Broadcom Unveils Industry’s First 2nm Custom Compute SoC on 3.5D XDSiP Platform
Broadcom Inc. (NASDAQ: AVGO), a global leader in semiconductor and infrastructure software solutions, has announced the shipment of the industry’s first 2 nanometer (nm) custom compute System-on-Chip (SoC) built on its advanced 3.5D eXtreme Dimension System in Package (XDSiP) platform. This milestone marks a significant leap in semiconductor integration, combining cutting-edge 2.5D and 3D-IC technologies through Face-to-Face (F2F) integration.
Revolutionizing Next-Generation XPU Design
The 3.5D XDSiP platform is engineered to serve as the foundation for next-generation XPUs, which are critical for powering artificial intelligence (AI) and high-performance computing (HPC) workloads. By leveraging modular, multi-dimensional stacked die architecture, Broadcom’s XDSiP enables consumer AI customers to deliver advanced XPUs with exceptional signal density, superior power efficiency, and ultra-low latency. These features are essential for meeting the immense computational requirements of gigawatt-scale AI clusters.
One of the key advantages of the XDSiP platform is its ability to allow compute, memory, and network I/O to scale independently within a compact form factor. This flexibility supports high-efficiency, low-power computing at scale, making it ideal for data centers and AI infrastructure where performance and energy efficiency are paramount.
Industry Collaboration and Technology Leadership
Broadcom’s first 3.5D custom compute SoC has been delivered to Fujitsu, highlighting the company’s commitment to innovation and execution in advanced semiconductor technologies. Frank Ostojic, Senior Vice President and General Manager of Broadcom’s ASIC Products Division, emphasized the significance of this achievement, noting that since the introduction of the 3.5D XDSiP platform in 2024, Broadcom has expanded its capabilities to support a broader range of XPUs. These advancements are set to benefit a wider customer base with shipments expected from the second half of 2026.
Naoki Shinjo, Senior Vice President and Head of Advanced Technology Development Unit at Fujitsu, described the launch as a transformative milestone for advanced semiconductor integration. By combining 2nm process technology with Face-to-Face 3D integration, the platform delivers unprecedented compute density and energy efficiency—key attributes for the next era of AI and HPC. This innovation is central to Fujitsu’s FUJITSU-MONAKA initiative, which aims to deliver high-performance, low-power processors for scalable and sustainable AI-driven solutions.
Enabling the Future of AI and High-Performance Computing
Broadcom’s 3.5D XDSiP platform sets a new standard for custom compute SoCs, enabling transformative breakthroughs in AI and high-performance computing. By pushing the boundaries of semiconductor integration and efficiency, Broadcom continues to drive the evolution of next-generation processors that will power the future of AI, data centers, and advanced computing applications.
