Nvidia's Nemotron-Cascade 2 wins math and coding gold medals with 3B active parameters — and its post-training recipe is now open-source

Nvidia's Nemotron-Cascade 2 is a 30B MoE model that activates only 3B parameters at inference time, yet achieved gold ...

Nvidia's DGX Station is a desktop supercomputer that runs trillion-parameter AI models without the cloud

Nvidia introduced the DGX Station at GTC 2026, a desktop supercomputer with 20 petaflops of AI performance and 748GB of ...
Computer Weekly

CUDA at 20: From billion-dollar gamble to agentic AI

As Nvidia marks two decades of CUDA, its head of high-performance computing and hyperscale reflects on the platform’s journey ...
Hosted on MSN

Cursor claims Nvidia engineers produce three times more code with AI tools

Cursor reports Nvidia engineers now commit three times more code than before Nvidia maintains that defect rates stayed flat despite the reported surge in output AI-assisted workflows contributed to ...

Nvidia Moves Beyond Chips With An Open-Source Platform For AI Agents

Nvidia's NemoClaw launch at GTC creates permanent chip demand—AI agents require far more compute than chatbots. Position now.

Nvidia Is Quietly Rewriting Biology

GPU behemoth Nvidia’s most important metric isn’t its data center revenue, hyperscaler capex trends, or the yield on its ...
Fudzilla

Nvidia’s anti-mining code is now broken

Nvidia’s code which borks its cards if they are used for Ethereum mining has been cracked. When Nvidia launched its Ampere Lite Hash Rate (LHR) graphics card with the feared Ethereum anti-mining ...