Anthropic discloses breaches, security-related training pause in IPO filing

[image1] Oracle announced Tuesday that it is expanding its artificial intelligence infrastructure partnership with Nvidia while adding chips from competing suppliers, including Google, Cerebras Systems and SambaNova Systems. The companies plan to expand their collaboration to a multi-gigawatt scale using Nvidia's next-generation Vera Rubin AI servers, Reuters reported. Financial terms of the agreement were not disclosed. Oracle also plans to offer Google's tensor processing units for the first time and use its own Acceleron networking chips in some Nvidia-based systems, reducing its reliance on Nvidia's networking technology. The expansion will give Oracle customers more options for running AI workloads, including inference, the process in which AI systems generate responses to user requests. The move comes as chipmakers compete for a larger share of the growing AI infrastructure market.

First reported by Fox News. This page is a short excerpt; the full text is theirs.