FuriosaAI

FuriosaAI’s Tensor Contraction Processor delivers high-performance, energy-efficient AI inference for LLMs.

Build AI applications at any layer with Furiosa AI. Drop-in vLLM replacement, pre-compiled model zoo, PyTorch backend, and kernel programming tools.

Deploy enterprise AI at scale with the NXT RNGD Server, combining FuriosaAI’s RNGD accelerators with optimized hardware for efficient inference.

RNGD is FuriosaAI’s flagship AI accelerator using Tensor Contraction Processor architecture.

The page explains AI agent concepts like autonomy, reasoning, and orchestration in enterprise AI systems.

FuriosaAI researches inference optimization, compiler innovation, and accelerator architecture for AI systems.

The FuriosaAI blog covers the company’s AI accelerators and NPUs shaping the AI industry.

The company examines challenges and benefits of integrating HBM3 into its RNGD accelerator for AI inference.

FuriosaAI shares updates on product launches, industry insights, and AI partnerships.

Join FuriosaAI as Algorithm - AI Research Engineer in Seoul, South Korea. Help build high-performance, energy-efficient AI compute for the Inference Era. Apply now.

Join FuriosaAI as Algorithm - AI Research Engineer Intern in Seoul, South Korea. Help build high-performance, energy-efficient AI compute for the Inference Era. Apply now.

Join FuriosaAI as Algorithm - AI System Researcher in Seoul, South Korea. Help build high-performance, energy-efficient AI compute for the Inference Era. Apply now.

FuriosaAI offers internships and residencies for engineers in sustainable AI computing.

FuriosaAI's mission is to build AI chips that efficiently run the world's advanced AI models, making AI computing more sustainable for the next generation.

The company provides information on its AI accelerators, products, and partnerships.

Furiosa offers hands-on access to its RNGD AI accelerators via hardware or API.

FuriosaAI designs high-performance, power-efficient AI accelerators (NPUs) used in data centers for computer vision, GenAI, LLMs, and demanding workloads.

FuriosaAI lawfully processes and safely manages data in compliance with the Personal Information Protection Act and other applicable laws and regulations.

GitHubRepository owner

FuriosaAI secures **$620 million in Series D funding**, achieving a post-money valuation near **$2.9 billion**. The Seoul-based company manufactures the **RNGD inference accelerator**, with **20,000 units** slated for mass production in 2026 and over **40,000 in 2027**. Korean firms like **LG Group, Upstage, and Samsung SDS** already deploy the hardware in live services. The funding supports scaling production of its **Tensor Contraction Processor (TCP)-based architecture** and software stack.

Oct 2026Jon Peddie ResearchNews article

FuriosaAI secures $620 million in Series D funding to develop AI chips targeting NVIDIA’s dominance in data center infrastructure. The round elevates the startup’s valuation to approximately $2.9 billion, positioning it as a competitor in high-performance computing. The company focuses on addressing power efficiency challenges in AI workloads, aiming to optimize data center operations.

Oct 2026VentureburnNews article

FuriosaAI Inc. secures **$620 million** in funding to scale production of its energy-efficient AI inference chips, targeting data center power constraints. The startup’s **Tensor Contraction Processor (TCP)**, based on its **RNGD architecture**, uses a **three-dimensional matrix** for tensor computations, reducing power consumption by **less than half** compared to Nvidia’s GPUs while maintaining compatibility with frameworks like PyTorch. Major South Korean firms—**LG Group, Upstage, and Samsung**—already deploy its chips for AI models, live search summaries, and neural processing services. The funding will accelerate production, aiming for **20,000 units in 2026** and **40,000 in 2027**, while advancing a **third-gen processor** co-developed with Broadcom.

Oct 2026SiliconANGLENews article