SiFive Unveils 2nd Gen Intelligence™ IP: RISC-V Cores Unite Scalar, Vector & Matrix Compute to Power AI from Far Edge to Data Center
SiFive, the pioneer in commercial RISC-V processor IP, today launched its next-generation Intelligence™ product family — a bold expansion of five AI-optimized RISC-V cores engineered to meet surging global demand for efficient, scalable compute across the entire AI spectrum. Anchored by two all-new additions — the X160 Gen 2 and X180 Gen 2 — and enhanced versions of the X280, X390 and XM series, this 2nd Generation lineup delivers unprecedented integration of scalar, vector, and (in the case of XM) matrix processing capabilities — all within a unified, configurable RISC-V architecture.
Designed for the realities of modern AI workloads — from deeply embedded IoT sensors to high-performance data center accelerators — these new cores give developers the flexibility to match silicon to application. The X160 Gen 2 and X180 Gen 2 are purpose-built for the “far edge” — think smart sensors, autonomous robots, automotive ECUs, and industrial controllers — where power and die area are at a premium. Despite their compact footprint, they deliver advanced AI inference capabilities previously reserved for larger systems, enabling real-time decision-making where it matters most.
“AI is catalyzing the next era of the RISC-V revolution,” said Patrick Little, CEO of SiFive. “We’re not just keeping pace — we’re setting it. Already, two Tier 1 U.S. semiconductor companies have adopted the new X100 series. With Gen 2 Intelligence IP, we’re adding configurability, performance boosts, and critical new features to accelerate our customers’ time-to-market and unlock AI innovation across verticals.”
Why Now? The AI Edge Explosion
The timing couldn’t be better. According to Deloitte, AI workloads are projected to grow by at least 20% across all computing environments — with edge AI seeing a staggering 78% surge. As AI moves out of the cloud and into devices, the need for efficient, adaptable, and licensable IP becomes critical.
“SiFive is strategically investing in the white space of edge AI compute — where innovation has lagged behind demand,” said Pat Moorhead, Founder and Chief Analyst at Moor Insights & Strategy. “Their Gen 2 products introduce novel architectural approaches that solve real-world efficiency and performance bottlenecks. This isn’t incremental — it’s foundational.”
The Vector Advantage: Efficiency Meets Performance
At the heart of SiFive’s AI acceleration is its advanced vector processing engine. Unlike traditional scalar CPUs that process one data element per instruction, vector engines crunch multiple data points simultaneously — slashing instruction overhead, reducing power consumption, and accelerating neural network operations. For edge applications where every microwatt counts, this means running complex AI models faster, cooler, and in a smaller silicon footprint.
But SiFive doesn’t stop at vectors. The XM Gen 2 core introduces scalable matrix compute — essential for accelerating transformer models, computer vision, and next-gen AI workloads that rely on dense linear algebra. Together, scalar, vector, and matrix units form a cohesive compute fabric optimized for AI’s evolving demands.
One Architecture, Infinite Configurations
What sets SiFive apart is architectural unity. All five Gen 2 cores — from the tiny X160 to the powerhouse XM — share the same RISC-V ISA, toolchain, and programming model. This allows developers to scale seamlessly across applications without rewriting software or retraining teams. Whether you’re building a battery-powered smart lock or a data center inference server, SiFive’s IP scales with you.
Beyond Compute: Accelerator Control Made Simple
SiFive’s X-Series cores also double as Accelerator Control Units (ACUs), managing external AI accelerators via standardized interfaces like SSCI and VCIX. This offloads control-plane complexity from the accelerator itself, letting customers focus on differentiating their data-plane innovations. The result? Faster development cycles, simplified software stacks, and more robust system integration.
Real-World Impact Across Industries
Automotive OEMs can deploy X180 Gen 2 cores in radar and vision systems for low-latency object detection. Industrial automation vendors can embed X160 Gen 2 into predictive maintenance sensors. Data center architects can pair XM Gen 2 with custom accelerators for high-throughput LLM inference. The use cases are as diverse as the markets SiFive serves.
With RISC-V adoption accelerating and AI becoming ubiquitous, SiFive’s 2nd Gen Intelligence IP arrives at the perfect inflection point — empowering innovators to build smarter, leaner, and more powerful AI solutions from the sensor to the server.
About SiFive
SiFive is the leader in RISC-V computing, delivering the custom silicon foundation that AI demands. As the gold standard for RISC-V CPUs, SiFive combines the flexibility of open standards with unmatched performance, enabling scalable AI solutions from the edge to the data center. With the broadest portfolio, industry-leading IP, and deep system expertise, SiFive helps the world’s leading technology companies rapidly build differentiated AI hardware—faster, at lower cost, and without compromise.
