SiFive raises $400M

SiFive closed a $400 million funding round led in part by Nvidia to build out RISC‑V chips aimed at the data‑centre market, positioning itself as a challenger to Arm. The financing signals continued fragmentation and competition in the data‑centre silicon layer rather than consolidation into a quiet oligopoly (techstartups.com).

SiFive just pulled in a $400 million funding round, and Nvidia is one of the names in it. The company said on April 9 that the round was oversubscribed and valued SiFive at $3.65 billion. (sifive.com) This is a fight over the basic language a processor speaks. In chip design, that language is called an instruction set architecture, and RISC-V is an open standard that anyone can build on. (riscv.org) Arm sells access to its chip building blocks through licenses. RISC-V starts from a different place, because the core instruction set is open and companies can customize around it without starting from a proprietary rulebook. (arm.com) (riscv.org) SiFive sits in the middle of that shift because it does not mainly sell finished chips. It sells processor core intellectual property, which means blueprints other companies can plug into their own silicon the way a carmaker buys an engine design instead of machining every part from scratch. (sifive.com) (reuters.com) The company was founded in 2015 by Krste Asanović, Yunsup Lee, and Andrew Waterman, the University of California, Berkeley researchers tied to the early RISC-V project. That link matters because SiFive has spent a decade turning a university architecture into something big customers can actually license and ship. (sifive.com) (tracxn.com) The target now is the data center, which is the warehouse-sized computer room behind cloud services and artificial intelligence systems. Reuters reported that SiFive is using this money to push into data-center central processor chips, where the prize is much bigger than the microcontrollers and embedded devices where RISC-V first spread. (reuters.com) SiFive said the new cash will accelerate its high-performance roadmap for data centers and expand engineering teams for what it called “agentic artificial intelligence” workloads. In plain terms, it wants processors that can handle the control work around artificial intelligence systems, not just the graphics chips that do the heavy matrix math. (sifive.com) (datacenterdynamics.com) Nvidia’s presence is the part that makes other chip companies pay attention. Nvidia already dominates artificial intelligence accelerators, so an investment in a RISC-V core supplier hints that future server designs may use more than one processor architecture inside the same rack. (reuters.com) (forbes.com) That is awkward timing for Arm because Arm has also been pushing harder into the data center. In March 2026, Arm launched its first physical data-center processor, a break from its long-running model of only licensing designs to others. (datacenterdynamics.com) So this round is not just a startup getting bigger. It is one more sign that the chip stack inside data centers is splitting into more camps at the same time that artificial intelligence demand is making every layer of that stack more valuable. (bloomberg.com) (sifive.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.