Paris, June 12, 2025 — The relentless march of artificial intelligence continues to reshape industries, but its true potential remains tethered to the underlying infrastructure that powers it.
As GTC Paris 2025 convenes, the spotlight falls not just on groundbreaking algorithms, but on the very hardware enabling them to thrive.
Amidst the buzz, PEGATRON, a global design, manufacturing, and services (DMS) provider, has emerged as a pivotal player, unveiling a suite of AI server solutions poised to redefine enterprise-scale AI.
These aren’t merely incremental upgrades; they represent a foundational shift, built upon the formidable NVIDIA Blackwell architecture, designed to push the boundaries of AI training, inference, and deployment.
The sheer computational demands of modern AI, particularly for large language models and complex simulations, necessitate an infrastructure that is not only powerful but also incredibly efficient and resilient.
PEGATRON’s latest offerings at GTC Paris address these critical needs head-on, showcasing a deep understanding of the challenges faced by cloud service providers and enterprise data centers alike.
At the vanguard of their showcase is the RA4802-72N2, a marvel of engineering anchored by the NVIDIA GB300 NVL72 rack system.
This behemoth houses an astonishing 72 NVIDIA Blackwell Ultra GPUs and 36 NVIDIA Grace CPUs, a configuration designed explicitly for what PEGATRON terms “AI factories.”
Imagine the implications: a system capable of multiplying AI production by up to 50 times.
This isn’t just about faster computations; it’s about accelerating discovery, dramatically shortening development cycles for new AI models, and enabling entirely new paradigms of AI-driven research and application.
Crucially, such immense power generates prodigious heat.
PEGATRON’s answer is an internally developed Coolant Distribution Unit (CDU), boasting a robust 310 kW cooling capacity with hot-swappable redundant pumps.
This bespoke cooling solution is a testament to PEGATRON’s integrated approach, ensuring not only peak performance but also the unwavering reliability essential for mission-critical AI workloads.
It’s a recognition that the future of AI isn’t just about processing power, but about the sustainable and reliable delivery of that power.
Beyond the colossal scale of the “AI factory” solution, PEGATRON is also catering to the burgeoning demand for high-performance AI reasoning and real-time applications.
The AS208-2A1, a compact 2U liquid-cooled GPU server, integrates the NVIDIA HGX B300 system with dual AMD EPYCTM 9005 CPUs.
This system epitomizes efficiency through its direct liquid cooling design, a trend gaining significant traction in high-density data centers due to its superior heat dissipation and energy efficiency compared to traditional air cooling.
Scalable within an NVIDIA MGX 48U rack to support up to 128 GPUs and 32 CPUs, the AS208-2A1 is meticulously optimized for the AI reasoning era.
This means it excels at handling complex tasks like agent systems, real-time inference, and video generation – the very applications that bring AI to life, enabling intelligent automation and immersive digital experiences.
Its high-efficiency cooling allows sustained performance under intense computational loads, crucial for applications where latency is paramount.
Recognizing that not all data centers are ready for a full transition to liquid cooling, PEGATRON wisely offers air-cooled alternatives.
Their 8U air-cooled GPU server, also featuring NVIDIA HGX B300 systems, provides flexibility without compromising on computational muscle.
Available with dual Intel® Xeon® 6 processors (AS801-2T1) or AMD EPYCTM 9005 GPUs (AS801-2A1), these systems offer robust performance for AI model training, scientific research, and advanced data analytics, ensuring that a broad spectrum of enterprise needs can be met.
For the diverse landscape of enterprise AI workloads, PEGATRON introduces the AS400-2A1.
This 4U RTX PRO server is equipped with 8 NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs and dual AMD EPYCTM 9005 CPUs.
Its versatility makes it a workhorse for everything from multimodal inference and physical AI to sophisticated design applications, scientific computing, and high-fidelity graphics and video processing.
A significant enhancement here is its compatibility with the NVIDIA MGX PCIe switch board featuring ConnectX-8 SuperNIC, enabling ultra-fast 800 Gbps networking.
In the world of distributed AI, where vast datasets must move seamlessly between GPUs, such high-bandwidth, low-latency networking is not just an advantage; it’s a necessity for achieving optimal performance, energy efficiency, and cost-effectiveness.
May Wang, Vice President of PEGATRON, aptly summarized the company’s vision: “As AI continues to transform industries, the need for powerful, scalable, and energy-efficient infrastructure is more critical than ever.”
Our new platforms, based on the NVIDIA Blackwell Ultra and RTX PRO architectures, empower organizations to achieve new levels of performance, reliability, and productivity.
Her words underscore PEGATRON’s understanding of the strategic imperative behind robust AI infrastructure.
It’s not just about raw power, but about delivering that power reliably and sustainably, enabling businesses to unlock the true value of their AI investments.
Founded on January 1, 2008, PEGATRON Corporation has built a formidable reputation as a DMS provider with extensive experience in product development and vertically integrated manufacturing.
This deep expertise in design and production, particularly in server manufacturing, has positioned them uniquely to address the evolving demands of the AI era.
Their current focus on developing cutting-edge servers for both current and future cloud service providers and enterprise data centers is a natural evolution, leveraging their core strengths to meet the computational challenges of tomorrow.
Visitors to GTC Paris 2025 can explore these innovations firsthand at PEGATRON’s booth G10, located at Paris Expo Porte de Versailles, through June 12.
For those seeking to peer into the future of AI infrastructure, PEGATRON’s showcase offers a compelling glimpse into the hardware backbone that will power the next wave of intelligent transformation.
A new documentary excavates the overlooked role of a Holocaust survivor who became a pivotal intelligence asset during the 1976 Entebbe hostage crisis.
G7 nations are committing $8 trillion to military expansion, creating a surge in demand for critical minerals that mirrors the growth of the artificial intelligence sector.
A deep dive into the corporate histories and legal records of Krishan Rattan and Ankiti Bose reveals a complex reality behind their high-profile investment platform.
