SiFive Bumps Up Performance for RISC-V HCP, AI Systems

SiFive Bumps Up Performance for RISC-V HCP, AI Systems

The News: SiFive has launched two new reduced instruction set computer (RISC-V) systems, the Performance P870 and Intelligence X390, designed to address requirements for high performance compute through low power, compute density, and vector compute capability. You can read the news release at the SiFive website.

SiFive Bumps Up Performance for RISC-v HCP, AI Systems

Analyst Take: SiFive, founded by the inventors of the RISC-V open standard, who came from the University of California, Berkeley, has set out to bring commercial grade RISC-V microprocessor intellectual property (IP) to market. The company employs more than 500 engineers.

SiFive’s RISC-V Market Strategy

Startup SiFive is considered a “unicorn” with a valuation of more than $250 billion after it raised more than $350 million in venture capital. However, SiFive competes in the semiconductor market against entrenched incumbents such as Intel, AMD, and ARM. The semiconductor market has been a tough place for newcomers to survive. SiFive aims to disrupt traditional models by offering a more flexible and customizable solution, accelerating the democratization of chip design and potentially reshaping the industry.

SiFive needs to produce innovative solutions for AI and machine learning (ML) to deliver increased performance per watt efficiency and better overall end-user experiences. This requires differentiating in areas such as vector computing for AI and high-performance applications processors. SiFive’s early push has been in the automotive, aerospace, and defense industries but it is looking to expand into more markets with its new upgraded systems.

The SiFive Performance P870 and SiFive Intelligence X390 are engineered for high performance compute across markets including consumer, automotive, and infrastructure sectors. When used in tandem, the products offer a combination of scalar and vector compute capabilities, suitable for data-intensive operations and AI applications.

The Performance P870 is SiFive’s third-generation applications processor. It is aimed at high-performance consumer applications or for use with a vector processor in the data center. It brings a 50% peak single thread performance upgrade (specINT2k6) over the previous generation SiFive Performance processors. The P870 is a six-wide out-of-order core that meets the RVA 23 profile and offers a shared cluster cache enabling up to a 32-core cluster. The P870 is fully compatible with Google’s platform requirements for Android on RISC-V.

The SiFive Intelligence X390 builds on the Intelligence X280 used by Google, NASA, and other organizations, which couples AI/ML applications with hardware accelerators in mobile, infrastructure, and automotive applications. SiFive claims the X390 brings a 4x improvement to vector computation with its single core configuration, doubled vector length, and dual vector arithmetic-logic units (ALUs). This enables quadruple the amount of sustained data bandwidth. The SiFive Vector Coprocessor Interface eXtension (VCIX) allows customers to add their own vector instructions and/or acceleration hardware to increase performance with custom instructions.

Where SiFive Fits in the Chip Landscape

SiFive, Intel, ARM, and AMD serve as critical entities within the semiconductor landscape but operate under different paradigms and target different market niches. Intel and AMD primarily use the x86 architecture, focusing on high-performance central processing units (CPUs) for general-purpose computing needs in consumer, enterprise, and data center markets. ARM takes a different approach, licensing its power-efficient architecture to other firms; it has a strong presence in mobile and embedded systems. SiFive employs the open-source RISC-V architecture, allowing for high customization and adaptability for specific use-cases such as IoT, edge computing, high-performance computing (HPC), and AI.

The business models also vary considerably among these players. Intel and AMD manufacture and sell their chips directly, while ARM and SiFive license their designs to other companies. However, SiFive distinguishes itself by leveraging an open-source architecture, making the scope for innovation more democratized. In terms of strategic implications, Intel and AMD need to evolve their legacy architectures continually to stay competitive, especially given ARM’s inroads into the server market. ARM’s $4.9 billion initial public offering (IPO)―the biggest in tech since 2019―shows the value of this market.

Disclosure: The Futurum Group is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.

Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of The Futurum Group as a whole.

Other insights from The Futurum Group:

ARM Goes Public with Successful Debut

Intel’s FPGA Unit Becoming a Separate Business To Drive Growth

AMD Datacenter & AI Conference Recap

Author Information

Steven engages with the world’s largest technology brands to explore new operating models and how they drive innovation and competitive edge.

Dave focuses on the rapidly evolving integrated infrastructure and cloud storage markets.

Related Insights
Lakebase and LTAP Challenge Database Orthodoxy, Are Monoliths Finally Obsolete?
July 2, 2026

Lakebase and LTAP Challenge Database Orthodoxy, Are Monoliths Finally Obsolete?

Databricks revolutionizes analytical platforms through Lakebase and LTAP, unifying transactional and analytical workloads. Research shows 73.6% of organizations are increasing spend, signaling a major shift from legacy databases....
NVIDIA Jetson in Lunar Orbit Signals Commercial GPUs Are Ready for Spaceflight
July 1, 2026

NVIDIA Jetson in Lunar Orbit Signals Commercial GPUs Are Ready for Spaceflight

Brendan Burke, Research Director at Futurum, analyzes how Firefly Aerospace's deployment of NVIDIA Jetson in lunar orbit proves commercial GPUs now support demanding long-duration spaceflight missions....
Applied Materials' Master Class Schools Memory Makers on Logic-Class Fabrication
June 30, 2026

Applied Materials’ Master Class Schools Memory Makers on Logic-Class Fabrication

Brendan Burke, Research Director at Futurum, frames Applied Materials' advanced packaging systems for AI-scale HBM and 3D stacking as a logic-memory convergence that could double DRAM equipment spending....
Will SCE’s Wildfire Recovery Program Set a New Standard for Utility Crisis Response?
June 30, 2026

Will SCE’s Wildfire Recovery Program Set a New Standard for Utility Crisis Response?

Southern California Edison's $700M wildfire compensation program reveals why utilities must adopt enterprise AI for claims processing, customer support automation, and workflow orchestration at scale during disaster recovery....
Microsoft Unifies CX on a Single Dynamics 365 Data Model
June 29, 2026

Microsoft Unifies CX on a Single Dynamics 365 Data Model

Keith Kirkpatrick, Vice President & Research Director, Enterprise Software & Di at Futurum, analyzes how Microsoft's integration of agentic AI into Dynamics 365 unifies customer service, contact center operations, and...
Qualcomm’s Investor Day 2026 Agentic and AI Inference To Drive 2x Revenue Growth by 2030
June 29, 2026

Qualcomm’s Investor Day 2026: Agentic and AI Inference To Drive 2x Revenue Growth by 2030

Olivier Blanchard and Brendan Burke, Research Directors at Futurum, explain the significance of Qualcomm's June 24 Investor Day announcements as the company continues to evolve into a full-stack AI platform...

Book a Demo

Newsletter Sign-up Form

Get important insights straight to your inbox, receive first looks at eBooks, exclusive event invitations, custom content, and more. We promise not to spam you or sell your name to anyone. You can always unsubscribe at any time.

All fields are required






Thank you, we received your request, a member of our team will be in contact with you.