Necessary Always Active
Necessary cookies are required to enable the basic features of this site, such as providing secure log-in or adjusting your consent preferences. These cookies do not store any personally identifiable data.
|
||||||
|
||||||
|
||||||
|
FuriosaAI, the Nvidia chip challenger, has signed its initial major contract with LG after turning down Meta’s $800 million acquisition offer, Yahoo Finance reported. The Seoul-based chip design startup won the deal after its AI chip, RNGD, got approved by LG AI Research following a rigorous evaluation that lasted seven months.
LG’s evaluation included efficiency and performance checks. Speaking to Bloomberg News, FuriosaAI CEO June Paik said LG, which is also a South Korean company, will use its chips to power its large-language AI model called Exaone.
“For the last eight years, we worked very hard from R&D to product phases and finally this commercialization phase. This signals that our product is ready for enterprise adoption,” Paik said.
As an AI chip startup, FuriosaAI views LG’s approval as an important validation. The company is among the few Korean semiconductor designers that are looking to ride the post-ChatGPT boom in developing AI infrastructure. FuriosaAI designed the RNGD chip to challenge industry leader NVIDIA, along with other AI chip startups like Groq, Cerebras Systems, and SambaNova.
Paik, who previously served at Advanced Micro Devices (AMD) and Samsung Electronics Co, founded FuriosaAI back in 2017. The Korean startup designs semiconductors that support AI inferencing. It claims that compared to graphics processing units, its chips can deliver 2.25 times better inference performance.
Paik recognized LG AI’s Exaone as the leading AI model saying that it will play a key role in shaping the broader AI landscape in the country. In his view, interest in the model and its chips remains strong both locally and abroad. He added that LG has already penetrated the global market and expects the company’s customers to buy chips manufactured by his AI startup. FuriosaAI chips are designed exclusively for AI computing.
The partnership between FuriosaAI and LG AI will involve deployment of RNGD chips across different industries, from finance to electronics. The chips will also be used to power LG’s enterprise AI agent called ChatExaone as LG plans to expand it to external clients.
FuriosaAI is focused on tapping a huge base of talent, government incentives, and suppliers within the semiconductor ecosystem These talents and incentives have emerged around SK Hynix and Samsung over the last 10 years.
Currently, FuriosaAI is working on acquiring customers abroad in the Middle East, the US, and Southeast Asia. According to Paik, the startup expects to enter into similar agreements over the next six months.
FuriosaAI captured public attention in March after news that it had rejected an acquisition offer from Meta. Disagreements between the two firms emerged after their organizational plans and business strategies clashed.
Following the collapse of the deal, Paik issued a statement underscoring FuriosaAI’s goal to pursue and advance sustainable AI. Meta has been working towards reducing its dependence on NVIDIA AI chips that it needs to train and run large language models. The social media giant plans to spend over $65 billion on AI initiatives this year.
LG AI unveiled a new hybrid AI model called Exaone 4.0 this month. The electronics giant is already marketing the latest model as the first hybrid-reasoning system. According to LG, Exaone 4.0 is capable of generating and assessing hypotheses.
LG also says that it is working to improve its Exaone AI models to position itself in the market. In March this year, the South Korean company launched its first reasoning AI model known as Exaone Deep. This model features 32 billion parameters to enhance AI reasoning and learning, with experts placing its performance at DeepSeek’s R1 level.
FuriosaAI plans to raise capital before going public in the future. Currently, the chip maker is negotiating a $48 million funding round with investors and expects to secure the investment by the end of July.