Nvidia Corp - NVDA STOCK NEWS

NVIDIA has launched its H100 Tensor Core GPU, designed to meet the increasing demand for generative AI training and inference. Major cloud providers, including Oracle, AWS, and Microsoft Azure, are now offering H100 GPU-based services. The H100 technology boasts significant performance improvements, featuring nine times faster AI training and up to thirty times faster inference compared to the A100. Early adopters, such as OpenAI and Meta, are leveraging the H100 for their AI initiatives. The DGX H100 supercomputers are now in full production and available worldwide, indicating strong market demand for advanced AI computing solutions.

NVIDIA has launched four new inference platforms designed for generative AI applications, enhancing developers' capabilities to create specialized AI services. These platforms utilize the latest NVIDIA Ada, Hopper, and Grace Hopper processors, including the L4 Tensor Core GPU and H100 NVL GPU. Notably, the L4 GPU offers 120x more AI video performance than CPUs, while the H100 NVL is optimized for large language models like ChatGPT, delivering up to 12x faster performance than its predecessor. Google Cloud is integrating these platforms into its machine learning service, Vertex AI, with early adopter organizations already leveraging their capabilities.

NVIDIA has launched the BioNeMo Cloud Service, expanding its generative AI cloud services aimed at revolutionizing drug discovery, protein engineering, and life sciences research. This service allows researchers to customize AI models using proprietary data, accelerating the drug development process. Early access customers like Amgen and several startups are already utilizing the platform for significant advancements in biologics discovery. BioNeMo includes pretrained AI models that facilitate rapid identification and design of drug molecules, potentially reducing costs and time in drug development.

NVIDIA has launched NVIDIA AI Foundations, a suite of cloud services enabling enterprises to create and customize generative AI models tailored to specific needs. Esteemed companies like Getty Images, Morningstar, and Shutterstock will leverage these services for applications across language, images, video, and 3D. The NeMo and Picasso services facilitate the development of proprietary AI applications, enhancing productivity in creative workflows. Additionally, notable partnerships, including one with Adobe, aim to redefine generative AI in industry.

NVIDIA has launched its DGX Cloud, an AI supercomputing service that enables enterprises to access dedicated clusters of NVIDIA DGX AI supercomputing resources via a web browser. This service streamlines the process of training advanced models for generative AI. Initially hosted on Oracle Cloud Infrastructure, it will expand to Microsoft Azure and Google Cloud. Notable clients like Amgen are utilizing DGX Cloud to enhance drug discovery processes, achieving training speeds up to 3x faster. Instances start at $36,999 per month, allowing organizations to scale AI development efficiently.

NVIDIA has partnered with Microsoft to offer cloud-based services through Microsoft Azure, giving enterprise users access to advanced AI supercomputing and industrial metaverse applications. The collaboration includes NVIDIA DGX Cloud, which will provide AI supercomputing capabilities, and NVIDIA Omniverse Cloud, a platform for developing 3D applications. Additionally, Microsoft 365 applications will be integrated with the Omniverse platform, enhancing digitalization processes across enterprises. With services available by mid-2023, this partnership aims to transform business operations by combining real-time data with advanced AI and digital twin technologies.

NVIDIA announced that Oracle Cloud Infrastructure (OCI) has integrated the BlueField-3 Data Processing Units (DPUs) into its networking stack. This third-generation DPU aims to offload data center tasks from CPUs, thereby enhancing performance, efficiency, and security. BlueField-3 enables up to 24% power reduction and increases compute power and storage processing capabilities up to 4x compared to prior models. NVIDIA's advancements are intended to meet the increasing demands of AI workloads and solidify OCI's position in sustainable, high-performance cloud infrastructure.

NVIDIA has made a significant breakthrough in computational lithography with its new cuLitho software library, which is being integrated into processes by industry leaders like TSMC, ASML, and Synopsys. This advancement aims to push the limits of semiconductor manufacturing as current methods reach physical constraints. cuLitho offers performance improvements of up to 40x, enabling greater energy efficiency and faster production rates, potentially allowing for 3-5 times more photomasks daily while using 9x less power. This innovation supports the development of 2nm technology and promises to reduce bottlenecks in chip design.

NVIDIA has partnered with Medtronic to enhance AI integration in healthcare, focusing on the GI Genius™ intelligent endoscopy module, the first FDA-cleared AI-assisted colonoscopy tool. This collaboration aims to improve patient care through NVIDIA's Clara™ healthcare platform, facilitating the development of real-time AI algorithms for diagnostic procedures. The integration of NVIDIA Holoscan and IGX technologies will support enhanced imaging and diagnostic capabilities. Initial systems utilizing this technology are set to launch later this year, marking a significant step in AI-enabled medical devices.