Latest NVIDIA News
Nvidia may guarantee $250bn for OpenAI data center
Nvidia is reportedly in talks to provide $250 billion in financing guarantees to help OpenAI lease a massive 10-gigawatt data center campus in Ohio. This move aims to secure the necessary infrastructure for OpenAI's future compute-intensive AI models.
Spirit v1.6 Briefly Tops RoboArena
Chinese physical AI startup Spirit AI briefly surpassed Nvidia on RoboArena, a global benchmark for embodied AI, with its Spirit v1.6 model. The ranking result has triggered controversy over possible manipulation and highlights the difficulty of evaluating autonomous systems.
NVIDIA Releases Full-Duplex VoiceChat 11B
NVIDIA has listed the NemotronLabs VoiceChat 11B model on Hugging Face, highlighting full-duplex voice interaction. The post provides limited details but signals a new open model option for real-time conversational audio applications.
NVIDIA Vera Storage Speeds Up AI Data Protection
NVIDIA presents benchmarks for Vera Storage focused on encryption, compression, integrity checking, and recovery in AI-native storage environments. The update highlights storage as a continuous component of agentic AI workflows, where repeated data operations support retrieval, memory, KV-cache reuse, and tool execution.
onsemi Profit Jumps 33% as AI Demand Accelerates
onsemi reported $1.604 billion in fiscal Q2 2026 revenue and $227 million in net income, up 9.18% and 33.18% year over year. The company expects AI data-center revenue to double in fiscal 2026 and is expanding its power portfolio, cloud infrastructure partnerships, and NVIDIA MGX ecosystem presence.
Isolated Kubernetes Tenants on Shared GPUs
NVIDIA Developer Blog explains how organizations can run isolated tenant Kubernetes clusters on shared GPU infrastructure. The approach addresses coordination problems caused by shared clusters, including conflicting CRD versions, overlapping RBAC, and the lack of team-level GPU budgets.