Yonhap Infomax NewsFeb 11
South Korea will invest 1 trillion won ($845 million) over five years to develop domestic AI semiconductors, aiming to reduce reliance on foreign chips and boost the local industry through the M.AX Alliance and new policy support.
#YonhapInfomax #AISemiconductor #MOTIE #FuriosaAI #KOnDevice #1TrillionWonInvestment #Economics #FinancialMarkets #Banking #Securities #Bonds #StockMarket
https://en.infomaxai.com/news/articleView.html?idxno=105004
Government to Invest 1 Trillion Won in 'K-On-Device AI Semiconductor' Development

South Korea will invest 1 trillion won ($845 million) over five years to develop domestic AI semiconductors, aiming to reduce reliance on foreign chips and boost the local industry through the M.AX Alliance and new policy support.

Yonhap Infomax
Pirates.BZ Legacy AccountJan 20
#FuriosaAI, a Seoul-based #AIchipdeveloper, is reportedly seeking $300 million to $500 million in a Series D funding round. The company’s flagship #RNGD chip is optimised for #tensorcontraction, a mathematical operation that can produce the same results as #matrixmultiplication more efficiently, leading to faster #AIperformance. https://siliconangle.com/2026/01/19/ai-chip-developer-furiosaai-reportedly-raising-500m/?Pirates.BZ #Pirates #Tech #Startup #News
GripNewsJan 15
🌗 隆重推出 Furiosa NXT RNGD 伺服器：資料中心級的高效率 AI 推論方案
➤ 無需改建資料中心，企業也能輕鬆駕馭大規模 AI 推論
✤ https://furiosa.ai/blog/introducing-rngd-server-efficient-ai-inference-at-data-center-scale
FuriosaAI 正式發表了旗下首款品牌化套裝解決方案「NXT RNGD 伺服器」，旨在打破現今企業在部署大規模 AI 推論時面臨的硬體瓶頸。這款伺服器專為大語言模型（LLM）與多模態模型設計，核心優勢在於極高的能源效率與相容性：它僅需 3 kW 功耗即可運作，能完美適配目前全球超過 80% 仍採用氣冷技術的舊型資料中心。透過預載的 Furiosa SDK 與相容於 vLLM 的執行環境，企業無需改建基礎設施或投資昂貴的液冷系統，即可從實驗階段快速轉向大規模商用部署。
+ 3 kW 的功耗設定非常精準，直接切中目前多數企業機房無法負荷高功率 GPU 的痛點。
+ 支援標準 PCIe 互
#AI 硬體 #資料中心 #LLM 推論 #FuriosaAI
Introducing Furiosa NXT RNGD Server: Efficient AI inference at data…

Furiosa RNGD Server is an enterprise-ready turnkey appliance designed to bring world-class performance to any on-prem or private cloud deployment.

FuriosaAI
WinFuture.deJan 5
#FuriosaAI startet die Massenproduktion seiner KI-Chips, die #Nvidia bei vergleichbarer Leistung mit deutlich weniger Energie Konkurrenz machen sollen. OpenAI und LG testen bereits die Technologie des Startups. https://winfuture.de/news,155961.html?utm_source=Mastodon&utm_medium=ManualStatus&utm_campaign=SocialMedia
FuriosaAI: Massenproduktion des Nvidia-Killers soll jetzt losgehen

Das Chip-Startup FuriosaAI will in den kommenden Tagen seine ersten praktischen Schritte gehen, um Nvidia anzugreifen. Das Unternehmen hat neue KI-Chips entwickelt, die sich mit denen des Marktführers messen lassen sollen.

WinFuture.de
Walk NewsDec 31
https://www.walknews.com/1164865/ FuriosaAI、「2025 Korea Tech Festival」で世界各国の通信社・機関から自社技術需要を確認 | AVING Newsのプレスリリース #FuriosaAI、「2025KoreaTechFestival」で世界各国の通信社・機関から自社技術需要を確認 #PRTIMES #SouthKorea #サービス #サイト #ニュースリリース #プレスリリース #代行 #大韓民国 #方法 #配信 #韓国
Japan Pop NewsDec 29
https://www.wacoca.com/news/2731737/ FuriosaAI、「2025 Korea Tech Festival」で世界各国の通信社・機関から自社技術需要を確認 | AVING Newsのプレスリリース #FuriosaAI、「2025KoreaTechFestival」で世界各国の通信社・機関から自社技術需要を確認 #PRTIMES #SouthKorea #サービス #サイド #ニュースリリース #プレスリリース #代行 #大韓民国 #方法 #配信 #韓国
Yonhap Infomax NewsDec 11
South Korea's AI chipmakers, including FuriosaAI and Rebellions, are gaining momentum as the government unveils a major K-Semiconductor strategy to challenge Nvidia's dominance and foster a self-reliant AI ecosystem.
#YonhapInfomax #FuriosaAI #Rebellions #AISemiconductor #Nvidia #KSemiconductorStrategy #Economics #FinancialMarkets #Banking #Securities #Bonds #StockMarket
https://en.infomaxai.com/news/articleView.html?idxno=95020
'Making Hay While the Sun Shines'—K-AI Semiconductor Firms Challenge Nvidia's Dominance

South Korea's AI chipmakers, including FuriosaAI and Rebellions, are gaining momentum as the government unveils a major K-Semiconductor strategy to challenge Nvidia's dominance and foster a self-reliant AI ecosystem.

Yonhap Infomax
Yonhap Infomax NewsDec 3
Industrial Bank of Korea (IBK) hosted a major networking event in Seoul, connecting over 500 startups and investors, and pledged to strengthen its support for venture capital and innovation, having supplied 2.5 trillion won ($1.9 billion) to more than 3,000 startups since January 2023.
#YonhapInfomax #IndustrialBankOfKorea #VentureCapital #StartupEcosystem #FuriosaAI #25TrillionWon #Economics #FinancialMarkets #Banking #Securities #Bonds #StockMarket
https://en.infomaxai.com/news/articleView.html?idxno=93738
N-gated Hacker NewsOct 25
🚀✨ Behold! The magical world where we squeeze a 5.8ms GPT-OSS-120B inference out of two RNGD cards—perfect for those who think GPUs are too mainstream. 😂🎩 #FuriosaAI and OpenAI's circus of acronyms promises a sustainable #AI future by selling you more #silicon you never knew you needed. 🙃🔮
https://furiosa.ai/blog/serving-gpt-oss-120b-at-5-8-ms-tpot-with-two-rngd-cards-compiler-optimizations-in-practice #Innovation #Sustainability #OpenAI #HackerNews #ngated
Serving gpt-oss-120b at 5.8 ms TPOT with two RNGD cards: compiler…

Here are the key optimizations that enabled two RNGD cards to achieve 5.8 ms per output token for gpt-oss-120b, running under 180 W, in just weeks.

FuriosaAI
Hacker NewsOct 25

Compiler optimizations for 5.8ms GPT-OSS-120B inference (not on GPUs)

https://furiosa.ai/blog/serving-gpt-oss-120b-at-5-8-ms-tpot-with-two-rngd-cards-compiler-optimizations-in-practice

#HackerNews #CompilerOptimizations #GPTOSS120B #Inference #AIResearch #MachineLearning #FuriosaAI

Serving gpt-oss-120b at 5.8 ms TPOT with two RNGD cards: compiler…

Here are the key optimizations that enabled two RNGD cards to achieve 5.8 ms per output token for gpt-oss-120b, running under 180 W, in just weeks.

FuriosaAI