Compal expands AI and HPC server lineup with NVIDIA MGX-based platforms at GTC 2025

Compal Electronics (Stock Ticker: 2324.TW) has introduced a trio of next-generation AI and HPC server platforms at GTC 2025, signaling its commitment to high-performance computing. The SX420-2A, SX220-1N, and SX224-2A are built on the NVIDIA MGX architecture, designed to cater to increasing enterprise demand for AI workloads, large-scale simulations, and high-load computing applications. By integrating NVIDIA’s latest GPU technology, Compal aims to enhance compute power efficiency, scalability, and AI-optimized performance in data centers worldwide.

What Sets Compal’s New AI-HPC Servers Apart?

Compal’s newly launched servers offer enterprise-grade AI acceleration, each tailored for different workload requirements. The SX420-2A 4U AI Server is designed for deep AI-HPC applications, incorporating a flexible rack design that supports both EIA 19” and ORV3 21” configurations. The system can house up to eight RTX PRO 6000 Blackwell GPUs, enabling high-throughput AI computing and deep learning model training at an unprecedented scale.

The SX224-2A 2U AI Server, which pairs NVIDIA MGX architecture with an AMD x86 platform, is engineered for versatile computing environments, supporting both AI-HPC and AI-graphics processing. It allows businesses to customize configurations for workload-specific performance optimizations, ensuring long-term compatibility with evolving AI infrastructure.

At the core of both models is the RTX PRO 6000 Blackwell GPU, which boasts 96GB of ultra-fast GDDR7 memory and a passively cooled thermal design to maintain stable performance under extreme loads. Designed for agentic AI, scientific computing, and high-end video rendering, this GPU-powered server architecture ensures sustained efficiency in AI and HPC computing.

How Does SX220-1N Improve Large-Scale AI Model Training?

The SX220-1N 2U AI Server is built to address the needs of giant-scale AI model training and high-performance computing. It features the NVIDIA GH200 Grace Hopper Superchip, a powerful solution that integrates GPU and CPU performance to enhance data processing speed.

This platform incorporates NVIDIA NVLink-C2C technology, a next-generation interconnect that creates a coherent memory pool across multiple GPUs and CPUs, significantly increasing bandwidth and memory efficiency. The result is an optimized AI server environment that enables enterprises to train large-scale AI models faster, improve data transfer speeds, and accelerate high-load computational tasks.

With growing demand for AI-driven solutions in fields such as biotechnology, financial modeling, and autonomous systems, Compal’s SX220-1N provides a future-proof computing infrastructure that can evolve with emerging enterprise AI and HPC needs.

Why Is Compal Betting on NVIDIA MGX Architecture for Future AI-HPC Solutions?

The NVIDIA MGX architecture is a scalable and modular computing framework designed to maximize GPU-driven performance in AI workloads. As AI applications become more complex, traditional computing architectures struggle to keep up with the growing demand for higher processing power.

By adopting NVIDIA MGX-based designs, Compal is positioning itself as a key player in the AI-HPC computing market, offering customizable server solutions that align with the future of AI-driven industries. The integration of modular computing components ensures that businesses can upgrade server configurations without investing in entirely new infrastructure, reducing long-term operational costs while maintaining cutting-edge AI performance.

How Is Compal’s Stock Performance Reacting to AI-HPC Expansion?

While Compal Electronics’ latest AI-HPC innovations have generated significant industry interest, the company’s stock performance remains under investor scrutiny. As of March 19, 2025, Compal Electronics (2324.TW) is trading at NT$37.10, marking a 3.76% decline from its previous close of NT$38.55.

Despite the recent downturn, Compal’s stock has seen a year-to-date gain of 3.79%, reflecting long-term investor confidence in its strategic focus on AI-HPC computing. Analysts maintain a neutral rating, with a hold recommendation and an average one-year price target of NT$38.56, suggesting limited near-term upside.

Fundamentally, Compal has delivered consistent earnings growth, with EPS increasing at an annual rate of 7.6% over the past five years. The company recently announced an increase in its annual dividend to NT$1.40 per share, providing a dividend yield of approximately 3.67%. Given its strong fundamentals and expansion into AI-HPC, analysts suggest that Compal remains a solid long-term investment, though short-term stock movements may be influenced by broader market trends and competition within the AI server sector.

What’s Next for Compal’s AI-HPC Business?

Compal’s continued push into high-performance AI computing is expected to play a pivotal role in shaping enterprise data center strategies. As demand for AI-optimized server infrastructure grows, the company is well-positioned to capitalize on next-generation AI developments.

Investors and industry leaders will closely watch Compal’s upcoming earnings report on March 20, 2025, which will provide insights into its AI-HPC revenue performance and future growth trajectory. While competitive pressures remain a challenge, Compal’s strategic alignment with NVIDIA’s AI ecosystem reinforces its position as a leading provider of enterprise AI solutions.

As the company showcases its latest AI-HPC innovations at GTC 2025, businesses evaluating scalable AI solutions can visit Compal’s Booth (#1809) to explore how these new server platforms are redefining AI-driven computing.