NVIDIA announces Blackwell Ultra B300 AI chip — 50% faster than B200
At GTC 2025, NVIDIA officially announced the Blackwell Ultra B300, a new GPU for data centers and artificial intelligence.
DeepSeek opens 3FS — file system code for AI training servers
During Open Source Week, DeepSeek made its Fire-Flyer Fire System (3FS) parallel file system publicly available. According to the company, 3FS reaches 7.3 TB/s aggregate read speed on its own server data clusters, where it has been used since 2019.
Apple asks Foxconn and Lenovo to build AI servers on Silicon — but in Taiwan
Apple plans to use their own processors Apple Silicon in its own artificial intelligence servers. The company asks Foxconn and Lenovo to assemble them in Taiwan.
Kioxia announces light-speed SSD data transfer — optical connection allows SSDs to be placed far from servers
Kioxia, one of the world’s leading SSD manufacturers, has announced new optical interface SSDs for next-generation green data centers.
13th and 14th generation Intel processors «glitchy» also on servers — suppliers advise AMD
It is already known that Intel processors of the 13th and 14th generations (Raptor Lake) have problems with stability particularly in games. However, youtuber Level1Techs found that the phenomenon is also common in data centers
«AI gold rush»: global IT investments will grow by 8% — to $5 trillion in 2024
Global IT investments are expected to increase by 8% in 2024 compared to 2023 and amount to $5.06 trillion.
NVIDIA unveiled the Blackwell architecture, B200 GPUs, NVLINK 7.2T interface, B200 NVL72 servers, and supercomputers based on it
At the GTC March 2024 Keynote event, NVIDIA introduced the Blackwell chip architecture and B200 GPUs based on it. The new GPUs with AI capabilities will replace the retiring H100 and H200 in data centers. Grace Blackwell GB200 chips will combine the Grace architecture with Blackwell.
Spelling error report
The following text will be sent to our editors: