Meta in Talks to Acquire South Korean AI Chip Startup FuriosaAI

Meanwhile, even OpenAI is building its custom chip.
Illustration by Mark Zuckerberg, founder and chief executive officer of Facebook Inc., speaks during the Facebook F8 Developers Conference in San Francisco, California, U.S., on Tuesday, April 12, 2016. Zuckerberg outlined a 10-year plan to alter the way people interact with each other and the brands that keep advertising dollars rolling at the world's largest social network. Photographer: Michael Short/Bloomberg via Getty Images

Meta is in talks to acquire FuriosaAI, a South Korean AI chip startup, to strengthen its custom chip efforts amid an NVIDIA GPU shortage. According to reports, the deal could be finalised this month, but other buyers are also in discussions.

Founded in 2017, FuriosaAI developed RNGD, an AI chip optimised for Llama 2 and Llama 3. The chip is said to consume less power than NVIDIA’s H100 GPUs. 

FuriosaAI has raised 170 billion won ($115 million) from investors, including Naver and DSC Investment. CEO June Paik, formerly of Samsung and AMD, holds an 18.4% stake.

The talks come as South Korean rival Rebellions merged with SK Hynix-backed Sapeon, forming the country’s first AI chip unicorn. 

South Korea is ramping up AI investments, positioning startups like FuriosaAI as key players. If Meta proceeds with the acquisition, it could reduce the reliance on NVIDIA’s high-cost GPUs.

Meanwhile, even OpenAI is building its custom chip.

The company plans to finalise the design soon and send it to TSMC for production, with mass production expected in 2026. A team of 40 engineers, led by former Google employee Richard Ho, is developing the chip in partnership with Broadcom. The chip will be used for training AI models and improved over time.

Tech giants Microsoft and Meta have faced challenges in producing AI chips. OpenAI’s move aligns with industry efforts to reduce dependence on NVIDIA, which controls 80% of the AI chip market. Microsoft and Meta plan to invest $80 billion and $60 billion in AI infrastructure next year, respectively.

The chip will use 3-nm technology with features similar to NVIDIA’s, including fast memory and networking. 

📣 Want to advertise in AIM? Book here

Picture of Aditi Suresh

Aditi Suresh

I hold a degree in political science, and am interested in how AI and online culture intersect. I can be reached at aditi.suresh@analyticsindiamag.com
Related Posts
Association of Data Scientists
GenAI Corporate Training Programs
Our Upcoming Conference
India's Biggest Conference on AI Startups
April 25, 2025 | 📍 Hotel Radisson Blu, Bengaluru
Download the easiest way to
stay informed

Subscribe to The Belamy: Our Weekly Newsletter

Biggest AI stories, delivered to your inbox every week.