75 articles
NVIDIA has released its Nemotron 3 Nano model as a fully managed, serverless offering on Amazon Bedrock, delivering hybrid mixture-of-experts performance designed for building scalable AI multi-agent systems.
Fresh benchmark data reveals Nvidia's GB200 NVL72 cuts AI token costs dramatically compared to rivals, proving that raw performance matters more than sticker price in modern AI infrastructure.
Big Tech concentration has defined recent markets, with Nvidia leading the $3 trillion AI boom. But market history shows peak dominance often marks the beginning of decades-long underperformance.
Chinese scientists just dropped a bombshell optical computing chip that crushes Nvidia's top GPU by over 100x in speed and efficiency. This breakthrough is pushing AI adoption way beyond the usual suspects, fueling a 10% earnings jump in the S&P 500's Impressive-493 as competition heats up across industries.
Nvidia's pushing for 16-layer High Bandwidth Memory by late 2025, sparking an intense tech race between Samsung, SK Hynix, and Micron to crack the manufacturing challenges first.
Nvidia's entering a non-exclusive licensing deal with Groq, snagging its AI inference-chip technology and hiring key talent, including founder Jonathan Ross. The partnership boosts Nvidia's AI game without raising antitrust red flags.
Nvidia's new 4D-RGPT model transforms how AI understands video content, delivering a 5.3% accuracy improvement in depth and motion recognition across major benchmarks.
NVIDIA unveils Nemotron 3 Nano, a 30-billion parameter AI model that activates only 3 billion parameters per pass, achieving 3.3x higher throughput than competitors while handling 1-million token contexts for advanced agentic reasoning tasks.
New TurboDiffusion framework delivers 100–200× faster AI video generation on NVIDIA RTX 5090, reducing creation time for 5-second clips to under 2 seconds through advanced optimization techniques.
Nvidia has reportedly agreed to acquire AI chip startup Groq for $20 billion in cash, according to CNBC coverage. The deal would exclude Groq's cloud business, focusing instead on the company's hardware technology and engineering team.
ByteDance and Tsinghua researchers claim their new TurboDiffusion framework accelerates AI video generation by up to 199× on a single Nvidia RTX 5090, potentially bringing real-time video creation to consumer GPUs.
Fresh benchmark data reveals Nvidia's Blackwell GB300 NVL72 delivers up to 1.9× faster training speeds per chip than Google's Ironwood TPU when running complex AI workloads.
New research reveals SonicMoE method delivers major efficiency improvements for AI models on Nvidia H100 GPUs, boosting speed while cutting memory usage nearly in half.
NVIDIA just dropped Nemotron 3 Nano, a ~30B mixture-of-experts model that scored 52 on the Artificial Analysis Intelligence Index while keeping inference lean with only 3.6 billion active parameters.
Nvidia's latest CUDA update brings new developer tools and performance features that help maintain the company's lead in AI infrastructure, even as competitors like AMD and Google push their own accelerator platforms.
A fresh report claims China's DeepSeek trained its upcoming AI model using Nvidia's Blackwell chips that were smuggled into the country, dodging U.S. export restrictions. The chips reportedly traveled through third countries before being taken apart and shipped into China piece by piece.
Over 7 gigawatts of new data center capacity will come online in 2025, with another 10 gigawatts breaking ground. Amazon, Nvidia, and other tech giants are racing to expand infrastructure as AI computing needs skyrocket.
NVIDIA rolled out GenMol v2 with a redesigned SAFE syntax that uses angle brackets to map molecular connections more clearly, streamlining three key research workflows.
NVIDIA introduces Alpamayo-R1, an open-source AI model that enhances autonomous vehicle reasoning by moving beyond simple object detection to context-aware understanding of real-world driving scenarios.
Fresh benchmark data reveals NVIDIA H100 and B200 delivering the lowest inference costs for Llama 3.3 70B, with Google TPU v6e and AMD MI300X falling behind in tokens-per-dollar efficiency.
We use cookies to improve your experience on our site and to show you relevant advertising. To find our more, read our privacy policy and cookie policy
Marina Lyubimova
Peter Smith
Eseandre Mordi
Victoria Bazir
Saad Ullah
Usman Salis
Sergey Diakov
