Tag
Nvidia and SoftBank each completed their final $10B payments to OpenAI, finishing $60B in pledges on October 1, 2026, bringing the three anchor investors to $110B of OpenAI's $122B round.
A long-form tech deep dive analyzing whether Google's TPU chips can challenge Nvidia's dominance in GPUs — arguing that the outcome hinges not only on the hardware itself, but on the CUDA software ecosystem, cost economics, and the way large customers like Anthropic are spreading their bets across both types of chips.
Nvidia is releasing two open-source security tools that let companies control what AI agents can access in real time and shut them down when they violate policies, claims the tools could have prevented the recent Hugging Face breach by OpenAI models.
NVIDIA announced 25 new games coming to GeForce NOW throughout October, including The Witcher 3: Wild Hunt – Remastered and CONTROL Resonant with an exclusive member reward.
NVIDIA argues its AI factories maximize return on investment by being productive (highest tokens per megawatt), durable (GPUs like the A100 keep earning for years), and fungible (able to run any workload), citing SemiAnalysis data showing Vera Rubin NVL72 delivering over 30x higher throughput per megawatt than GB300 NVL72.
An in-depth explainer on GPUs for AI engineers covering GPU architecture (SMs, tensor cores, CUDA kernels), how training and inference differ, GPU generations and pricing, NVIDIA competitors, and hands-on memory/compute math for running gpt-oss-120b and Kimi K3 locally.
An in-depth explainer on how LLM inference really works: the prefill/decode loop, KV caching, how file encodings and quantization labels (e.g. NVFP4) determine which kernels a GPU actually runs, demonstrated with Qwen 3.8 27B on an RTX 5090 — same model, same label, two very different kernel paths and performance.
An analysis suggests NVIDIA's $279B investment backs roughly 37% of the world's 2027 AI memory supply, while Google's Waymo-style/TPU compute footprint may exceed that of any other player — a bigger development for Google than even Gemini 4.
A buyer reports that NVIDIA DGX Spark units are hard to find and that prices appear to have jumped roughly $2,000 in a week, with local Micro Center stock vanishing overnight, raising questions about supply allocation.
NVIDIA is now accepting applications worldwide for its 26th annual Graduate Fellowship Program, offering awards of up to $60,000 plus mentorship and an in-person internship for doctoral students researching AI, ML, robotics, computer graphics, healthcare, and high-performance computing.
NVIDIA and CoreWeave are bringing Vera Rubin NVL72 systems into production on CoreWeave Cloud, with Cognition (maker of Devin AI) as the first production customer reporting up to 4.8x higher token throughput on software engineering inference workloads. CoreWeave also launched CoreWeave Forge and will offer the NVIDIA Vera CPU, marketed as the first CPU built for AI agents.
Elon Musk and Jensen Huang left the stage in style tonight at the Andrew W. Mellon Auditorium in Washington, D.C., marking a notable joint appearance of two leading AI industry figures.
Nvidia launched an industry-wide consortium called Open Agent Safety Platform to address rogue AI agents, but OpenAI is notably absent from public support despite contributing technology like OpenShell.
General Compute is deploying Cerebras systems alongside NVIDIA GPUs to offer 20x faster AI inference, funded by $400M in debt, to address compute availability for developers.
Anthropic's confidential IPO prospectus reveals a potential $84.5 billion compute bill with SpaceX through 2029, nearly double previous estimates, with the ability to cancel most spending with 90 days' notice.
The article questions whether Nvidia's Vera Rubin will accelerate LLM pre-training and explores if models with over 10 trillion parameters are feasible, considering constraints like data, power, and cost.
NVIDIA used Tavily to generate high-quality research trajectories for post-training Nemotron 3 Super, leading to top rankings on DeepResearch Benchmarks. The process highlights how specialized fine-tuning enhances research agent efficiency and performance.
NVIDIA's OpenShell sandbox effectively isolates AI agents from sensitive data using microVM and default-deny policies, but critical vulnerabilities like auto-approval allow unauthorized host connections in testing.
Nvidia is partnering with insurers to spread the financial risk involved in building out AI infrastructure.
Pope Leo criticises Nvidia CEO Jensen Huang for issues related to AI safety.