Altera and Arm have announced an expansion of their long-standing collaboration to deliver programmable, high-efficiency compute solutions tailored for next-generation AI data centers.
The enhanced partnership builds on more than two decades of joint development across embedded, industrial, and communications markets. The latest phase extends this collaboration into data center infrastructure by integrating Altera’s field-programmable gate array (FPGA) technologies with Arm’s AGI CPU, based on the Arm Neoverse CSS V3 platform. The combined solution is designed to enable system architects to develop low-latency, scalable, and highly adaptable compute platforms optimized for AI-driven workloads.
As AI applications continue to evolve, data centers are increasingly requiring specialized compute architectures capable of handling diverse and dynamic workloads. Traditional CPU- and GPU-centric designs are being complemented by programmable acceleration technologies such as FPGAs, which offer flexibility and deterministic performance for tasks including data pre-processing, networking, and AI inference orchestration.
Altera’s FPGA solutions are already widely deployed across data centers in formats such as PCIe accelerator cards, SmartNICs, and data processing units (DPUs). These deployment models allow programmable acceleration to be positioned close to the data path, enabling improved latency, enhanced security, and faster deployment cycles.
The integration with Arm’s AGI CPU is expected to further extend these capabilities by combining efficient general-purpose compute with programmable acceleration. This approach supports heterogeneous computing architectures, which are increasingly seen as essential for managing the complexity and scale of modern AI workloads.
According to Arm, its AGI CPU provides an energy-efficient compute foundation for AI infrastructure, while collaborations with ecosystem partners such as Altera help broaden its applicability across high-performance environments. The joint solution aims to address key data center requirements, including real-time processing, workload adaptability, and scalability.
Altera’s leadership emphasized that the expanded collaboration represents a strategic step toward enabling a new class of compute platforms designed specifically for AI data centers. By combining its established footprint in FPGA-based infrastructure with Arm’s advanced CPU architecture, the companies aim to meet growing demand for flexible and performance-optimized solutions.
