The AI semiconductor design company FuriosaAI announced on the 31st that it successfully raised $125 million (approximately 173.6 billion KRW) in a Series C round of investment.
FuriosaAI stated on its website that the cumulative investment amount now totals $246 million (about 341.7 billion KRW).
This round of funding included participation from Korea Development Bank, Industrial Bank of Korea, KStone Partners, PI Partners, and Kakao Investment.
Doyoung Kim, CEO of Kakao Investment, evaluated FuriosaAI by saying, “FuriosaAI has established a robust alternative to GPUs for AI inference, and their tensor contraction processor chip architecture delivers essential performance and power efficiency for next-generation AI applications.”
FuriosaAI plans to immediately use the newly secured funds to expand the production of its main chip, Renegade, accelerate market entry, and develop next-generation chips.
The company added, “We are making bold investments to secure solutions most suited for future AI agent systems and inference models,” and intends to expand the limits of high-performance, ultra-efficient AI computing through next-generation chip design and mass production.
Recently, LG AI Research Institute announced the adoption of FuriosaAI’s Renegade chip for its large language model (LLM), Exaone.
According to LG’s experiments, the Renegade demonstrated 2.25 times better performance per watt in inference execution of Exaone, a medium-sized model, compared to existing GPUs.
Meanwhile, FuriosaAI has appointed Jihoon Kang, an associate professor at Korea Advanced Institute of Science and Technology (KAIST) in the computer science department, as the Chief Research Officer (CRO), and Youngjin Cho, former vice president at Samsung Semiconductor’s Memory Business Unit, as the Vice President of the hardware division.