Anthropic is in early talks to acquire Israeli startup Decart AI for around $6 billion, its largest acquisition ever, according to reports. The deal is primarily aimed at reducing Anthropic's $1.25 bi…
Amazon has paid the full $50 billion investment to OpenAI, despite none of the milestones being met, securing a 5% stake in the company, according to Financial Times. Amazon also committed up to $33 b…
ZML, a Paris startup, released LLMD on July 8, a free, open-source inference server that runs LLaMA, Gemma, Qwen, and Mistral on NVIDIA CUDA, AMD ROCm, Google TPU, Intel oneAPI, and Apple Metal from a…
Ray AI libraries (Serve, Data, Train) now support Google TPU slices through a topology field that reserves a whole ICI-connected slice, preventing multi-host deployment hangs. Ray Serve serves LLMs vi…
The FBI has issued a Request for Information for AI supercomputer hardware featuring Nvidia B300 GPUs and Google TPUs, to be deployed at its CJIS Division in Clarksburg, West Virginia, for LLM trainin…
ZML launched LLMD, a free inference server that runs LLaMA, Gemma, Qwen, and Mistral models on NVIDIA, AMD, Google TPU, Intel, and Apple hardware from a single Docker image. Built in Zig and compiled …
Google released Gemma 2, an open model family with 9B and 27B parameter sizes, featuring architectural changes like hybrid attention and Grouped-Query Attention for improved inference efficiency. The …
Researchers at UC San Diego, led by Hao Zhang, have successfully implemented a diffusion-style speculative decoding method called DFlash on Google TPUs, achieving an average 3.13x increase in tokens p…