NVIDIA AI News - Page 8 of 16
310 articles • Page 8 of 16
Blackwell Ultra speeds up AI; Nvidia Rubin platform slated for months-away launch
Nvidia's current AI accelerator is getting faster. Its next one will make it look slow.
Mythic raises USD 125M to scale US-made AI chips that claim 100× NVIDIA efficiency
Forget about beating NVIDIA at its own game. The real threat might come from ignoring the game entirely.
AMD Bets on Software to Close NVIDIA's CUDA Ecosystem Lead
For years, the GPU war was fought on raw teraflops. But the battle lines have shifted.
DFlash speculative decoding boosts NVIDIA Blackwell inference up to 15×
NVIDIA Blackwell delivers 15 petaflops of dense NVFP4 compute, a staggering amount of raw power.
Anthropic may keep supplying Claude to NSA despite Pentagon risk flag
The Pentagon flagged Anthropic as a supply chain threat. The NSA needs chips it doesn’t have. And yet, a deal is nearly done.
Attention output GEMM reduces blended Fprop speedup to 1.47× in NVFP4 training
The 1.47x speedup for NVFP4 training isn't fake. It's just the real answer after you pay the bill.
Red Hat unveils AI 3, hybrid cloud-native platform for enterprise inference
The enterprise AI landscape is getting another heavyweight contender. Red Hat, known for its open-source infrastructure solutions, is stepping into...
TokenSpeed-Kernel Delivers Top Performance on AMD GPT-OSS 120B via Gluon Kernels
LLM models and inference hardware are changing at breakneck pace. Why does that matter? Because speed alone isn’t enough any more.
Musk overhauls xAI as Nvidia unveils Nemotron 3 Super, a 120B reasoning model
Elon Musk is rebuilding his artificial intelligence startup xAI from the ground up, a person familiar with the matter told The Rundown AI.
NVIDIA JetPack 7.2 lets developers quickly deploy Yocto‑based AI at the edge
Edge AI development has long been a fragmented headache. Different hardware meant a fresh software puzzle each time—a maddening scramble of kernels...
Nvidia invests USD 4 B in photonics, taps Lumentum and Coherent optics for AI GPUs
Nvidia just dropped $4 billion on photonics. Why? Because electrons are no longer fast enough.
NVIDIA releases Cosmos 3 with Super‑Text2Image and Nano‑Policy‑DROID
NVIDIA’s Cosmos 3 isn’t just another model drop. It’s a two-tower mixture-of-transformers foundation model that fuses physical reasoning, world...
ComputeEval 2025.2 expands to 232 CUDA challenges, upping LLM test difficulty
The standard tests for large language models used to be forgiving. Not anymore. ComputeEval 2025.2, NVIDIA’s new benchmark, just made them brutally...
Rubin Observatory sends 800,000 alerts on first night, reaching astronomers in minutes
The Vera C. Rubin Observatory logged 800,000 astronomical alerts on the first night its new system ran.
Arm's first AGI CPU, up to 136 cores, to power Meta AI datacenters this year
Arm isn't just renting out the blueprints anymore. It’s pouring the foundation. This year, its debut CPU—packing up to 136 cores—will land inside...
Omniverse Workflows Boost Vision AI Accuracy Using Synthetic Data, Fine‑Tuning
Vision AI models fail in boring, predictable ways. They choke on a new camera angle, a weirdly lit warehouse, a product they haven't seen before.
NVIDIA BioNeMo Toolkit Enables AI Scientist to Align, Fold, and Dock Molecules
Lab work is slow. The AI scientist is not. It runs without sleep or salary, folding proteins and aligning sequences in a silent digital loop.
Meta plans facial recognition for AI smart glasses, amid privacy concerns
Meta is quietly reopening the door to facial recognition, this time through the lens of its AI-powered smart glasses.
NVIDIA releases NvRTX 5.7.4 with DLSS 4.5 support for UE5.7.4
Nvidia's latest update is a patch note pretending to be a physics paper. The NvRTX 5.7.4 release adds DLSS 4.5 support for Unreal Engine 5.7.4.
NVIDIA MCG Toolkit hits 61% completion, parsing code, configs, repo structure
Nobody documents their AI work. The model seems to function just fine without it.