News from the AI & ML world
DeeperML
Dashveenjit Kaur@TechHQ
//
Nvidia is significantly expanding its "AI Factory" offerings, collaborating with Dell Technologies to power the next wave of AI infrastructure. This includes the development of the next-generation NERSC-10 supercomputer, dubbed Doudna, based on the Nvidia Vera Rubin supercomputer architecture. This new system is designed to accelerate scientific discovery across fields such as fusion energy, astronomy, and life sciences, benefiting around 11,000 researchers. The Doudna supercomputer will be built by Dell and aims to deliver a tenfold increase in scientific output compared to its predecessor, Perlmutter, while significantly improving power efficiency.
The Doudna supercomputer represents a major investment in scientific computing infrastructure and is named after CRISPR pioneer Jennifer Doudna. Unlike traditional supercomputers, Doudna's architecture prioritizes coherent memory access between CPUs and GPUs, enabling efficient data sharing for modern AI-accelerated scientific workflows. This innovative design integrates simulation, machine learning, and quantum algorithm development, areas increasingly defining cutting-edge research. The supercomputer is expected to be deployed in 2026 and is crucial for maintaining American technological leadership in the face of escalating global competition in AI and quantum computing.
In addition to expanding AI infrastructure, Nvidia is also addressing the China AI market by working with AMD to develop export rules-compliant chips. This move comes as the U.S. government has restricted the export of certain high-performance GPUs to China. Nvidia CEO Jensen Huang has emphasized the strategic importance of the Chinese market, noting that a significant portion of AI developers reside there. By creating chips that adhere to U.S. trade restrictions, Nvidia aims to continue serving the Chinese market while ensuring that AI development continues on Nvidia's CUDA platform.
ImgSrc: techhq.com
References :
The Doudna supercomputer represents a major investment in scientific computing infrastructure and is named after CRISPR pioneer Jennifer Doudna. Unlike traditional supercomputers, Doudna's architecture prioritizes coherent memory access between CPUs and GPUs, enabling efficient data sharing for modern AI-accelerated scientific workflows. This innovative design integrates simulation, machine learning, and quantum algorithm development, areas increasingly defining cutting-edge research. The supercomputer is expected to be deployed in 2026 and is crucial for maintaining American technological leadership in the face of escalating global competition in AI and quantum computing.
In addition to expanding AI infrastructure, Nvidia is also addressing the China AI market by working with AMD to develop export rules-compliant chips. This move comes as the U.S. government has restricted the export of certain high-performance GPUs to China. Nvidia CEO Jensen Huang has emphasized the strategic importance of the Chinese market, noting that a significant portion of AI developers reside there. By creating chips that adhere to U.S. trade restrictions, Nvidia aims to continue serving the Chinese market while ensuring that AI development continues on Nvidia's CUDA platform.
ImgSrc: techhq.com
Share:
References :
- futurumgroup.com: Olivier Blanchard, Research Director at Futurum, shares insights on Dell and NVIDIA’s upgraded AI Factory and how it enables enterprises to deploy high-performance, full-stack AI infrastructure with integrated services and tools.
- insideAI News: Report: NVIDIA and AMD Devising Export Rules-Compliant Chips for China AI Market
Classification:
- HashTags: #NvidiaAI #Supercomputing #ChinaAIMarket
- Company: Nvidia
- Target: Researchers, Enterprises
- Product: AI Factory
- Feature: AI infrastructure, Supercomput
- Type: AI
- Severity: Informative