SAN JOSE, Oct. 14, 2025 /PRNewswire/ — Wiwynn (6669), an innovative cloud IT infrastructure provider for data centers, is showcasing its next-generation AI servers, developed in collaboration with Wistron, and advanced direct liquid cooling (DLC) solutions at OCP Global Summit 2025 (Booth B13, San Jose Convention Center).

“At OCP Global Summit, we’re proud to demonstrate our latest accelerated solutions and cooling innovations,” said William Lin, President and CEO at Wiwynn. “Together with industry leaders, we’re advancing the frontiers of AI computing performance and thermal efficiency. These collaborations enable a comprehensive suite of cutting-edge products that address the surging needs of both modern and legacy data centers in the AI era.”

Next-Generation AI Systems

• NVIDIA GB300 NVL72: Wiwynn and Wistron are among the first to be offering NVIDIA GB300 NVL72 systems. This liquid-cooled, rack-level AI system is built on 72 NVIDIA Blackwell Ultra GPUs and features NVIDIA ConnectX-8 800Gb/s SuperNICs, delivering unprecedented AI factory output for reasoning model inference, leading the next wave of accelerated computing.
• NVIDIA HGX B300: The cutting-edge 10U system is built with eight NVIDIA Blackwell Ultra GPUs. Featuring 2.1TB of HBM3e memory and NVIDIA ConnectX-8 800Gb/s SuperNICs, the NVIDIA HGX B300 system delivers breakthrough performance for demanding AI workloads.
• AMD Instinct™ MI350 Series GPUs and platforms: Powered by AMD Instinct™ MI350 series (AMD CDNA™ 4) and built on a platform featuring EPYC™ CPUs and an AMD Pollara 400 AI Network Interface Card, the system delivers up to 35x improvement in inference performance compared to its prior generation.
• Double-Wide Rack Architecture: Doubles traditional rack width to support larger, higher‑power AI chips, integrating HVDC power, advanced liquid cooling, and signal‑integrity‑optimized layouts to address data, power and thermal challenges in high‑density computing. The new architecture is engineered to accommodate next-generation accelerators such as AMD Instinct™ MI400 Series GPUs, enabling rack scale performance for frontier AI training and inference. (More Details about the Architecture)

Advanced cooling innovations

• Double-side Cold Plate: Designed for future high‑power ICs, Wiwynn’s double‑sided cold plate supports up to 4 kW thermal capability for the primary AI-accelerator and essential vertical power‑delivery ICs concurrently. Leveraging Wiwynn’s proprietary microchannel designs and electrochemical 3D-printing technology, it is expected to provide up to a 40% improvement in thermal performance.
• Two-phase Cold Plate: Designed for two-phase liquid cooling systems using dielectric, eco-friendly refrigerants, Wiwynn’s patented 3D‑printed wicking‑fin cold plate design harnesses phase change to boost heat transfer and reduce leakage‑related downtime—delivering a high‑efficiency, high‑reliability solution for large‑scale liquid‑cooled AI/HPC.
• 300 kW AALC Sidecar: Developed with Shinwa Controls Co.,Ltd., the double-rack-width solution, further boosts AALC (Air-Assisted Liquid Cooling) capability with high reliability and serviceability, enabling liquid-cooled AI racks in air-cooled data centers without infrastructure changes.  
• Future Technologies Symposium preview: Wiwynn and Fabric8Labs will present “Next-generation cold plate with electrochemical 3D-printed microchannels for Ultra-high heat fluxes beyond 350W/cm²” , detailing how advanced thermal technologies and IC package co‑design significantly enhance the cold plate thermal capability for liquid cooling system.

Cutting-edge Networking Solutions

• NVIDIA Spectrum-X: The companies are showcasing NVIDIA Spectrum-X  Ethernet for AI networking, building upon NVIDIA Spectrum-4 MAC technology. This platform integrates SONiC and NVIDIA Cumulus to deliver advanced Ethernet connectivity for multi-tenant, hyperscale AI clouds, enabling flexible datacenter deployment scenarios.
• Broadcom Tomahawk 6: The revolutionary 102T datacenter switch adopted the latest Broadcom Ethernet Switch, supporting 1.6Tbps x 64 ports in a 4RU air-cooled system designated for modern datacenter.

Visit Wiwynn @OCP 2025 (Booth B13, San Jose Convention Center ) , join our engineering workshop sessions and download our latest whitepapers  to explore the latest innovations shaping the future of AI infrastructure and data center technologies.

About Wiwynn

Wiwynn is an innovative cloud IT infrastructure provider of high-quality computing and storage products, plus rack solutions for leading data centers. We are committed to the vision of “unleash the power of digitalization; ignite the innovation of sustainability”. The company aggressively invests in next-generation technologies to provide the best TCO (Total Cost of Ownership), workload, and energy-optimized IT solutions from cloud to edge.

For more information, please visit Wiwynn website, Facebook, and Linkedin.

SAN JOSE, Calif., Oct. 14, 2025 /PRNewswire/ — As AI compute demand continues to surge and data center energy consumption reaches new highs, balancing performance, cost, and sustainability has become a core challenge for cloud providers and enterprises. Compal Electronics (Compal; Ticker: 2324.TW), a pioneer in high-performance AI server platforms, announced that it will present its latest data center technology blueprint at OCP Global Summit 2025, showcasing innovations that span AI strategic collaboration, CXL-based memory solutions, and advanced liquid cooling, underscoring its comprehensive vision for future infrastructure.

In recent years, the server market has shifted beyond raw compute performance, placing increasing emphasis on energy efficiency and scalability. Compal continues to deepen its collaboration with NVIDIA, addressing enterprise requirements for high performance, flexibility, and efficiency, helping accelerate AI factory and inference platform deployment, and shortening time-to-market.

“AI has moved beyond proof-of-concept to become a central pillar of enterprise digital transformation. Our collaboration with NVIDIA focuses on integrating compute, networking, and cooling technologies to help customers build scalable and sustainable AI infrastructure. We warmly invite attendees to visit us at OCP25 and experience our latest AI platform demonstration accelerated by NVIDIA,” said Alan Chang, Vice President of the Infrastructure Solutions Business Group at Compal.

The showcase will also feature PCIe-based memory pooling technology built on CXL (Compute Express Link) and RDMA (Remote Direct Memory Access) architectures, demonstrating next-generation data center innovations in memory expansion, dynamic resource allocation, and low-latency interconnects.

By leveraging advanced interconnect technologies such as PCIe Gen5 (Peripheral Component Interconnect Express, Generation 5), NVMe-oF (Non-Volatile Memory Express over Fabrics), and CXL Fabric, the solution enables high-bandwidth data exchange and resource pooling across server nodes. This architecture further integrates memory and storage tiers, driving the realization of the Storage-as-Memory concept, allowing data to be accessed at near main-memory speed while enhancing overall system performance, scalability, and efficiency.

The on-site demonstration will highlight multi-terabyte memory pooling capabilities and cache-coherency technologies, offering attendees a firsthand look at how these innovations enable a more efficient, flexible, and energy-optimized AI data center infrastructure.

To tackle the thermal demands of high-performance GPUs, Compal will unveil two innovative liquid cooling servers at the OCP Global Summit 2025: the OG720-2A-L2 ORv3 7OU 8x AMD Instinct™ GPUs Direct-to-Chip Liquid-Cooled AI Server and the SG223-2A-I 2U 8x GPU Immersion Cooling Server. These demonstrations will highlight how both solutions sustain optimal performance and energy efficiency under maximum load, enabling data centers to achieve a partial Power Usage Effectiveness (pPUE) below 1.1. The systems provide a scalable and efficient framework for next-generation high-density AI and HPC deployments.

OCP Global Summit 2025 will take place in San Jose Convention Center, California, USA. Visitors to Compal’s Booth C35 will gain insight into how AI, CXL, and advanced liquid cooling technologies are shaping the next decade of data centers — and will have the opportunity to engage directly with Compal’s engineering team to discuss optimal solutions for performance, efficiency, and cost.

About Compal
Founded in 1984, Compal is a leading manufacturer in the notebook and smart device industry, creating brand value in collaboration with various sectors. Its groundbreaking product designs have received numerous international awards. In 2025, Compal was recognized by CommonWealth Magazine as one of Taiwan’s top 7 manufacturers and has consistently ranked among the Forbes Global 2000 companies. In recent years, Compal has actively developed emerging businesses, including cloud servers, auto electronics, and smart medical, leveraging its integrated hardware and software R&D and manufacturing capabilities to create relevant solutions. More information, please visit https://www.compal.com

Media Contact
Jack Wang   Vice president and Spokesperson   +886-2-8797-8588   Investor@compal.com

ZHONGSHAN, China, Oct. 13, 2025 /PRNewswire/ — At the heart of the Guangdong–Hong Kong–Macao Greater Bay Area, Sunwen West Road Pedestrian Street in Zhongshan—a landmark that carries over a century of history—is making a dazzling return, showcasing the city’s unique cultural charm in a brand-new light. Built in the early 20th century, Sunwen West Road has long been a cradle of Xiangshan culture and a hub of commercial prosperity.

Its origins can be traced back to the Southern Song Dynasty (1127-1279), and by the Ming (1368-1644) and Qing (1644-1911) periods, the area had already become a bustling marketplace. In the 1930s, this 500-meter stretch was lined with more than a hundred shops, reflecting the city’s vibrant trading heritage as a hometown of overseas Chinese. In 1995, it was designated Guangdong’s first national-level historic and cultural block. However, over time, the old street faced declining foot traffic and outdated businesses amid rapid urban development.

To restore its vitality, Zhongshan has launched a comprehensive preservation and revitalization project since 2021. Guided by the principle of “restoring the old to its original appearance,” the project team meticulously refurbished every architectural detail of the iconic qilou (arcade-style) buildings—from terrazzo floors and gilded wooden beams to traditional shop signs and original timber structures. The renovation not only preserved the street’s historic charm but also introduced new cultural and creative spaces, heritage workshops, and specialty dining, breathing new life into the century-old street.

On September 28, 2025, Sunwen West Road Pedestrian Street reopened to the public. To capture this historic moment, the film production team employed an innovative one-take drone shot, guiding viewers on a smooth aerial journey through arcade corridors, cultural and creative spaces, and traditional performance scenes—like traveling through time itself. In post-production, game-inspired interactive visual effects were creatively integrated into the live-action footage, allowing younger audiences to experience and appreciate traditional culture in a fresh, immersive way.

This production is not just a technical breakthrough—it is also an innovative experiment in cultural storytelling. It shows international audiences that heritage preservation and modern creativity can blend seamlessly, allowing historic districts to retain their authenticity while embracing new vitality. With the opening of the Shenzhen–Zhongshan Link, the successful revitalization of Sunwen West Road Pedestrian Street will serve as a new window for the world to experience Zhongshan’s rich history and culture, while offering a distinctive “Zhongshan model” for global urban renewal.

Video – https://www.youtube.com/watch?v=gi6rLB_yZEs

Committed to providing solutions based on open standards that create business value through consumer experience innovation

PALO ALTO, Calif., Oct. 13, 2025 /PRNewswire/ — Plume today announced its membership in the prpl Foundation, a leading industry global alliance advancing open-source and open-standard development for next-generation connected devices. As a Gold member, the second highest tier, Plume will collaborate with industry peers to accelerate innovation in cloud-based services and broadband and IoT intelligence for Internet service providers (ISP) and provide exceptional consumer experience.

Plume joins prpl Foundation to Drive Ecosystem Collaboration and Openness

Plume is participating in this week’s Global prpl Summit 2025 in Paris, France, (13-14 October) – the annual gathering of prpl members advancing open standards for the connected home. There, the company will share an early demonstration of their technology now running on the prpl platform – a first step toward delivering its hallmark consumer experience innovations through open standards.

As part of its support of industry standards, Plume is adopting a more open and collaborative approach with its ISP customers and technology partners, meeting them where their technology needs are as they increasingly deploy hardware and services powered by prpl. Plume’s cloud-based platform — with advanced services and APIs for Wi-Fi management, customer care, network intelligence, and security — delivers added intelligence and value to prpl-based deployments.

“Integrating our cloud-based platform with the prpl framework gives us the flexibility to deliver new levels of intelligence across the connected home,” said Plume CTO Chris Griffiths. “It’s an important step toward realizing a more open, software-defined future – one that elevates the consumer experience while enabling our ISP partners to innovate faster.”

“Plume’s membership underscores the continued industry commitment to advancing open standards for the connected home,” said Dr. Leonard Dauphinee, president of prpl Foundation. “We look forward to their participation and contributions to the prpl community as we work together to expand the benefits of openness and interoperability.”

At the NetworkX exhibition in Paris, Plume will also host customer meetings in its private meeting room (Pavilion 7 – Atrium Room 738) and will discuss the future of broadband and IoT experiences, including the responsible adoption of AI, self-healing home networks, and the development of a network intelligence operating system that will define the next era of connectivity and the consumer experience.

About Plume
Plume created the first managed Wi-Fi platform for Internet Service Providers (ISPs) in 2016 and continues to lead today with a best-in-class solution for cloud-managed Wi-Fi, security, and experience. With almost half a billion devices connected to its platform, Plume is a trusted technology partner for more than 400 ISPs around the world, helping them deliver better Wi-Fi experiences and services for subscribers while streamlining their own operations and support systems. Plume leverages OpenSync®, an open-source framework that comes pre-integrated and supported on the leading silicon, CPE, and platform SDKs, and supports leading industry standards like RDK-B and prplWave. This approach allows Plume to offer turnkey and modular solutions, giving ISPs ultimate flexibility, while enabling simpler and faster deployments. Plume is leading the way in providing ISPs an intelligence and innovation edge to stay competitive and adapt to the changing needs of subscribers and the market. Discover more about how Plume is empowering ISPs by visiting https://www.plume.com.

About prpl Foundation
The prpl Foundation is a collaborative open-source community dedicated to open APIs and open industry standards, focusing on carrier-grade software for broadband customer premise equipment as specified by operator members. The prpl Foundation brings together service providers, original equipment manufacturers, system integrators, chip vendors and application developers to harmonize home network architecture, then commonizes a reference implementation of standardized open-source infrastructure built atop open APIs and open industry standards. prpl enables a service delivery ecosystem including innovative third-party applications that can be quickly and easily deployed across different CPE platforms. Learn more about prpl Foundation at https://prplfoundation.org/members/.

SAN JOSE, Calif., Oct. 13, 2025 /PRNewswire/ — MiTAC Computing Technology, a global leader in high-performance and energy-efficient server solutions, is proud to announce its participation at the 2025 OCP Global Summit (October 13–16, San Jose, CA) , Booth C14. Under the theme “From AI Server to Cluster – Open for Growth. Built to Cool.” MiTAC will showcase modular, scalable infrastructures advancing from single AI servers to full clusters, powered by cutting-edge liquid-cooling and sustainable designs.

Teaming up with AMD, Broadcom, CoolIT, Intel, Micron, Murata, NVIDIA, and Solidigm, MiTAC continues to drive open computing innovation and energy-efficient data center evolution.

From Server to Cluster | From Air-Cooled to Liquid-Cooled, From AI to HPC – Showcasing Complete Data Center Solutions

At the summit, MiTAC Computing presents both OCP ORv3 and EIA standard racks, covering next-generation open and enterprise data center architectures.

OCP ORv3 Liquid-Cooled Rack — Supports up to 14 C2811Z5 multi-node servers powered by AMD EPYC™ 9005 series processors. Integrated Murata 33kW Power Shelf MWOCES-211-P-D and CoolIT 200kW CHx200+ In-Rack CDU deliver efficient power and thermal management for high-density deployments. The system’s open, modular design with MiTAC Lake Erie storage modules provides seamless compute, storage, and network integration, enabling an effortless upgrade path from server to cluster.

EIA Air-Cooled Rack — Configured with MiTAC G8825Z5 AI servers using AMD Instinct™ MI350X/MI325X GPUs, this 45U rack integrates a Dell Z9864F-ON switch featuring the Broadcom Tomahawk 5 chipset for 800G high-speed interconnect, along with MiTAC GC68C-B8056 management server and MiTAC TS70A-B8056 storage server. The result is rapid, compatible AI and HPC cluster deployment within existing infrastructure, enabling enterprises to scale efficiently from single server to cluster.

Live Demo: Open Firmware for Transparent and Secure Data Centers

At Booth C14, MiTAC Computing will host a live demonstration featuring OpenBMC and Open Platform Firmware (OPF), developed in collaboration with the Open Source Firmware Foundation, ISVs, and the open hardware community. The demo showcases Redfish-based unified server management, with Coreboot, LinuxBoot, and UEFI options that reduce POST time by up to 50%. It also integrates Broadcom MegaRAID 9560-16i / 9660-16i RAID cards to enhance data reliability and security. MiTAC Computing will further unveil an OPF proof-of-concept (PoC) on AMD EPYC™ 9005/9004 processor platforms, validating real-hardware scalability and transparency for future data center evolution.

After showcasing its cluster-level solutions and open firmware innovations, MiTAC Computing will also present a complete server portfolio spanning AI, HPC, cloud, and enterprise applications, highlighting its ability to scale from single servers to full clusters and deliver end-to-end solutions.

AI Computing Platforms | Liquid Cooling and GPU Acceleration for Large-Scale Training and Generative AI

• G4527G6: A 4U server based on the NVIDIA MGX™ architecture, with dual Intel^® Xeon^® 6767P CPUs, supporting up to 8 NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs and Solidigm D7-P5520 SSDs. Combining GPU parallel computing with high-speed storage, it is designed for deep learning, computer vision, and enterprise AI development.
• G4826Z5: Making its debut at the summit, this high-density liquid-cooled GPU platform supports up to eight AMD Instinct™ MI355X GPUs and AMD EPYC™ 9005/9004 series processors. With 24 DDR5-6400 memory slots and full CPU/GPU liquid cooling, it ensures consistent performance for dense deployments. Equipped with the Broadcom P2200G network adapter, the G4826Z5 delivers high-speed, low-latency connectivity and stands as MiTAC’s flagship liquid-cooled system at this year’s exhibition.

HPC and Cloud Best Practices | OCP-Compliant Platforms for Scalable Workloads

• C2811Z5: An OCP-compliant, liquid-cooled, high-density multi-node server powered by AMD EPYC™ 9005 series processors. Each node supports up to 12 DDR5-6400 memory slots and 3TB capacity, with NVMe E1.S storage paired with Micron 9550 NVMe SSDs, for consistent performance in bandwidth-intensive workloads. Designed for HPC workloads such as simulation and engineering, its liquid cooling boosts thermal efficiency and system reliability.
• Capri 3: An OCP-based cloud server platform offering modularity and flexible expansion, integrated with the Broadcom N1400GD network adapter for high-speed, stable connectivity. Ideal for cloud virtualization, software-defined storage, and data lake architectures, it provides agile scalability for modern data center deployments.

Enterprise Data Processing & Storage | Reliable Performance and Flexible Expansion for Data-Intensive Applications

• R1520G6: A 1U enterprise server powered by Intel^® Xeon^® 6700P processors, equipped with Micron DDR5 DRAM and Micron 6550 ION NVMe SSDs. Optimized for memory-intensive and sustained read/write workloads, it delivers high performance and reliability for demanding enterprise environments.
• R2520G6: A 2U dual-processor server supporting up to 24 NVMe U.2 SSDs, featuring Solidigm D7-PS1010 PCIe 5.0 drives for long-term performance stability. Designed for data storage, big data analytics, and AI data preprocessing, it enables enterprises to turn massive datasets into real-time insights for smarter decision-making.

In addition to its full rack-level solutions and live demonstrations, MiTAC Computing will host two Executive Sessions on October 14 to explore the future of AI clusters and the architectural evolution of sustainable data centers.

MiTAC Computing sincerely invites industry leaders and partners to visit Booth C14 to experience its latest server and cluster solutions firsthand and join the company in exploring the innovative journey “From Server to Cluster.”

For product catalogs and more information, please visit:
MiTAC Computing 2025 OCP Global Summit Landing Page
MiTAC Computing Technology Corporation website: https://www.mitaccomputing.com/
Intel Platform Brochure
AMD Platform Brochure

With optional home-integrated smart panel and up to a month of critical backup power, EcoFlow’s new system delivers fast, reliable home energy backup

SEATTLE, Oct. 13, 2025 /PRNewswire/ — EcoFlow, a leading provider of eco-friendly power solutions, today debuts DELTA Pro Ultra X, the company’s most powerful and easy-to-install whole-home backup power solution. In just seven days from scheduling to installation, whole-home power is now more accessible and practical for households than ever before. Moreover, DELTA Pro Ultra X delivers enhanced performance, scalable storage and smarter home integration, reflecting EcoFlow’s ongoing commitment to improving whole-home power solutions.

Stronger and smarter than ever

DELTA Pro Ultra X delivers 12kW of continuous output, scalable up to 36kW, for a 60% increase over its predecessor, DELTA Pro Ultra. Advanced semiconductors and low-impedance components ensure up to 97% energy conversion efficiency, allowing the system to power high-demand appliances such as water heaters, induction ovens and EV chargers without the need for soft starters. Storage is fully expandable from 12kWh to 180kWh, providing enough power for up to one month of essential home backup.

DELTA Pro Ultra X also comes with an upgraded Smart Home Panel 3, a 200-amp, and a 32-circuit home panel with intelligent energy management, which can be installed as either a main panel or subpanel. Homeowners can monitor and control each circuit remotely via the EcoFlow App, keeping critical circuits running while shutting off non-essential loads and extending critical runtime by up to 42%. The panel also supports smart timer functions that can, for example, run devices during the day and turn them off at night to save energy. It is compatible with previous-generation DELTA Pro Ultra batteries, third-party generators, solar panels and EV chargers, enabling seamless integration into any home setup.

Paired with the Smart Home Panel 3, DELTA Pro Ultra X adapts to daily energy use to reduce electricity costs up to $6,000 per year. The AI-driven Time-of-Use Mode automatically charges and discharges the battery based on real-time pricing and forecasts in the user’s region, storing energy during off-peak hours and avoiding high-peak rates. Compatible with solar, users can harness the sun to save even more. With 10kW solar input, the system can recharge two battery packs to 80% in approximately an hour.

Unlike conventional whole-home power systems in the industry, which typically require months of installation, DELTA Pro Ultra X transcends rooftop limitations by integrating DIY yard solar—delivering faster deployment and greater cost efficiency. Notably, the entire setup process takes just one day, with installation limited to the Smart Home Panel 3. Further enhancing its flexibility, the system features a portable design that enables placement in optimal locations. Its modular, quick-clip expansion architecture also allows users to add storage capacity or accessories without the need for complex rewiring or full reinstallation, streamlining upgrades and adaptability.

Trusted, Nationwide Installer Network

With a long-standing presence in the U.S. and an extensive network of over 425 certified installer partners covering nearly every state, EcoFlow ensures efficient, expert setup for households nationwide. The DELTA Pro Ultra X system can be installed in under seven days from scheduling, with just one day needed for the on-site setup by EcoFlow-certified electricians. Paired with transparent pricing, EcoFlow’s professional service helps households save 80% in time and cost compared to traditional whole-home battery systems. The modular design allows for easy power expansion and relocation, adapting to changing household needs without incurring future installation costs.

“At EcoFlow, our mission is to make reliable, smart energy solutions easy and accessible for every household,” said Brian Essenmacher, North American Head of Business Development for EcoFlow. “With DELTA Pro Ultra X, families can enjoy next-generation power backup that keeps their homes running smoothly, without the usual complexity, time or costs involved. As we continue to grow and cultivate our presence across North America, we’re proud to bring peace of mind and true energy independence to households everywhere.”

Availability

DELTA Pro Ultra X will be available starting Oct. 13 on the EcoFlow website. Customers can purchase the DELTA Pro Ultra X bundle (inverter and two batteries) for $7,999. The Smart Home Panel 3 is available for purchase at $2,999, or as a bundled package with DELTA Pro Ultra X for $10,899 (inverter, Smart Home Panel 3 and two batteries) until Dec. 31.

For more detailed information, please visit EcoFlow US or EcoFlow Canada.

About EcoFlow

EcoFlow is a leading provider of eco-friendly energy solutions, committed to powering a new world. Since its founding in 2017, EcoFlow has aimed to be the FIRST in power solutions — Flexible, Innovative, Reliable, Simple, and Thorough — for individuals and families, whether at home, outdoors or on the go. With a smart manufacturing center in China, and headquarters in the USA, Germany and Japan, EcoFlow has empowered over 5 million users in 140 markets worldwide. For more information, visit https://www.ecoflow.com/us.

SAN JOSE, Calif., Oct. 13, 2025 /PRNewswire/ — PEGATRON, a globally recognized total server solution provider, delivers end-to-end expertise across AI, HPC, cloud, networking, storage, and cluster-scale data center deployments, will present its latest innovations at the OCP Global Summit 2025, highlighting a full portfolio of next-generation platforms designed for AI, HPC, and professional visualization workloads. The showcase features NVIDIA GB300 NVL72, NVIDIA HGX B300, AMD Instinct™ MI355X platforms and NVIDIA RTX PRO servers.

NVIDIA GB300 NVL72 – Accelerating AI Reasoning with Extreme-scale Performance

At the forefront of PEGATRON’s lineup is the RA4802-72N2, featuring the NVIDIA GB300 NVL72 platform with 72 NVIDIA Blackwell Ultra GPUs and 36 NVIDIA Grace CPUs. Featuring 18 compute trays, 9 NVIDIA NVLink switch trays, and 18 NVIDIA BlueField-3 DPUs with NVIDIA ConnectX-8 SuperNICs, the system delivers 130 TB/s aggregate NVLink Switch bandwidth to accelerate large-scale AI training and inference.  PEGATRON has officially begun shipping its RA4802-72N2 system, built on the NVIDIA GB300 NVL72 platform, delivering industry-leading AI performance with streamlined deployment for large-scale data centers.

NVIDIA HGX B300 Platform – Scaling AI from Server to Rack to Cluster

At the core of PEGATRON’s AI reasoning solutions is the NVIDIA HGX B300 platform, accelerated by NVIDIA Blackwell Ultra GPU, which delivers 7X more AI compute than NVIDIA Hopper, enabling next-generation AI reasoning and the most demanding workloads across data centers. The PEGATRON AS402-2T1-8H2 is a powerful 4U liquid-cooled system with 8 NVIDIA Blackwell Ultra GPUs, featuring 2.1 TB HBM3e memory, dual Intel® Xeon® 6 processors, and NVIDIA ConnectX-8 SuperNICs delivering 800 Gb/s ultra-fast networking. For organizations requiring air cooled configurations, the AS801-2T1-8H2 offers an 8U solution with the same GPU and CPU configuration, giving enterprises deployment flexibility while maintaining exceptional performance. Scaling further, the RA4400-64H1 is a 44U liquid-cooled rack solution that integrates 8 individual 4U systems, delivering a total of 64 NVIDIA Blackwell Ultra GPUs and 16 Intel® Xeon® 6 processors for unprecedented compute density.

AMD Instinct™ MI355X Platform – Breakthrough AI Supercomputing with Ultra High-Density 128-GPU per Rack

PEGATRON expands its AMD Instinct™ portfolio with the AS501-4A1-16I1, a high-density liquid-cooled system featuring 4 AMD EPYC™ 9005 processors and 16 AMD Instinct™ MI355X GPUs in a 5OU system, equipped with 288 GB HBM3E memory per GPU and 8 TB/s bandwidth. Scaling up to the RA5100-128I1, an ultra high-density liquid-cooled rack solution with 128 GPUs and 32 CPUs, provides a powerful foundation for AI training, generative AI, HPC, and scientific computing.

NVIDIA RTX PRO Servers – The Ultimate Universal Data Center Computing Platform for Enterprise and Industrial AI

Completing PEGATRON’s portfolio are the two new server platforms powered by  NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. The AS205-2T1 is a compact 2U, 2-socket system featuring 4 NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs (TDP 600W) and dual Intel® Xeon® 6 processors, designed for top-tier performance in space-constrained environments. For higher density and scalability, the AS400-2A1-CX8 is a powerful 4U, 2-socket system with 8 NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs (TDP 600W), dual AMD EPYC™ 9005 processors, and NVIDIA ConnectX-8 SuperNICs with PCIe Gen 6 switch for maximum throughput. Together, these systems integrate NVIDIA Blackwell GPUs to deliver cutting-edge AI performance, photorealistic rendering, generative AI, digital twin simulations, and complex industrial workflows.

“AI is transforming every industry, and the demand for advanced infrastructure is accelerating rapidly,” said May Wang, General Manager of PEGATRON Server Business. “PEGATRON is ready for the AI-driven future with platforms engineered for next-generation performance and proven in real-world deployments, enabling enterprises to scale seamlessly from server efficiency to cluster-scale supercomputing.”

PEGATRON continues to push the boundaries of AI and HPC infrastructure, delivering platforms that combine performance, efficiency, and scalability. We invite industry leaders, partners, and innovators to visit us at the OCP Global Summit 2025, Booth A33, and experience how PEGATRON is powering the future of intelligent data centers.

For more information, please visit PEGATRON SVR website and follow us on LinkedIn and YouTube.

https://svr.pegatroncorp.com 

https://www.linkedin.com/showcase/pegatron-svr/?originalSubdomain=tw

https://www.youtube.com/@pegatroncorp.6158

About PEGATRON

PEGATRON Corporation (hereafter referred to as “PEGATRON”), with abundant product development experience and vertically integrated manufacturing, is committed to providing clients with innovative design, systematic production, and manufacturing services to comprehensively and efficiently satisfy all our customers’ needs.

Drawing on accumulated experience in server design, manufacturing, and deployment, PEGATRON focuses on developing a variety of state-of-the-art servers including liquid-cooled/air-cooled server solutions based on x86 and ARM architectures, racks, and AI clusters, that meet the requirements of present and future Cloud Service Providers’ data centers, as well as enterprise-grade data centers.  

PEGATRON Corporation Website: https://www.pegatroncorp.com/