AI news, product launches and company developments.
OpenAI & AnthropicCosmos 3 is a new physical AI model developed by NVIDIA that enables AI systems to think before acting in real-world environments. It uses an open-world foundation model to better understand and predict physical interactions. This advancement helps improve AI decision-making in robotics and autonomous systems.
NVIDIA Cosmos 3 is introduced as the first open omni-model designed for physical AI reasoning and action. It enables AI systems to better understand and interact with the physical world. This advancement opens new possibilities for robotics and AI applications requiring physical environment awareness.
NVIDIA is enhancing local AI agents on RTX PCs and DGX systems to better support personalized workflows. These agents run locally, improving privacy and responsiveness while automating tasks and generating content. This development leverages popular open source projects like OpenClaw and Hermes to expand AI capabilities on personal devices.
AI chip startup Groq is reportedly raising $650 million in funding. The company is shifting its focus from hardware to AI inference technology. This move follows Nvidia's $20 billion not-acqui-hire of Groq's talent.
AI factories convert power into intelligence in real time using token-based processing. They enable scalable, autonomous AI agents that operate continuously in enterprises. This approach focuses on optimizing performance per watt and cost per token for economic efficiency.
The US government plans to spend $9 billion on Nvidia superchips to advance its AI capabilities. This investment aims to keep the country competitive in the global AI arms race. Nvidia's technology is seen as crucial for maintaining leadership in AI development.
Telecommunications companies face challenges in adopting AI effectively. NVIDIA's 2025 report highlights key readiness gaps in the industry. Addressing these gaps is crucial for telcos to leverage AI benefits fully.
AWS demonstrates building high-performance generative AI systems using Strands Agents, NVIDIA NIM, and Amazon Bedrock AgentCore. The system supports parallel reasoning, context persistence, and traceable execution paths with GPU acceleration and serverless orchestration. This architecture enhances scalability, performance, and observability for...
Nemotron-Labs has developed a new diffusion language model that aims to achieve text generation at near speed-of-light speeds. This approach leverages diffusion techniques to improve generation efficiency and quality. The innovation could significantly accelerate natural language processing tasks.
NVIDIA GTC Taipei at COMPUTEX 2026 brings together developers, researchers, and industry leaders to explore the latest AI breakthroughs. The event covers diverse topics including AI factories, scaling infrastructure, and agentic AI. These advancements highlight the evolving landscape of AI technology across multiple industries.
NVIDIA has delivered its first Vera CPUs, designed specifically for AI agents, to leading AI labs including Anthropic, OpenAI, and SpaceXAI. These CPUs aim to accelerate agent-based AI workloads. The delivery also included Oracle Cloud Infrastructure, highlighting broad industry adoption.
NVIDIA Cosmos Predict 2.5 has been fine-tuned using LoRA and DoRA techniques to improve robot video generation. This advancement enhances the model's ability to create more accurate and realistic robotic videos. The update is significant for robotics and AI video generation applications.
