Korean Startup FuriosaAI Wins First Major Client with LG for AI Chip RNGD

FuriosaAI, a Seoul-based startup, has secured its first major customer, LG AI Research, for its AI chip RNGD. The chip will power LG's Exaone large-language models, and FuriosaAI plans to deploy RNGD servers across various industries. The startup rejected an $800 million acquisition offer from Meta in March and aims to raise capital before an IPO.

Seoul-based startup FuriosaAI has secured its first major customer, LG AI Research, to power its AI chip, RNGD. The collaboration will see FuriosaAI deploy RNGD servers across various industries, including electronics, finance, telecommunications, and biotechnology. This strategic partnership follows FuriosaAI's decision to decline an $800 million acquisition offer from Meta in March.

The RNGD chip, optimized for running large language models (LLMs), will be integrated into LG AI Research’s recently unveiled EXAONE 4.0 platform. The partnership aims to address the growing need for sovereign AI and reduce the reliance on third-party suppliers like Nvidia. LG AI Research has reported that RNGD outperforms traditional GPUs, delivering 2.25 times better inference performance and improved energy efficiency [1][2].

FuriosaAI’s RNGD chip is not a general-purpose GPU but is designed specifically for AI computing, which allows for more efficient and cost-effective solutions. The Seoul-based startup, with a global team of just 15 employees, plans to expand its customer base and potentially raise capital before an IPO. The CEO of FuriosaAI, June Paik, stated that the partnership with LG AI Research will open up business opportunities beyond South Korea.

This development is significant as it marks one of the few public endorsements of a rival to Nvidia by a major enterprise. FuriosaAI’s hardware is noted for its lower cost compared to traditional solutions, making it an attractive option for enterprises looking to optimize their AI infrastructure [1][2].

FuriosaAI's RNGD chip has successfully passed rigorous performance tests with LG AI Research’s ExaOne models. The RNGD server solution, which aggregates the power of eight RNGD accelerators into a single, air-cooled 4U chassis, is now available to enterprise customers leveraging LLMs [2].

Moving forward, FuriosaAI and LG AI Research are committed to enabling businesses to deploy advanced models and agentic AI sustainably, scalably, and economically. The partnership aims to provide a powerful and sustainable AI infrastructure stack for advanced AI applications.

References:
[1] https://techcrunch.com/2025/07/21/instead-of-selling-to-meta-ai-chip-startup-furiosaai-signed-a-huge-customer/
[2] https://gamesbeat.com/lg-ai-research-taps-furiosaai-to-get-2-25x-better-llm-inference-performance-vs-gpus/