FuriosaAI Closes $125M Funding to Scale AI Inference Chip Production

FuriosaAI secured $125 million in Series C bridge funding to accelerate production of its RNGD (“Renegade”) AI inference chip and advance development of its next-generation silicon. The Seoul-based company, founded in 2017, is now valued at $735 million and has raised $246 million to date. Investors in this round included Korea Development Bank, Industrial Bank of Korea, Keistone Partners, and Kakao Investment.

The company recently won a major design contract with LG AI Research, which selected RNGD after benchmarking showed 2.25x higher LLM inference performance per watt compared with GPUs. FuriosaAI has also expanded its executive team, bringing in Jeehoon Kang, a KAIST professor specializing in parallel systems, as Chief Research Officer, and Youngjin Cho, a former Samsung Corporate VP and SoC expert, as VP of Hardware.

FuriosaAI argues that GPU-based architectures are increasingly unsustainable due to high power consumption and infrastructure costs. Its RNGD chip is designed to deliver better performance-per-watt for AI inference in enterprise and data center environments. The new capital will fund global scaling of RNGD production and support the company’s roadmap for future chip designs.

• $125M Series C bridge funding; total raised: $246M
• Company valuation: $735M
• Major RNGD design win: LG AI Research
• RNGD chip delivers 2.25x better LLM inference performance-per-watt vs. GPUs
• New leadership hires: Jeehoon Kang (Chief Research Officer) and Youngjin Cho (VP of Hardware)
• Investors: Korea Development Bank, Industrial Bank of Korea, Keistone Partners, Kakao Investment

“AI today is dependent on a broken business model, where the infrastructure costs and constraints of power-hungry GPUs are a critical, rapidly worsening roadblock. We’re committed to solving this so that AI is truly sustainable – not just environmentally but economically as well,” said June Paik, co-founder and CEO of FuriosaAI.

🌐 Why it Matters: FuriosaAI is positioning RNGD as a purpose-built inference accelerator that addresses one of AI’s most pressing bottlenecks—scaling LLMs without runaway energy and infrastructure costs. The company’s win with LG AI Research validates its architecture, which integrates a redesigned hardware and software stack optimized for parallel processing and compiler-driven efficiency. Unlike GPUs, which evolved from graphics workloads, RNGD is engineered from first principles for AI inference, achieving higher throughput and lower power per watt. With enterprise adoption and fresh funding, FuriosaAI is emerging as a credible alternative to GPU-dominated compute, reflecting a broader shift toward specialized silicon in AI infrastructure.

🌐 We’re tracking the latest developments in networking silicon. Follow our ongoing coverage at: https://convergedigest.com/category/semiconductors/