|
|
|
|
|
01.10.26 - 04:42
|
Nvidia ties AI agent security to BlueField-4, Vera CPU (Digitimes)
|
|
|
Over the past year, Nvidia has repeatedly observed AI agents drifting beyond approved workflows. In some cases, agents quietly escalated read permissions to write access, sent requests to unauthorized endpoints, or clicked unauthorized CAPTCHA-bypass links....
|
|
|
|
|
|
|
|
|
01.10.26 - 04:06
|
Memory shortages deepen as AI demand keeps server prices climbing (Digitimes)
|
|
|
The memory market is entering the fourth quarter of 2026 with a widening split between weak consumer demand and intense server buying, underscoring how artificial intelligence workloads are reshaping supply. Spot prices for DDR5 server modules continue to rise, while some lower-grade products are softening in the channel....
|
|
|
01.10.26 - 04:06
|
Applied Materials and Kioxia team up on AI-era memory tech (Digitimes)
|
|
|
Applied Materials and Kioxia are joining forces at a Silicon Valley research hub to develop next-generation memory technologies that could shape AI systems used worldwide. The collaboration underscores efforts to boost memory density, speed, and energy efficiency, which are increasingly important as global demand for data-intensive computing and advanced chip packaging rises....
|
|
|
01.10.26 - 04:06
|
AI chip packaging demand meets labor pushback: ASE′s NT$5.6B Taiwan snack factory deal sparks strike vote (Digitimes)
|
|
|
ASE Technology Holding (ASEH) announced on July 30 that it would spend roughly NT$5.67 billion (approx. US$177.97 million) to acquire land and factory buildings in the Zhongli Industrial Park from Taiwanese snack maker Kuai Kuai, a move meant to support future capacity expansion at its Zhongli plant. The deal, however, has unexpectedly become a flashpoint, with the Kuai Kuai union recently holding a strike vote and protesting outside ASE's Zhongli facility....
|
|
|
01.10.26 - 04:06
|
DIGITIMES Insight: AI scaling is moving from chip shrinkage to system integration (Digitimes)
|
|
|
TSMC said the next phase of AI performance growth will depend less on shrinking a single chip and more on combining multiple chips, memory, and advanced packaging inside larger systems. The shift could reshape semiconductor competition, as demand for compute, bandwidth, power delivery, and cooling rises in tandem. The company outlined the trend at its heterogeneous integration forum during SEMICON Taiwan 2026, saying AI computing transistors in a single CoWoS package are set to rise more than 48-fold by 2029 from 2024 levels, while total HBM bandwidth is expected to increase more than 34-fold....
|
|
|
01.10.26 - 04:06
|
Google to test AI computing in space with first orbital TPU launch (Digitimes)
|
|
|
Google is preparing to send its in-house AI accelerator, the Tensor Processing Unit (TPU), into low Earth orbit for the first time as it explores whether AI computing can eventually expand beyond terrestrial data centers and into space. CEO Sundar Pichai recently announced that Project Suncatcher will begin its first in-orbit test on October 1, with computing power roughly equivalent to that of a single ground-based server....
|
|
|
01.10.26 - 04:06
|
Applied Materials′ EPIC Center adds Kioxia for AI memory, 3D stacking push (Digitimes)
|
|
|
Applied Materials announced that Kioxia will join its Equipment and Process Innovation and Commercialization (EPIC) Center in Silicon Valley, California. The two companies will collaborate on developing advanced memory cells and structures, multi-chip stacking architectures, and new materials engineering technologies to improve the memory density and performance required in the artificial intelligence (AI) era....
|
|
|
01.10.26 - 04:06
|
HPE secures US$1.2 B order from Vultr for first AMD Helios AI rack systems (Digitimes)
|
|
|
Hewlett Packard Enterprise announced a major US$1.2 billion order from Vultr, which HPE describes as the world's largest privately held cloud infrastructure company, on September 30. The agreement marks HPE's inaugural order for the AMD Helios™ AI Rack by HPE, which will be deployed across Vultr's US cloud data centers to power large-scale model training and inference workloads for global enterprises and service providers....
|
|
|
01.10.26 - 04:06
|
Nextronics sees AI, robotics, medical demand reshaping revenue mix (Digitimes)
|
|
|
Connector maker Nextronics Engineering said its humanoid robot-related products began shipping in 2026, while North American customers are preparing for mass production and have set a 2027 production target. The company expects those programs, alongside AI-related communications, to lift revenue sharply in 2027 and further shift its business mix toward higher-growth segments....
|
|
|
|
|
|
|
01.10.26 - 03:06
|
Jabil targets 24% growth in FY27 as AI hardware ramp reshapes business (Digitimes)
|
|
|
Jabil's fourth quarter of fiscal 2026 marked the point at which AI data-center hardware became the clear center of gravity for the US contract manufacturer. Revenue for the three months ended August 31 rose 29% from a year earlier to US$10.6 billion. The GAAP operating margin widened to 5.7% from 4.1% as new capacity came online faster than planned and more hyperscale customers moved into volume production....
|
|
|
01.10.26 - 02:48
|
Synopsys and OpenAI to share revenue from AI model built to run chip design tools (Digitimes)
|
|
|
Synopsys is betting that the surest way to keep generative AI from eroding its electronic design automation (EDA) franchise is to build the model together with the AI developer and share the proceeds. The chip design software maker said on September 30, 2026, that it has signed a multi-year agreement with OpenAI to jointly develop GPT-Synopsys, a specialized model optimized to operate Synopsys' EDA tools, under an arrangement that Synopsys CEO Sassine Ghazi told Reuters was structured so that "it will not be cannibalizing our business."...
|
|
|
|
|
|
|
|
Börsentreff auf Facebook |
Börsentreff auf X |
Börsentreff auf Instagram
