Malaysia
Singapore
Philippines
Stay connected with BizTech Community—follow us on Instagram and Facebook for the latest news and reviews delivered straight to you.
South Korean AI chip manufacturer FuriosaAI has reportedly declined an $800 million acquisition offer from Meta, opting instead to continue its independent expansion. The rejection comes amid increasing competition in the AI chip industry, as major tech firms seek to reduce reliance on Nvidia for large language model (LLM) training and inference solutions.
Strategic Differences Led to Breakdown in Negotiations
According to reports, the decision to walk away from the Meta deal was not due to pricing concerns but rather disagreements over post-acquisition business strategy and organizational structure. Meta, like many other tech giants, is aggressively working to develop proprietary AI chips to support its growing AI infrastructure. Acquiring FuriosaAI would have provided a significant boost to these efforts, offering Meta a more efficient alternative to traditional GPUs.
FuriosaAI’s Growing Presence in AI Hardware
Founded in 2017 by June Paik, a former Samsung Electronics and AMD executive, FuriosaAI has emerged as a major player in AI chip development. The company has produced two AI chips—Warboy and Renegade (RNGD)—designed to rival Nvidia’s dominance in AI processing. Recent testing of the RNGD chip, conducted in partnership with LG AI Research and Aramco, suggests it could offer superior efficiency for AI inference models.
LG AI Research is reportedly planning to integrate RNGD chips into its AI infrastructure, with FuriosaAI aiming to bring the chips to market by the end of the year. The startup claims its RNGD platform achieves a throughput of 3,200–3,300 tokens per second when running Meta’s LLaMA 3.1-8B model, consuming just 181 watts per card—significantly lower than high-end Nvidia GPUs, which can use up to 1,200 watts each.
Independent Growth and Fundraising Efforts
Rather than selling to Meta, FuriosaAI is focusing on expanding its own business. The company announced plans to invest up to $65 billion in AI-powered initiatives by 2025 and is currently in discussions with investors to raise approximately $48 million (KRW 70 billion). The funding round is expected to close this month.
The Future of AI Chip Competition
The AI chip market is becoming increasingly competitive, with companies like Groq and Cerebras also developing alternatives to Nvidia’s GPU-centric approach. As AI adoption grows, the demand for efficient inference solutions is expected to rise, making FuriosaAI’s technology highly valuable.
For Meta, missing out on FuriosaAI underscores the challenges of securing a strong foothold in AI hardware. While the company continues developing its own AI chips, acquiring a proven startup would have accelerated its progress. Instead, FuriosaAI remains independent, positioning itself as a key player in the evolving AI infrastructure landscape.