At the recent NVIDIA GTC conference, Jen-Hsun Huang showcased NVIDIA's latest advances in the areas of AI arithmetic, software services, the Omniverse cloud platform, and embodied intelligence, and declared a new era of AI technology.
Six Keywords
Blackwell architecture
-
Superchip named after mathematician David Blackwell - Blackwell -
The chip, which integrates 208 billion transistors, is a new computing platform that heralds a new era in AI and high-performance computing, capable of training a 1.8 trillion covariate GPT model in 90 days, with one-quarter the power consumption -
New GPU architecture optimized for large AI models improves training, inference performance and energy efficiency in the data center -
Featuring six revolutionary technologies, including improved GPU performance, high-speed data processing and security technologies
-
A set of optimized cloud-native microservices, model customization services based on the CUDA platform, designed to simplify and accelerate the deployment process of generative AI models, reduce time-to-market, and provide enterprises with fast, flexible AI solutions -
This new service model will revolutionize the way software is developed, and in the future developers may no longer need to write code line by line, but through communication with AI to mobilize a series of NIMs to complete the training and application deployment, making NVIDIA the "fab" of AI software.
-
NVIDIA has announced a embodied intelligence program called GR00T, which aims to drive groundbreaking work in robotics and embodied intelligence -
Dedicated to enabling humanoid robots to see and interact, showcasing NVIDIA's efforts to advance AGI (generalized artificial intelligence) -
Advances in Human-Robot Interaction Technology through Robot Understanding of Natural Language and Simulation of Human Actions
-
NVIDIA Updates Omniverse Cloud Platform to Simulate Products in a Fully Fidelity, Fully Digital Way, Opening New Possibilities for Digital Twin Technology to Improve Product Design Efficiency and Optimize Performance -
Through digital twin technology, the AI platform integrates digital simulation, enabling AI to conduct cyclic testing in an accurately simulated physical environment, providing a self-learning "gym" for robots, and promoting the learning and application of AI in the physical world.
-
An in-vehicle computing platform designed for generative AI applications and expected to integrate with the Blackwell architecture to deliver high-performance autonomous driving solutions -
As an advanced version of DRIVE Orin, DRIVE Thor boasts up to 2,000 TOPS of computing power on a single chip, realizing an eight-fold increase in performance -
This powerful chip is capable of providing the necessary computing support for both autonomous driving systems as well as in-car entertainment information systems. According to an announcement on NVIDIA's official website, DRIVE Thor is expected to enter mass production in 2025 -
So far, five Chinese automakers have confirmed plans to adopt this latest chip technology. BYD plans to adopt NVIDIA's technology to streamline its factory operations and supply chain processes, and for creating virtual display spaces; GAC EAN's premium brand Hao Platinum, as well as Xiaopeng Automobile, also announced on the same day that they will integrate the DRIVE Thor chip in their future models
-
The emergence of a new type of software, born out of breakthroughs in generative artificial intelligence technology, symbolizes the rise of a new industry, one that is computer-written, creating an unprecedented category of software that is very different from the data center work of the past -
By analogy with the discovery of electricity at the beginning of the industrial revolution, the new "electronic tokens" we are now creating, AI, produced by AI factories, have enormous value -
The AI-generating nature of the new software and its ability to be mass-produced heralds a change in the way computing is done and how software is built -
Nvidia's core values lie in the convergence of computer graphics, physics, and artificial intelligence, all of which are reflected in the Omniverse platform's virtual world simulations
Generative AI will lead the new industrial revolution

For more, please refer to
AIGC Big Model: ushering in an era of dual revolution in engineering and science
Which tool in the country can flat out replace ChatGPT?
AI+Healthcare Mega Modeling|Exploring the Future of AIGC in Domestic Healthcare
What is the best ai drawing tool to use?
gpts stunning debut: elite gathering, let countless intelligences be your backbone