Can Anyone Beat Nvidia? This South Korean Firm Is Going to Try

Can Anyone Beat Nvidia? This South Korean Firm Is Going to Try

Despite China’s best efforts, no one’s going to steal the AI training crown from Nvidia’s hardware any time soon, but inference is another matter entirely. Whole industries are gearing up to compete with Team Green on that front. South Korea-based FuriosaAI’s RNGD platform, for example, is entering mass production and could offer stiff competition when it comes to energy savings and performance, The Wall Street Journal reports.

The RNGD chip is a neural processor designed to power the kind of matrix computations and parallel operations that running AI inference workloads require. That gives companies like FuriosaAI, with bespoke neural processing hardware, a real chance to capture some of Nvidia’s dominance in the inference space. Just as Chinese companies are looking to fill the gap left by trade-blocked Nvidia chips, FuriosaAI is looking to provide an alternative for Western markets eager for options.

Founded by former AMD GPU and Samsung Electronics engineer June Paik, and several of his previous Samsung colleagues, FuriosaAI has been developing machine learning hardware since 2017, but it’s finally set to launch its first inferencing hardware. The RNGD, short for “Renegade,” is designed to shake up an industry that Paik feels has become too reliant on Nvidia. “A market dominated by a single player—that’s not a healthy ecosystem, is it?” he tells the WSJ.

FuriosaAI RNGD rack server.

FuriosaAI pitches its RNGD system as not just faster and more efficient, but readily available, too. (Credit: FuriosaAI)

The company has now secured enough funding to value it at over $700 million, giving it a strong foundation to launch its underdog attempt to steal market share from under Nvidia. Meta tried to acquire the company in 2025 for a cool $800 million, but Paik refused. LG has reportedly tested out the AI hardware and claims it has shown excellent performance, with other companies now said to be in talks with FuriosaAI about leveraging its new hardware when available.

In its collaboration with LG in 2025, FuriosaAI showed that the company’s EXAONE AI model could run on RNGD hardware with up to 2.25 times the inference performance of competing GPUs. “We can support a wide variety of AI models efficiently. But unlike GPUs, which are still fundamentally general-purpose processors, our architecture is natively built for AI computing. We do not develop our chip for rendering or mining,” Paik said at the time.

Recommended by Our Editors

With mass production of the RNGD chip beginning this January, it won’t be long before FuriosaAI will be able to make good on some of its promises. How Nvidia and its other competitors respond will be interesting to see.



Newsletter Icon

Get Our Best Stories!

Your Daily Dose of Our Top Tech News


What's New Now Newsletter Image

Sign up for our What’s New Now newsletter to receive the latest news, best new products, and expert advice from the editors of PCMag.

By clicking Sign Me Up, you confirm you are 16+ and agree to our Terms of Use and Privacy
Policy.

Thanks for signing up!

Your subscription has been confirmed. Keep an eye on your inbox!

About Our Expert

Autor

  • Gaby Souza é criador do MdroidTech, especialista em tecnologia, aplicativos, jogos e tendências do mundo digital. Com anos de experiência testando dispositivos e softwares, compartilha análises, tutoriais e notícias para ajudar usuários a aproveitarem ao máximo seus aparelhos. Apaixonado por inovação, mantém o compromisso de entregar conteúdo original, confiável e fácil de entender