Latest news with #CUDA-X
Geek Culture
22-05-2025
- Business
- Geek Culture
Computex 2025: NVIDIA Supercharges AI With World's Largest Quantum Supercomputer, AI Infrastructure & More
NVIDIA CEO Jensen Huang kicked off Computex 2025 in Taipei with a bold vision for an AI-driven future. He emphasized that AI is now fundamental infrastructure, likening it to electricity and the internet, and described how AI data centers are evolving into 'AI factories' that generate valuable tokens. 'They're not data centers of the past,' Huang explained. 'These AI data centers, if you will, are improperly described. They are, in fact, AI factories. You apply energy to it, and it produces something incredibly valuable, and these things are called tokens.' Huang showcased NVIDIA's CUDA-X platform, highlighting its role in advancing AI, 6G, and quantum supercomputing. He outlined the growing power of AI, leading to 'agentic AI,' which can think and act, and 'physical AI,' which understands the world—culminating in general robotics. To meet surging demand, Huang announced innovations such as NVIDIA Blackwell-powered systems, NVLink Fusion to optimise AI data centers, and collaborations with major global tech players. He also detailed NVIDIA's partnership with Taiwan, including an AI supercomputer project with Foxconn and the Taiwan government. New product announcements included DGX Spark, a personal AI supercomputer; DGX Station, capable of running trillion-parameter models; NVIDIA RTX PRO Servers for enterprise AI; and NVIDIA AI Data Platform for intelligent storage. He also introduced tools for humanoid robotics, including Isaac GR00T-Dreams and the GR00T N1.5 foundation model. The NVIDIA DGX Spark and DGX Station. Huang shares that companies worldwide are investing $5 trillion in new factories, and optimizing their design is crucial for maximising efficiency. Leading Taiwanese manufacturers, including TSMC, Foxconn, and others, are leveraging NVIDIA Omniverse to create digital twins, driving the next phase of industrial physical AI in semiconductor and electronics production. Quantum Research Supercomputer. He also unveiled the opening of the Global Research and Development Center for Business by Quantum-AI Technology (G-QuAT), home to ABCI-Q, the world's largest quantum research supercomputer. This breakthrough system, powered by 2,020 NVIDIA H100 GPUs and Quantum-2 InfiniBand networking, aims to accelerate quantum-GPU computing to tackle complex challenges in industries like healthcare, energy, and finance. This supercomputer will enable researchers in Japan to explore core challenges in quantum computing and drive progress toward real-world use cases. Huang ended the keynote by unveiling NVIDIA Constellation, a new Taiwan office to be built in Beitou-Shiulin, and emphasising Taiwan's crucial role in AI development. He described the industry's trajectory as a 'once in a lifetime opportunity' to build AI factories, AI agents, and robotics within a unified architecture. Yonk is a geek who is fortunate enough to have an equally geeky Star Wars fan for a wife, who owns a LEGO Millennium Falcon encased in a glass coffee table as their home's centre-piece. AI Artificial Intelligence Computex Computex 2025 NVIDIA
TECHx
20-05-2025
- Business
- TECHx
AI Factories Are Real and Expanding: Jensen Huang
Home » COMPUTEX » COMPUTEX 2025 » AI Factories Are Real and Expanding: Jensen Huang At COMPUTEX 2025, NVIDIA founder Jensen Huang delivered a groundbreaking keynote. He revealed a new era where AI factories replace traditional data centers. This shift is changing everything in computing. Jensen Huang opened with a bold claim: intelligence is the new industrial output. The old world of general-purpose CPUs is giving way to accelerated computing powered by GPUs. Data centers no longer just store data. They actively produce intelligence. Traditional CPUs have reached their limits. According to Huang, they cannot keep up with AI's growing demands. NVIDIA's accelerated platforms, such as the upcoming GeForce RTX 5060 and next-gen DGX systems, unlock new performance levels. Moreover, Huang introduced NVIDIA's east-west networking architecture. This innovation connects thousands of AI processors across large clusters. Thanks to Mellanox technology, NVIDIA now leads in networking critical for AI factories. What sets Huang's keynote apart is his vision beyond technology. He described how AI tools are becoming accessible to creators, researchers, and industries worldwide. From healthcare breakthroughs to digital art and autonomous machines, AI is transforming sectors at scale. Huang emphasized democratizing intelligence. Anyone, anywhere, will soon be able to build, train, and deploy AI models. This accessibility marks a major milestone in AI development. He also highlighted NVIDIA's CUDA-X platform. Huang showcased a vast ecosystem of partners using CUDA-X for applications ranging from AI-driven 6G to quantum supercomputing. He explained how a growing install base fuels more innovation and benefits. Looking ahead, Huang outlined AI's future phases: agentic AI that thinks and acts, physical AI that understands the real world, and general robotics. These advancements drive an increasing demand for computing power. To meet this demand, Huang shared NVIDIA's latest innovations, including Grace Blackwell NVL72 systems and advanced networking tech. He mentioned large AI factory investments by CoreWeave, Oracle, Microsoft, xAI, and others worldwide. 'Factories are built because you know the answer,' Huang said. 'The more you buy, the more you make.' Taiwan plays a central role in this revolution. Huang highlighted NVIDIA's partnership with Foxconn Hon Hai and the Taiwan government. Together, they are building an AI factory supercomputer powered by NVIDIA's Blackwell infrastructure. This system will support Taiwan's researchers, startups, and industries such as TSMC. 'Having world-class AI infrastructure in Taiwan is crucial,' Huang added. He also announced NVLink Fusion, a new architecture allowing hyperscalers to build semi-custom compute solutions. This technology removes traditional bottlenecks and enables scalable, optimized AI systems. Finally, Huang introduced NVIDIA Constellation, a new Taiwan office. A dramatic video showed NVIDIA's headquarters launching into space and landing in Taiwan, symbolizing growth and commitment. In closing, Huang thanked NVIDIA's partners and Taiwan's tech community. He called the AI opportunity 'extraordinary' and 'once in a lifetime.' 'We are creating a new industry supporting AI factories, agents, and robotics — all on one architecture,' he said. At COMPUTEX 2025, Jensen Huang proved the AI future is here. It's fast, intelligent, and transforming the world — one AI factory at a time.
Yahoo
18-03-2025
- Business
- Yahoo
Synopsys Accelerates Chip Design with NVIDIA Grace Blackwell and AI to Speed Electronic Design Automation
Announcing Up to 30x Projected Performance Gains on the NVIDIA Grace Blackwell Platform to Speed Circuit Simulations for Next-Generation Semiconductors News Highlights: Today's GTC keynote featured Synopsys as part of the ecosystem with full-stack EDA suite acceleration across design, verification, and manufacturing on NVIDIA GPU and NVIDIA CUDA-X libraries. Leveraging the NVIDIA GB200 Grace Blackwell Superchip, Synopsys PrimeSim is projected to accelerate circuit simulation by up to 30x.1 Leveraging the NVIDIA B200 Blackwell architecture, Synopsys Proteus is projected to accelerate computational lithography simulations by up to 20x.1 NVIDIA NIM inference microservices integration driving 2x faster time-to-information with generative AI-powered Copilot.2 Optimizing more than 15 Synopsys solutions for NVIDIA Grace CPU in 2025. SUNNYVALE, Calif., March 18, 2025 /PRNewswire/ -- Synopsys Inc. (Nasdaq: SNPS) today announced the next phase of its work with NVIDIA to accelerate chip design up to 30x with the NVIDIA Grace Blackwell platform. To achieve this speedup, Synopsys announced at the GTC global AI conference that it is using NVIDIA CUDA-X libraries to optimize its solutions for next-generation semiconductor development. The company is also expanding support for the NVIDIA Grace CPU architecture and enabling more than 15 Synopsys solutions in 2025. "At GTC, we're unveiling the latest performance results observed across our leading portfolio when optimizing Synopsys solutions for the NVIDIA Blackwell platform to accelerate computationally-intensive chip design workflows," said Sassine Ghazi, president and CEO of Synopsys. "Synopsys technology is mission-critical to the productivity and capabilities of engineering teams from silicon to systems. By harnessing the performance of NVIDIA accelerated computing, we can help customers unlock new breakthroughs and deliver innovation even faster." "Chip design is one of the most complex engineering challenges in human history," said Jensen Huang, founder and CEO of NVIDIA. "With NVIDIA Blackwell and CUDA-X, Synopsys is cutting simulation times from days to hours—advancing chip design to power the AI revolution." Synopsys and NVIDIA are advancing a multi-year collaborative effort to accelerate electronic design automation (EDA) workloads. Synopsys is further applying NVIDIA accelerated compute architectures, including the NVIDIA GB200 Grace Blackwell Superchip, to achieve significant, projected runtime gains for workflows including circuit simulation, computational lithography, Technology Computer-Aided Design (TCAD), physical verification, and materials engineering. These accelerated workflows include: Circuit Simulation: Synopsys PrimeSim™ SPICE simulation workloads are projected to achieve a 30x speed up utilizing the NVIDIA Grace Blackwell platform. Today, customers can achieve up to 15x speed up utilizing NVIDIA GH200 Superchips. NVIDIA accelerated computing architectures enable the simulation of challenging circuits to achieve signoff with SPICE-level accuracy, reducing runtimes from days to hours. Computational Lithography: The production-proven choice for accelerating computational lithography for more than two decades, Synopsys Proteus™ provides optical proximity correction (OPC) software and inverse imaging technology (ILT) to resolve challenges at leading technology nodes. By leveraging NVIDIA technologies, Synopsys is delivering game-changing technology to advance this computationally-intensive manufacturing process. Today, Synopsys Proteus is optimized for NVIDIA H100 GPUs and integrated with the NVIDIA cuLitho library, achieving a 15x speed up of OPC. Leveraging the NVIDIA Blackwell platform, Synopsys Proteus is expected to accelerate computational lithography simulations by up to 20x. TCAD Simulation: Early results applying GPU-enabled capabilities and NVIDIA CUDA-X libraries to the Synopsys Sentauras™ TCAD process and device simulation solution is projected to accelerate time to results up to 10x. This solution is currently under development and is expected to be available to customers later this year. Materials Engineering: Synopsys QuantumATK® delivers atomic-scale modeling for semiconductor and materials research and development. Utilizing CUDA-X libraries on the NVIDIA Hopper architecture can accelerate time to results up to 100x, enabling customers to simulate and analyze a wide range of materials with greater efficiency. Synopsys plans to continue enabling accelerated computing on NVIDIA platforms throughout its portfolio. Advancing Synopsys Solutions with NVIDIA AI SoftwareSynopsys and NVDIA's efforts to accelerate chip design extend to speeding chip design with generative AI using NVIDIA NIM microservices: Generative AI Software for Chip Design: Today, customers using Synopsys' Gen AI-powered knowledge assistant, Copilot, are realizing an average 2x productivity improvement compared to prior methods. The integration of NVIDIA NIM microservices is projected to enable an additional 2x speedup for even faster time to answers.2 Optimizing Synopsys EDA with Grace CPU Additionally, Synopsys is enabling more than 15 solutions using the Grace CPU architecture for Synopsys EDA workloads spanning circuit simulation, physical verification, static timing analysis, and functional verification. The company plans to further increase support on the Grace CPU architecture in 2025. Synopsys at GTC 2025Synopsys is demonstrating at GTC 2025 March 18th through 21st in booth #222 in the Design and Simulation Pavilion. The company is also presenting sessions on semiconductor manufacturing and materials engineering and AI-driven chip design. For more information about Synopsys' presence at GTC, visit 1 Compared to CPU-based models.2 Copilot knowledge assistance without NVIDIA NIM integration. About SynopsysCatalyzing the era of pervasive intelligence, Synopsys, Inc. (Nasdaq: SNPS) delivers trusted and comprehensive silicon to systems design solutions, from electronic design automation to silicon IP and system verification and validation. We partner closely with semiconductor and systems customers across a wide range of industries to maximize their R&D capability and productivity, powering innovation today that ignites the ingenuity of tomorrow. Learn more at © 2025 Synopsys, Inc. All rights reserved. Synopsys, the Synopsys logo, and other Synopsys trademarks are available at Other company or product names may be trademarks of their respective owners. Forward-Looking StatementsThis press release contains forward-looking statements, including, but not limited to, statements regarding strategies related to products, technology and services and planned product releases and capabilities. These statements involve risks, uncertainties and other factors that could cause our actual results, time frames or achievements to differ materially from those expressed or implied in such forward-looking statements. Such risks, uncertainties and factors that could affect Synopsys' results is included in filings we make with the SEC from time to time, including in the sections entitled "Risk Factors" in our latest Annual Report on Form 10-K and in our latest Quarterly Report on Form 10-Q. The information provided herein is as of March 18, 2025. Synopsys undertakes no duty to, and does not intend to, update any forward-looking statement, whether as a result of new information, future events or otherwise, unless required by law. Editorial ContactKelli WheelerSynopsys, View original content to download multimedia: SOURCE Synopsys, Inc.
