Samsung, AMD Deepen 5G/6G AI Pact
Samsung and AMD are reinforcing their strategic collaboration to advance AI-powered 5G and 6G network technologies. The partnership focuses on integrating AMD's compute capabilities with Samsung's radio access network (RAN) solutions for upcoming commercial deployments. This highlights the growing convergence of AI, high-performance computing, and next-generation telecom infrastructure.
This expanded partnership is moving from the lab to live commercial deployments, with Canadian carrier Videotron selecting Samsung's 5G and 4G core gateway solutions running on AMD EPYC 9005 Series CPUs. The deal signifies a major step in deploying AI-powered, virtualized network architectures in real-world environments. At the core of the collaboration is the validation of Samsung's AI-powered virtualized Radio Access Network (vRAN) running on AMD EPYC processors without the need for additional hardware accelerators. Multi-cell testing at Samsung's R&D lab confirmed the setup could achieve commercial-grade performance, a significant proof point for the viability of software-driven, open network architectures that reduce dependency on proprietary hardware. This CPU-centric approach to AI-RAN directly challenges the narrative that powerful GPUs are essential for AI workloads in telecommunications infrastructure. Competitors like Ericsson are focusing on their own custom silicon (ASICs) with integrated neural network accelerators, also aiming to avoid GPU dependency. Meanwhile, Nokia has notably embraced a GPU-centric strategy through a major partnership with NVIDIA. The move toward virtualized, software-defined networks allows operators greater flexibility and the ability to mix and match components from different vendors, breaking the traditional single-vendor lock-in model. For AMD, this deepens the impact of its 2022 acquisition of Xilinx, whose Zynq UltraScale+ RFSoC (Radio Frequency System-on-Chip) platform is critical for 5G radio and Massive MIMO deployments. Looking ahead, this collaboration is foundational for 6G, which is envisioned as an "AI-native" network from the ground up. Unlike 5G, where AI is largely applied to optimize existing infrastructure, 6G aims to embed AI and machine learning into every layer of the network fabric for self-optimizing, real-time automation. For enterprise customers, Samsung is leveraging the partnership to showcase a "Network in a Server" (NIS) solution, a fully virtualized edge-AI platform powered by AMD CPUs. This has been verified with a major Japanese operator for use cases like video analysis and services based on Integrated Sensing and Communication (ISAC) technology.
