… This represents one of the largest AI factory initiatives announced to date. Additionally, the Blue Swan platform from BavariaAI will include about 1,000 GPUs to help develop multimodal AI models for use in the medical field, robotics, and various areas of scientific research. …
… Powered by the NVIDIA RTX Spark, this new desktop workstation delivers up to one petaflop of AI compute and comes with 128GB of unified memory. According to Microsoft, it can run models with up to 120 billion parameters locally, significantly reducing the need to rely on cloud GPU instances. …
… Today, OpenAI and Broadcom announced Jalapeño, OpenAI’s first custom “Intelligence Processor” designed specifically for large language model inference. The new chip is the first product from a multi-generation compute platform being developed by OpenAI. …
… Players should see better lighting and stability in their ray-traced games using this model without using the detail from the original image. DLSS 4.5 Ray Reconstruction will be available sometime in August 2026. …
… In addition to AI infrastructure, Apple also worked with Google to use technologies behind the Gemini family of models to build the next generation of Apple Foundation Models to power Apple Intelligence features across on-device and cloud workloads. …
… Other announcements at Build 2026 include the following: Microsoft unveils new security tools for IT admins and developers building AI products Microsoft announces Scout, an OpenClaw-powered personal agent for enterprise customers Microsoft unveils MAI-Thinking-1 reasoning and MAI-Code-1 coding mod… …
… New features in Search: The search giant announced that Preferred Sources are now available in AI Overviews and AI Mode. AI responses will display carousels for developing topics, along with a small text summary, and more web article links will get the "Highly Cited" label. …
… An AI model is as good as the data it's trained on. Google is reportedly reaching out to Android developers to get their private codebases to train models and improve development tools, while giving them some cash in return. …
… The standard N1 lineup appears to be aimed at more mainstream devices. One variant is tipped to feature a 12-core CPU paired with a 20 SM GPU containing 2560 CUDA cores, while another could ship with a 10-core CPU and a 16 SM graphics configuration for 2048 CUDA cores. …
… Microsoft announced a few Edge-related updates this week. The company is killing Copilot Mode in favor of native AI features on desktop and mobile devices. Also, Microsoft acknowledged the problem with passwords stored in memory as plain text and promised to fix it with an emergency patch. …
To show you the most relevant results, we’ve omitted some entries very similar to those already shown. Repeat the search with the omitted results included.