Expanding its RISC-V artificial intelligence portfolio to handle escalating workloads, SiFive makes a move.

Expansion of SiFive's RISC-V Intelligence Family to Cater to Growing AI Tasks

In the rapidly evolving world of artificial intelligence (AI), SiFive has made a significant stride with the announcement of its second generation Intelligence Family. This new family of cores, including the X160 Gen 2, X180 Gen 2, upgraded versions of X280, X390, and XM, is designed to cater to a wide range of AI, machine learning (ML), and edge computing applications.

The X390, for instance, is tailored for industrial edge and mobile applications, while the XM Gen 2 can scale to thousands of tera operations per second (TOPS), making it suitable for data-center-class generative AI workloads. The XM Gen 2 also delivers wide vector and matrix compute capabilities, which are tuned for large language models and other intensive applications.

SiFive's announcement coincides with the AI Infrastructure Summit in Santa Clara, California, taking place on September 9-11, where the company can be found at booth #908. The company has integrated its software stack with popular AI runtimes and libraries such as TensorFlow Lite, ONNX Runtime, and Llama.cpp, further broadening its appeal.

Research suggests that AI is accelerating the adoption of RISC-V, with shipments of RISC-V AI processors forecast to climb rapidly over the next several years. Deloitte data points to a 20% increase in AI workloads across all computing domains, with edge workloads projected to rise nearly 80%.

SiFive's vector-enabled RISC-V architecture processes multiple data items in parallel, reducing instruction overhead and power consumption. All five of the new products from SiFive's second generation Intelligence Family are available for licensing immediately.

Despite the excitement surrounding SiFive's new offerings, there is no publicly available, comprehensive list of companies currently using these cores or in the licensing process (pre-release). Typical users are semiconductor manufacturers, SoC licensees, and system integrators with a focus on AI, ML, and edge computing. For current, specific information, it is recommended to contact SiFive directly or to check their official communication channels.

In the AI race, success for SiFive will hinge on execution, particularly in aligning with software frameworks and proving scalability in production silicon, expected in Q2 2026. The company's unique proposition lies in its approach, which pairs efficient scalar/vector compute with flexible accelerator control interfaces, as demonstrated by the introduction of new interfaces, SSCI and VCIX, to enable advanced integration with AI and other accelerators, allowing their cores to act as Accelerator Control Units (ACUs). The X160 Gen 2, optimized for IoT and edge compute, features a small footprint and energy efficiency, making it a strong contender in its segment.

Stay tuned for more updates from SiFive as they continue to make strides in the AI and edge computing market.