AI Computing Infrastructure, Built for Scale
In 2025, Foxlink Group further expanded its investment in AI supercomputing infrastructure through DULINK, reinforcing its long-term commitment to next-generation accelerated computing.
As a Foxlink Group subsidiary, DULINK draws on the Group’s global industrial footprint, technology capabilities, and experience in large-scale operations. It is designed to support demanding workloads across AI model training, inference, and high-performance computing.
By integrating accelerated compute, high-speed networking, scalable storage, and enterprise-oriented operations, DULINK provides a scalable foundation for advanced AI development and deployment.
Energy efficiency and sustainability are central to DULINK’s long-term vision. Through more efficient computing, power, cooling, and resource utilization, DULINK aims to support responsible AI innovation at scale.
Latest GPU clusters with ultra-fast interconnect technology.
Elastic infrastructure that grows with you.
Enterprise-grade security and 99.9% availability.
Supported by a leading global manufacturer.
DULINK provides the computing foundation required to develop, scale, and operate increasingly complex AI and data-intensive workloads.
High-density GPU computing infrastructure designed for parallel processing and large-scale computational workloads.
Scalable computing capacity for foundation models, generative AI, multimodal models, fine-tuning, and other training-intensive applications.
Infrastructure designed to support high-throughput inference and the deployment of AI models in production environments.
Private cloud and on-premises solutions for enterprises.
From the hardware to the software, we provide a complete stack to accelerate your AI journey.
High-density GPU clusters, high-speed interconnect, scalable storage, power and liquid cooling.
Kubernetes-based scheduling, multi-tenant isolation and elastic allocation of accelerated resources.
Training and inference frameworks, distributed runtimes, libraries and optimized model toolchains.
Managed endpoints, routing, authentication and usage governance for model and data services.
Copilots, assistants, search and domain applications delivered to end users and business systems.
End-to-end security and compliance
Scale up or down in seconds
Maximize compute and efficiency
Expert support whenever required
DULINK welcomes collaboration with technology and ecosystem partners to scale accelerated computing infrastructure and support the growing demand for advanced AI.
