NVIDIA CEO Jensen Huang Unveils Major AI and Computing Innovations at COMPUTEX

NVIDIA founder and CEO Jensen Huang emphasized the transformative potential of generative AI and accelerated computing during his keynote address ahead of COMPUTEX 2024 in Taipei. Speaking to an audience of over 6,500 industry leaders, press, and tech enthusiasts, Huang highlighted how these innovations are set to redefine the future of computing.

Generative AI and Accelerated Computing

Huang underscored the significance of generative AI, stating that it is reshaping industries and unlocking new opportunities for innovation and growth. He pointed out that NVIDIA's accelerated platforms are already in full production, featuring capabilities powered by NVIDIA RTX technology. This includes AI PCs, consumer devices, and enterprise AI factories leveraging NVIDIA’s comprehensive computing platform.

New Semiconductor Roadmap

In a significant announcement, Huang revealed a roadmap for new semiconductors arriving on a one-year cycle. The Rubin platform, set to succeed the upcoming Blackwell platform, will feature new GPUs, an Arm-based CPU named Vera, and advanced networking solutions like NVLink 6 and CX9 SuperNIC.

“Our company operates on a one-year rhythm, pushing technology to its limits,” explained Huang. NVIDIA’s creative team utilized AI tools from the NVIDIA Inception startup program to craft the COMPUTEX keynote, showcasing the transformative impact of their technology.

Sustainable Computing

Huang emphasized that “accelerated computing is sustainable computing,” highlighting how the combination of GPUs and CPUs can achieve up to a 100x speedup with only a threefold increase in power consumption. This results in 25x more performance per watt compared to CPUs alone, offering significant cost and energy savings.

Industry Collaboration

Major computer manufacturers, especially from Taiwan, have embraced NVIDIA’s GPUs and networking solutions. Companies such as ASRock Rack, ASUS, GIGABYTE, and Supermicro are developing cloud, on-premises, and edge AI systems. The NVIDIA MGX modular reference design platform now supports Blackwell, optimized for large language models and data processing.

AMD and Intel are also supporting the MGX architecture, allowing server system builders to use these reference designs to save development time while ensuring performance consistency.

Next-Generation Networking with Spectrum-X

Huang introduced annual releases for Spectrum-X products to meet the growing demand for high-performance Ethernet networking for AI. NVIDIA Spectrum-X enhances network performance by 1.6x compared to traditional Ethernet fabrics, accelerating AI workload processing and deployment.

CoreWeave, GMO Internet Group, Lambda, Scaleway, STPX Global, and Yotta are among the first AI cloud service providers to integrate Spectrum-X into their infrastructures.

NVIDIA NIM for AI Development

NVIDIA NIM (inference microservices) allows developers to create generative AI applications easily. These services can be deployed on clouds, data centers, or workstations, maximizing infrastructure investments. Enterprises running Meta Llama 3-8B in a NIM see up to 3x more generative AI tokens on accelerated infrastructure.

Nearly 200 technology partners, including Cadence and Cloudera, are integrating NIM into their platforms to expedite generative AI deployments for specific applications like digital human avatars and code assistants.

AI Assistants with GeForce RTX AI PCs

NVIDIA’s RTX AI PCs are set to revolutionize consumer experiences with over 200 AI laptops and more than 500 AI-powered apps and games. Microsoft and NVIDIA are collaborating to bring generative AI capabilities to Windows native and web apps.

Project G-Assist, an RTX-powered AI assistant technology demo, showcases context-aware assistance for PC games and applications.

NVIDIA Robotics

NVIDIA is leading the $50 trillion industrial digitization shift with autonomous operations and digital twins. Taiwanese manufacturers are transforming their factories using NVIDIA technology, such as Foxconn’s use of NVIDIA Omniverse for creating digital twins.

Huang emphasized the importance of robotics and AI, stating, “The next wave of AI is physical AI. AI that understands the laws of physics and can work among us.” The NVIDIA Isaac platform provides tools for developers to build AI robots, enhancing factory and warehouse efficiency.

Tribute to Taiwan

Huang concluded his keynote by paying tribute to Taiwan and NVIDIA’s many partners there, expressing his gratitude and admiration.