FuriosaAI, a South Korean AI chipmaking company, has reportedly turned down an acquisition deal from Meta worth $800 million. The negotiations broke down due to disagreements over post-acquisition business strategy and organizational structure, with the company choosing to focus on developing and producing its AI chips.
Meta, along with other tech companies working on AI, is trying to reduce reliance on chipmaking giant Nvidia. Companies like Positron AI have been funded with the aim of providing competition. Nevertheless, Nvidia remains on a rising streak, with its Q4 earnings report revealing it to be the second most valuable company.
Read: What you need to know about the state of the metaverse today (March 19, 2022)
FuriosaAI, which was founded in 2017 by June Paik, has developed two AI chips called Warboy and Renegade (RNGD) to take on the likes of Nvidia and AMD. The startup has said it has completed testing the RNGD chips, which are said to be best suited for reasoning models, in partnership with LG AI Research and Aramco. LG AI Research reportedly plans to use RNGD chips in its AI infrastructure, and the startup plans to launch the chips later this year.
Meanwhile, the startup is reportedly in the talks to raise approximately $48 million (KRW 70 billion). It aims to complete the funding this month.
An industry insider commented, “Despite the lure of global capital, FuriosaAI’s decision to focus on technological independence and long-term growth sends a strong message to the Korean AI ecosystem.” The company reportedly has plans to accelerate the commercialization and global expansion of its Renegade chip in the near future.
READ: Metaverse: Sony embraces virtual world, others are not far behind (May 19, 2022)
Meta, among other tech giants, is focused on developing cutting-edge AI models and dominating the AI race. In a statement, Zuckerberg said he expected 2025 to be the year “when a highly intelligent and personalized AI assistant reaches more than 1 billion people,” and that he hoped Meta AI to be that AI assistant.
However, developing the latest AI technologies require highly sophisticated chips or graphic processing units (GPUs) to support their workloads. Earlier this month, Meta began testing its first in-house chip for training AI, in hopes that the new chip would help bring down significant infrastructure costs as Meta places expensive bets on AI to drive growth.