Thinking Machines, founded by former OpenAI CTO Mira Murati, has partnered with Nvidia to develop gigawatt-scale AI data centers. This collaboration aims to meet the escalating compute demands of futu...
A confidential AI infrastructure hyperscaler secured $2 billion in funding on March 9, 2026, aiming to accelerate AI compute development amid delays in Oracle and OpenAI's Stargate project. The invest...
OpenAI and Oracle's Project Stargate is pioneering AI data centers with power demands reaching 4.5 gigawatts, equivalent to four nuclear reactors. This scale challenges existing energy grids and infra...
NVIDIA’s NVFP4 low-precision training and support for long-context models, combined with OpenAI and Oracle’s Project Stargate, are driving a new era in AI infrastructure. This shift emphasizes integra...
This week’s AI infrastructure roundup highlights NVIDIA’s NVFP4 format, coding agents that reduce runtime costs, and OpenAI’s GPT-5.4 with enhanced professional skills. Together, these advances show h...
OpenAI released GPT-5.4 on March 1, 2026, featuring a one million token context window and enhanced professional capabilities. The model targets complex, regulated workflows with improved coding, tool...
NVIDIA’s new NVFP4 low-precision format and OpenAI’s GPT-5.4 model launch together signal a shift in AI infrastructure, balancing precision and throughput to handle larger, more complex AI tasks effic...
NVIDIA’s integrated approach to AI infrastructure—combining precision format innovations like NVFP4, extreme hardware-software co-design, and enhanced programming frameworks—delivers substantial gains...
NVIDIA has announced a significant breakthrough in its hardware-software co-design efforts, resulting in a substantial inference boost for Sarvam AI’s sovereign models. The industry has responded posi...
AMD is accelerating MI450 GPU production and strengthening partnerships with hyperscalers like Meta and OpenAI, aiming to disrupt the AI infrastructure market. This dual approach targets both hardware...