Semiconductor News & Analysis Feed

20 articles
2026-08-18
news.google.com 2026-08-18
2026-08-18
news.google.com 2026-08-18
2026-08-18
247wallst.com 2026-08-18
As the artificial-intelligence infrastructure expands rapidly, the semiconductor industry is facing an unprecedented supply-demand imbalance. High-bandwidth memory (HBM), a core component in AI accelerators, is under significant strain, affecting the entire sector. While companies like Micron Techno
2026-08-11
biz.chosun.com 2026-08-11
UBS analysts suggest that although NVIDIA may reduce the amount of high bandwidth memory (HBM) in its next-generation AI accelerator Rubin Ultra, overall HBM demand in 2027 could still exceed forecasts. By lowering HBM per chip, NVIDIA can ease supply constraints and increase shipments. UBS notes th
2026-08-11
www.tomshardware.com 2026-08-11
Nvidia is reportedly testing lower-memory configurations of its upcoming Rubin Ultra accelerator due to ongoing shortages of high-bandwidth memory (HBM). According to The Information, some versions are being tested with as little as 192 GB of memory and are switching from the originally announced HB
2026-08-11
tomshardware.com 2026-08-11
NVIDIA is reportedly testing lower-memory configurations of its upcoming Rubin Ultra accelerator due to a global shortage of high-bandwidth memory (HBM). According to The Information, the company is evaluating versions with 192GB or 256GB of memory, using HBM4 instead of the originally announced HBM
2026-08-08
news.google.com 2026-08-08
2026-08-07
news.google.com 2026-08-07
2026-08-07
news.google.com 2026-08-07
2026-08-07
news.google.com 2026-08-07
2026-08-07
digitimes.com 2026-08-07
2026-08-07
news.google.com 2026-08-07
2026-08-07
news.google.com 2026-08-07
2026-08-07
news.google.com 2026-08-07
2026-08-07
news.google.com 2026-08-07
2026-08-07
news.google.com 2026-08-07
2026-08-05
www.techtimes.com 2026-08-05
NVIDIA's upcoming Rubin Ultra AI chip, originally positioned as its 2027 performance flagship, may ship with less on-package high-bandwidth memory (HBM) than its predecessor due to ongoing DRAM shortages. According to TrendForce, NVIDIA is evaluating four HBM configurations, including one that reduc
2026-08-04
finance.biggo.com 2026-08-04
2026-07-09
wccftech.com 2026-07-09
NVIDIA's latest Rubin AI servers, especially the Ultra variant, are setting new benchmarks for cost in the high-end AI computing space. According to BofA Global Research estimates, the Rubin Ultra rack is priced at an estimated $21 million, with HBM4e memory alone accounting for around $1.5 million.
2026-07-06
tomshardware.com 2026-07-06
NVIDIA's Kyber NVL144 rack delivery has been pushed to 2028, a delay of over 12 months that impacts the planned rollout of Rubin Ultra GPUs in 2027. The setback stems from manufacturing challenges with a critical PCB midplane component, which connects eight Oberon racks and replaces traditional cabl