Korean chip start-up FuriosaAI has rejected Meta’s $800 million takeover offer. Instead, the artificial intelligence platform chose to grow as an independent company, according to a report published by Bloomberg. Meta Platforms Inc was in talks with the Seoul-based start-up about its acquisition since the beginning of the year.
FuriosaAI, one of the only Asian start-ups that have attracted Meta, is led by June Paik who worked with Samsung Electronics Co and Advanced Micro Devices Inc.
The eight-year old start-up develops semiconductors for AI inferencing, or services. Its second-generation processor, RNGD, has been designed to challenge productions from Nvidia, Groq Inc., SambaNova Systems Inc., and Cerebras Systems Inc. To note, RNGD is built on a 5-nanometer process of Taiwan Semiconductor Manufacturing Co. The processor uses HBM3 memory chips supplied by SK Hynix Inc.
The report also indicated at FuriosaAI’s capital raising plans before pursuing its IPO (initial public offering) plans. The start-up is expected to close an extended Series C funding round in a month. Currently, it has about 150 employees, including 15 in its Silicon Valley office.
On Monday, shares of South Korea-based venture capital firm DSC Investment Inc. --- a major supporter of FuriosaAI --- tumbled over 16%. Previously, the stock had surged following reports of potential takeover by Meta in February.
Meta is ramping up its investment in artificial intelligence infrastructure as it looks to keep pace in the rapidly evolving AI race against giants like OpenAI and Google, as well as emerging players like China’s DeepSeek. In mid-January, CEO Mark Zuckerberg announced plans to spend up to $65 billion this year, with significant allocations toward building a large-scale data center and expanding Meta’s AI talent pool.
Just a week later, he told investors that the company expects to ultimately invest hundreds of billions of dollars in AI infrastructure over the long term, Bloomberg reported. Meanwhile, the California-based company ‘Menlo Park’ is also working on its own chips designed for its AI workloads. The technology will power the recommendations and ranking ads on social media platforms like Facebook and Instagram.
