-cooled server solutions. Supermicro's global complete rack-level supply capabilities are expanding through several advanced integrated manufacturing facilities in North America, Europe and Asia. Supermicro is actively considering future manufacturing expansions and locations to meet growing customer demand for Supermicro's portfolio of artificial intelligence and high-performance computing rack-scale solutions.

"With our global footprint, we are now able to deliver 5,000 racks per month to support the large number of liquid-cooled racks that require fully integrated storage, with up to 100 racks per rack," said Charles Liang, Supermicro president and CEO. kilowatts of orders. We estimate that as CPU and GPU operating temperatures continue to rise, up to 20% of new data centers will adopt liquid cooling solutions due to growing global adoption. AI data centers are in high demand, and complete rack-scale and liquid-cooled server solutions should be considered early in the design and execution process. This will help shorten delivery times to meet the demand for AI and AI data centers. The urgent need for hyperscale data centers.”

Supermicro maintains high-specification "Golden SKU" inventory to meet the rapid supply needs of global deployments. Large cloud service providers and enterprise data centers running the latest generative AI applications will quickly benefit from reduced global delivery times. Supermicro's broad range of servers can be perfectly integrated from the data center to the edge (Internet of Things) to increase adoption and attract more customers.

With the recent launch of the MGX series products, equipped with NVIDIA GH200 Grace™ Hopper™ super chip and NVIDIA Grace™ CPU super chip, Supermicro will continue to expand servers optimized for artificial intelligence. Combined with existing product lines, including NVIDIA HGX 8-GPU solutions optimized for large language models, and NVIDIA L40S and L4 products, plus Intel Data Center MAX GPUs, Intel® Gaudi®2, and AMD Instinct™ MI series of GPUs, Supermicro can handle all fields of artificial intelligence training and inference applications. Supermicro's All-Flash storage server is equipped with NVMe E1.S and E3.S storage systems, which accelerates data access for various artificial intelligence training applications and enables faster operation. For high-performance computing applications, the Supermicro SuperBlade equipped with GPU can reduce the running time of high-profile simulation applications and reduce power consumption at the same time.

When liquid cooling technology is integrated into a data center, it can reduce the power usage efficiency (PUE) of the data center by up to 50% compared to the current industry average. Reducing the energy footprint and resulting lower power usage efficiency when running generative artificial intelligence or high-performance computing simulation applications can significantly reduce data center operating expenses.

Through the rack-level integration and deployment services provided by Supermicro, customers can quickly install from proven reference designs while considering their unique business goals, and can then work with Supermicro certified experts to co-design specific Optimized solutions for workloads. Once the servers are delivered, simply connect the racks to power, network and liquid cooling infrastructure for a plug-and-play solution. Supermicro is committed to providing complete data center IT solutions, including on-site delivery, deployment, integration and benchmarking, to achieve optimal operational efficiency.

Supermicro’s broad server product portfolio includes:

SuperBlade ® - Supermicro's high-performance, density-optimized and energy-efficient multi-node platform optimized for artificial intelligence, data analytics, high-performance computing, cloud and enterprise workloads.

GPU server with PCIe GPU support - A system that supports advanced accelerators to significantly improve performance and save costs. Designed for high-performance computing, artificial intelligence/machine learning, rendering and virtual desktop infrastructure (Virtual Desktop Infrastructure) workloads.

Universal GPU Server - An open, modular, standards-based server that delivers superior performance and availability with GPUs. Supported GPU specification options include the latest PCIe, OAM and NVIDIA SXM technologies.

Petascale Storage - Industry-leading storage density and performance, equipped with EDSFF E1.S and E3.S drives, enabling superior capacity and performance in a single 1U or 2U chassis.

Hyper - the flagship performance rack server built to handle the most challenging workloads while providing the storage and I/O flexibility to meet the needs of a wide variety of custom applications.

Hyper-E - Offers the powerful performance and flexibility of the flagship Hyper series, designed for optimized deployment in edge environments. Edge-friendly features, including short-depth chassis and front-facing I/O, make Hyper-E suitable for edge data centers and telecom racks.

BigTwin ® - 2U 2-node or 2U 4-node platform with dual processors per node and tool-less hot-swappable design, delivering superior density, performance and serviceability. These systems are ideal for cloud, storage and media workloads.

GrandTwin ™ - Designed for single-processor performance and memory density, with front-end (cold aisle) hot-plug nodes and front or rear I/O for easier maintenance.

FatTwin ® - Advanced, high-density multi-node 4U dual-machine architecture with 8 or 4 single-processor nodes optimized for data center compute or storage density.

Edge Server - High-density processing power in a compact form factor optimized for telecom rack and edge data center setups. Optional DC power supply and higher operating temperatures up to 55 degrees Celsius (131 degrees Fahrenheit).

CloudDC - an all-in-one platform for cloud data centers, with flexible I/O and storage configurations, and dual AIOM slots (supports PCIe 5.0; complies with OCP 3.0 standards) to achieve maximum data processing capacity.

WIO - Offers a wide range of I/O matching options to deliver a system truly optimized for specific enterprise needs.

Mainstream - Cost-effective dual-processor platform for everyday enterprise workloads.

Enterprise Storage - Ideal for large-scale storage workloads, leveraging 3.5-inch rotating media for high density and superior total cost of ownership (TCO). Front-load and front/rear-load configurations allow drives to be easily linked, while tool-less brackets simplify maintenance.

Workstations - Supermicro workstations deliver data center-class performance in a portable, under-desk form factor ideal for artificial intelligence, 3D design, and media and entertainment workloads in offices, research labs, and field offices.