SiFive Hits $3.65B Valuation After NVIDIA Backs Its $400M RISC-V AI Chip Push
In This Article
01 The $400M Round and Who’s Behind It
02 What Makes SiFive Different
03 NVIDIA’s Strategic Bet on RISC-V
$3.65B Valuation
NVIDIA Strategic Backer
RISC-V Open Architecture
SiFive, the UC Berkeley spinout that has spent a decade proving that open-architecture RISC-V chips can compete with Arm in commercial deployments, has closed a $400M funding round led by Atreides Management with NVIDIA as a strategic investor. The round values SiFive at $3.65 billion — the highest valuation ever achieved by an independent RISC-V company — and signals that AI data center demand for open, customizable CPU architecture has reached an inflection point.
The $400M Round and Who’s Behind It
The round was led by Atreides Management, a tech-focused hedge fund that has backed several semiconductor and infrastructure companies. NVIDIA’s participation as a strategic investor — rather than a financial one — is the detail that has drawn the most attention. Strategic investment from NVIDIA typically signals product integration roadmap alignment, not just a financial bet, and the company’s backing validates SiFive’s position in the AI compute supply chain in a way pure VC dollars cannot.
According to TechCrunch’s reporting, the round was oversubscribed — meaning investor demand exceeded the amount SiFive was seeking. At $3.65B, SiFive is now the highest-valued independent RISC-V company in the world, surpassing previous milestones in the open-chip ecosystem.
What Makes SiFive Different
SiFive was founded in 2015 by the original creators of the RISC-V architecture at UC Berkeley. Rather than manufacturing chips itself, SiFive operates as an IP licensing company — similar to Arm’s business model — but built on the open RISC-V instruction set architecture, which means customers don’t pay the per-unit royalties that Arm charges. For companies building AI accelerators and data center CPUs at scale, the royalty difference across millions of units is commercially significant.
SiFive’s approach allows customers to take its optimized processor cores and customize them for specific workloads without licensing restrictions. This flexibility has made SiFive cores increasingly popular for AI inference accelerators, where workload-specific customization can deliver meaningful efficiency gains over general-purpose CPU designs.
NVIDIA’s Strategic Bet on RISC-V
NVIDIA’s investment in SiFive isn’t isolated — it fits directly into the company’s NVLink Fusion initiative, which enables third-party CPU architects to tightly integrate with NVIDIA’s GPU ecosystem. SiFive’s RISC-V cores are confirmed as compatible with the NVIDIA CUDA software stack, meaning customers can deploy SiFive CPUs alongside NVIDIA GPUs with full software interoperability.
For NVIDIA, backing SiFive hedges against a scenario where Arm’s licensing terms become commercially unfavorable for data center customers and the industry pivots to RISC-V as the CPU standard. By being a strategic investor in the leading RISC-V IP company, NVIDIA ensures its GPU software stack remains compatible regardless of which CPU architecture wins.
Why RISC-V Is Winning in AI Data Centers
The AI compute buildout has created a unique environment for RISC-V adoption. Hyperscalers are designing custom silicon at unprecedented scale, and the ability to modify the CPU core itself — not just the surrounding chip design — is a meaningful advantage for optimizing specific AI workloads. RISC-V’s open architecture enables this level of customization; Arm’s licensing model restricts it.
China’s accelerated domestic AI chip development has also been a significant driver: US export controls have pushed Chinese AI companies to adopt RISC-V for components they cannot source from Arm licensees subject to US restrictions. This geopolitically-driven demand has materially expanded the RISC-V ecosystem and accelerated the maturation of its software tools.
Related Coverage
→ OpenAI IPO Q4 2026: The ChatGPT Enterprise Pivot Explained
Frequently Asked Questions
What is SiFive and what does it do?
SiFive is a semiconductor IP company founded in 2015 by UC Berkeley researchers, including the original creators of the RISC-V architecture. It licenses optimized RISC-V processor cores to chip designers, similar to Arm’s model but without per-unit royalties.
Why did NVIDIA invest in SiFive?
NVIDIA invested strategically to ensure SiFive’s RISC-V cores remain compatible with its CUDA software stack and NVLink Fusion program. This positions NVIDIA favorably regardless of whether Arm or RISC-V becomes the dominant CPU architecture for AI data centers.
What is RISC-V and how does it differ from Arm?
RISC-V is an open-source instruction set architecture (ISA) that anyone can implement freely. Arm is a proprietary ISA that requires per-unit licensing royalties. RISC-V’s open nature allows chip designers to customize cores at the architecture level without licensing restrictions.
What is SiFive’s valuation after this round?
SiFive is valued at $3.65 billion following the $400M oversubscribed round, making it the highest-valued independent RISC-V company in history.
How does SiFive fit into AI data center infrastructure?
SiFive’s RISC-V cores are used for CPU components in AI accelerators and data center processors. Their customizability, CUDA compatibility, and royalty-free structure make them an increasingly attractive alternative to Arm-based designs for hyperscale AI infrastructure builds.
Track Every Major Startup Round
Networkcraft covers the funding rounds, valuations, and strategic moves shaping deep tech. Get the weekly brief.