企業導入 AI 邁入新階段！因應資安疑慮、雲端降速與模型變動，企業算力需求正從雲端走向「地端落地」。麗臺科技於 COMPUTEX 2026 打造出一整條從工廠邊緣到資料中心的 AI 算力光譜，要助企業完美掌控代理式 AI 的核心機密與推論時效。
由於現在各類記憶體都優先分配到AI產業且價格飆漲，市場幾乎已經確認2026年不會有稱為GeForce RTX 50 Super的小改版產品線，然而或許是受到預期下一代消費級GPU的更新周期也再次延長，NVIDIA計畫以GeForce RTX 50 Super作為GeForce RTX 60將比預期更晚推出的墊檔產品的傳聞再起，但這次的RTX 50 Super可能與過往的做法不太相同。記憶體顆粒升級3GB、性能可能維持相同RTX 50 Super is back on track. This time includes 5060 12G ( or maybe it will have a new
由於無論AMD、Intel或NVIDIA皆把開發重心放到AI與運算領域，對於消費級市場而言可預期硬體的更迭速度放緩，但在產品生命週期會持續進行軟體增強；根據荷蘭媒體Tweakers引述AMD合作夥伴說法，兩家廠商出現有些不同的說法，但最快要到2027年第二季，最晚甚至要到2028年初。RDNA 5最快2027年中、但2027年底到2028年初機率更高▲Radeon RX 9070雖是在2025年CES公布，但實際上市卻是一季之後Tweakers其中一家詢問的AMD夥伴樂觀認為AMD最快可在2027年第二季或第三季推出基於RDNA 5架構的GPU，但另一家AMD夥伴則保守預估最快會在2027年底
Montar um PC gamer em 2026 ficou mais fácil em um ponto e mais complicado em outro. Fácil porque tanto AMD quanto Intel têm processadores fortes em várias faixas de preço. Complicado porque comprar a CPU “mais poderosa” nem sempre melhora o desempenho nos jogos. O erro mais comum continua o mesmo: gastar demais no
The post Ryzen ou Intel: qual CPU escolher para PC gamer em 2026? appeared first…
C
YPlasma has showcased its ultrathin solid-state cooling module based on dielectric barrier discharge (DBD) plasma actuators on NVIDIA Jetson Orin Nano at Computex 2026. Ultrathin solid-state cooling solutions replacing thicker and noisier mechanical fans have been demonstrated on consumer hardware in the past, with solutions such as xMEMS µCooling fan-on-a-chip for SSDs and Frore Systems Airjet…
Asus has had a busy Computex 2026, debuting a new generation of AI creator ProArt PCs, a new ProArt Mini PC designed for creative professionals, and a new Qualcomm Snapdragon-powered mini PC, the Asus QN10.
[Read More]
微軟除了在NVIDIA GTC Taipei的NVIDIA RTX Spark發表環節公布高效能筆電Surface Laptop Ultra以外，在Windows Build大會還公布一款同樣搭載NVIDIA RTX Spark處理器的迷你超級電腦RTX Spark Dev Box，鎖定Windows環境開發者的全能強大利器，但又小巧可放置在桌面。RTX Spark Dev Box與Surface Laptop Ultra同樣未公布售價級具體上市時間可視為Windows環境下的DGX Spark▲可視為Windows環境下的DGX Spark，但造型更為自由▲RTX Spark應該並未被要求須
C
Sharpa Wave is a high-end dexterous robotic hand with 22 DoF, high-sensitivity dynamic tactile array
NVIDIA just announced the Isaac GR00T Reference Humanoid Robot in a half-baked press release for Computex 2026, with an “available-soon reference workflow” and availability sometime by the end of the year with the Unitree H2 humanoid chassis. One component of the kit that appears to be available now, albeit in limited quantity, is the Sharpa Wave high-end dexterous robotic hand with 22 degrees of…
Arquitectura NVIDIA Grace Blackwell 3nm. CPU: hasta 20 núcleos Grace GPU de nivel RTX 5070 desarrollada en colaboración con MediaTek, hasta 6.144 núcleos CUDA 128 GB de memoria unificada 1 Petaflop de IA local delgado, ligero, apenas se ralentiza sin cable En el directo movió Forza Horizon 6 y 007 First Light en el escenario […]
Ver post completo: NVIDIA acaba de lanzar su primer procesador de…
Microsoft and Nvidia made joint announcements today. Microsoft is launching a brand-new Surface Laptop Ultra, the most powerful Surface Laptop ever built, and it is powered by Nvidia's new RTX Spark system-on-a-chip, a "new superchip that reinvents Windows PCs for the era of personal AI agents."
[Read More]
The first Windows laptop and desktop computers with NVIDIA’s new RTX Spark “superchip” are expected to arrive this fall. Major PC makers including Asus, Dell, HP, Lenovo, Microsoft and MSI have already announced plans to release PCs based on the new platform… although we don’t have prices or exact release dates yet. It’s a safe […]
The post Windows PCs with NVIDIA RTX Spark chips coming this…
NVIDIA and Microsoft are turning local agent security into a platform problem
在Computex 2026期間的GTC Taipei，NVIDIA執行長黃仁勳不僅回顧在GTC 2026介紹的Vera Rubin系統，更花費相當時間介紹Vera CPU以及闡述在Vera CPU在Vera Rubin系統及代理式AI的重要。而；看到NVIDIA執行長黃仁勳對NVIDIA研發的CPU侃侃而談、且是在冠上GTC的NVIDIA重要活動，是認識NVIDIA幾十年、過往把重心聚焦在GPU的認知之下難以想像的光景。從GPU技術公司、軟體定義公司轉型AI基礎設施公司▲NVIDIA不再僅是GPU公司或軟體定義公司，已轉型為AI基礎設施公司NVIDIA雖然以GPU技術為核心，但隨著NVIDI
Computex 2026’da PC pazarında yeni bir dönem başlatacak Nvidia RTX Spark (şimdiye kadar hep N1/N1X olarak geçiyordu) işlemci ailesi tanıtıldı. Artık PC’ler için yalnızca ekran kartı (GPU) üreten bir firma olmaktan çıkan Nvidia, ARM temelli RTX Spark serisiyle oldukça iddialı görünüyor. Üst düzey Nvidia yöneticilerinden Mark Aevermann’ın “şimdiye kadar üretilmiş en verimli bilgisayar çipi” dediği…
NVIDIA於Computex 2026前夕的GTC Taipei主題演講除了聚焦全新Windows on Arm PC級平台RTX Spark以外，也一併宣布在與微軟大力合作下，不僅止於筆電型態，合作夥伴也將推出搭載RTX Spark執行Windows的桌上型電腦、迷你電腦，在不同型態的裝置體驗兼具裝置端代理式AI的全新AI PC，NVIDIA以Arm架構的DGX S抗衡AMD Ryzen AI Max平台的策略不言而喻。與微軟合作為NVIDIA超級晶片打通Windows相容的任督二脈▲DGX Spark除了筆電還將涵蓋到桌上型電腦、迷你PC在NVIDIA與微軟大力合作下，進一步打通Arm架
Last year NVIDIA introduced a mini PC that the company called an “AI Supercomputer” thanks to its high-performance processor that combined a 20-core Arm-based CPU with NVIDIA Blackwell graphics and up to 128GB of unified memory with 273 GB/s bandwidth. Designed for running local AI models, the NVIDIA DGX Spark sold for $3,000 and up […]
The post NVIDIA’s N1 and N1X chips will bring…
The Alinx HEA13 combines an AMD Virtex UltraScale+ XCVU13P FPGA with support for NVIDIA Jetson AGX Orin and Jetson Thor modules. The platform links the FPGA and Jetson module through a PCIe Gen3 x8 interface for applications such as robotics, industrial vision, edge AI, and compute acceleration. The FPGA subsystem is based on the AMD […]
Nvidia çok uzun zamandır N1 ve N1X isimli iki farklı “SoC” üzerinde çalışıyor ve performansları çok merakla beklenen bu donanımlar Computex 2026’da resmi olarak geliyor. Şirketin buradaki paylaşımla sinyalini verdiği yeni işlemcilerin, spesifik olarak Jensen Huang’ın 1 Haziran tarihli Computex sunumunda tanıtılması bekleniyor. Dell ve Lenovo başta olmak üzere çok sayıda üreticinin N1 serisi…
聯發科搶在2026年Computex前夕舉辦媒體與分析師預先說明會，在會中並未宣布新產品，而是聚焦在聯發科當前的布局；聯發科總經理暨營運長陳冠州在開場時提到，借助聯發科長年於運算、通訊、多媒體的技術基礎以及長年攜手台積電在晶片設計、封裝等的經驗，在從AI邁向代理式發展的當下實現自端到雲的全面運算布局。▲聯發科總經理暨營運長陳冠州▲聯發科資深副總經理暨數據中心與運算事業群總經理Vince Hu▲聯發科副總經理暨車用平台事業部總經理張豫臺在分享活動中，聯發科聚焦兩大潛力領域，分別是資料中心與車用電子；資料中心由資深副總經理暨數據中心與運算事業群總經理Vince Hu負責解說，車輛電子領域由副總經理
NVDIA在正式出貨Vera Rubin系統前，已經在2026年5月將首批Vera CPU系統交給矽谷知名客戶，也象徵NVIDIA的自研Arm架構CPU不再只是GPU的附庸，而是對於客戶具有吸引力的產品；根據知名科技媒體Phoronix的評測，NVIDIA Vera CPU不僅相對Grace CPU提升63%性能，甚至面對AMD EPYC 9575F還有10%的領先，更壓倒性贏過128核心的Intel Xeon 6980P。從Grace的Arm Neoverse轉向Vera支援超執行緒的自研Olynpus架構不同於NVIDIA Grace採用Arm Neoverse架構並專注於NVLink的整
輝達執行長黃仁勳於員工大會揭示台灣新總部「Constellation」前衛設計，包含無電梯、停車場地下化，目標2030年啟用。他亦回應內部提問，並確認將設立對外開放的實體Gear Store。
Bugün gönderilen GeForce Graphics Drivers 610.47 WHQL sürücü güncellemesiyle birlikte uzun yıllardır hayatımızda olan “Nvidia Control Panel” emekliye ayrıldı. Bu konuda sürücü notları üzerinden yaptığı açıklamada, “Control Panel, 20 yılın ardından resmi olarak emekliye ayrılıyor. Kontrol panelinin tüm işlevleri ana Nvidia uygulamasına taşındı.” diyen şirket, kontrol panelini kullanmaya devam…
Nvidia çok uzun zamandır N1 ve N1X isimli iki farklı “SoC” üzerinde çalışıyor. Performansları çok merak edilen bu SoC’leri taşıyan ilk dizüstü modellerinin artık resmi olarak Computex 2026’da görücüye çıkarılması bekleniyor. Spesifik olarak Nvidia CEO’su Jensen Huang’ın 1 Haziran tarihli Computex sunumunda tanıtılması beklenen N1 ve N1X, MediaTek’in desteğiyle hayat buluyor. Dell ve Lenovo başta…
因應AI伺服器功耗劇增，嘉實多(Castrol)將液冷定位為唯一解方。其策略從供應冷卻液，升級為建構整合感測硬體與數位平台的標準化服務生態系，目標成為AI時代的熱管理基礎設施操盤者，掌握市場話語權。
AI skills becoming career essentials? As AI rapidly reshapes industries and workplaces, Nvidia CEO Jensen Huang says graduates without AI skills could fall behind in their careers.
C
Earlier this month, I received the Khadas Mind 2 (Intel Core Ultra 7 155H) mini PC with two accessories connecting through the Mind Link (PCIe x8) connector: The Mind xPlay Display and Keyboard combo, and the Mind Graphics 2 dock, adding a range of interfaces and an NVIDIA GeForce RTX 5060 Ti GPU with 16GB of VRAM. After going through the specifications, an unboxing, and a partial teardown in the…
Supermicro美超微繼2026年3月爆出公司高層聯手承包人員走私高階伺服器到中國後，又於2026年5月22日爆出海巡署在偵辦走私案件過程發現由游姓、陳姓、王姓等三名涉嫌人成立不同空殼公司及人頭負責人將NVIDIA高階晶片伺服器出口到中港澳地區，涉嫌偉造文書與懲治走私條例移送基隆地檢署；美超微再次發出公告表示強化出口管制法規執行與落實。▲美超微的官方聲明美超微的公告強調致力於與合作夥伴共同保護美國先進技術及智慧財產權，包含公司技術再最高道德與法律標準被妥善果里及應用，其中包括出貨前與出貨後的驗證，並因應與解決高需求技術被非法轉移到受限制市場風險，強調最近事件凸顯整體產業共同解決方案的必要。
為應對 AI 龐大耗能，Elon Musk 旗下 xAI 採用天然氣發電，與其潔淨能源承諾相悖，引發外界對其「宏偉計畫」的質疑。
NVIDIA 執行長黃仁勳揭示軟硬整合策略，聚焦 OpenClaw 代理式 AI 生態，並強調 Vera Rubin 將是史上最大量產。此舉旨在透過完整解決方案，建構從硬體到軟體的競爭壁壘，鞏固市場領導地位。
Recorded at HumanX, Ryan sits down with Garima Kapoor and Anand Babu Periasamy, co-founders and co-CEOs of MinIO, to chat about eliminating the storage bottlenecks that leave GPUs underutilized, their partnership with NVIDIA on the new STX reference architecture, and why modern AI infrastructure is converging on S3-compatible object storage.
Page 1
Older →