Desk
NVIDIA hardware, software, and developer news.
NVIDIA cut the cost of making giant training runs exactly reproducible to about 2 percent, which turns a debugging luxury into something labs can afford to leave switched on.
NVIDIA is consolidating GPU-driven networking into one shared layer, a dull-sounding cleanup that makes AI clusters faster and NVIDIA's stack even harder to step away from.
Look past the trillion-parameter headline: Le Chonk's sharpest claim is that it will do defensive security work closed models refuse, and that claim carries both promise and risk.
NVIDIA's green contexts let one program carve a GPU into dedicated lanes, and the latency win for urgent work is big enough that inference and robotics teams should take notice.
NVIDIA says telecom carriers are betting on open models for control and local fit, a sensible trend from a vendor with an obvious stake in how it's framed.
A new walkthrough shows how to train a robot policy from NVIDIA’s huge Cosmos3-DROID dataset without downloading it, a quiet but meaningful step toward making robot learning affordable outside big labs.
NVIDIA is showcasing startups using AI at every step of breast cancer care, from two-minute wearable ultrasound to pathology-based treatment predictions — promising, but it’s a vendor’s highlight reel, not a trial readout.
MarkTechPost’s IsaacTeleop walkthrough is a Colab tutorial with no headset: NumPy-built inputs, a graph retargeter, and the plumbing that turns XR hands into robot actions.
A seven-year-old streaming box just got a $100 sticker shock — and Nvidia’s own spokesperson pointed at memory prices driven by the AI boom it helps create.
IBM’s agentic coding platform Bob is GA for self-hosted and air-gapped shops — bring your own model, keep the code off the public internet.
Prime Intellect opens Prime Inference — serverless and reserved Blackwell serving for open frontier models, tuned for agentic turns and near-zero tool-call errors.
DOJ arrests Earthmade CEO Greg Lui over an alleged $300M+ Nvidia A100/H100 diversion ring via Malaysia and Singapore — while Nvidia calls smuggling a rounding error.