NVIDIA Research predicts what’s next in AI, from better weather predictions to digital humans

NVIDIA’s CEO, Jensen Huang, revealed some of the most exciting technological innovations during his keynote presentation at Computex 2024 in early June. 

Surpassing Apple to become the second most valuable company in the U.S., NVIDIA and their research team is working on finding breakthroughs that will drive further adoption of AI.

Below we explain a few of NVIDIA’s innovations mentioned during the presentation and briefly describe what makes them important for the future of AI.

1. Blackwell and GB200 GPUs

NVIDIA’s new generation of GPUs, known as Blackwell, are extremely powerful Graphics Processing Units (GPUs). They offer training performance that is 4x better than previous generations, 30x the inference performance, and 25x better energy efficiency. 

Why is it important? With MoE (mixture-of-expert) models like GPT-4 (which relies on high inference performance) becoming more common, tech companies, research institutions, and any sector making major investments in AI will benefit from increased performance and lower energy costs.

Fun Fact: NVIDIA’s new GPU, named Blackwell, is named after mathematician and statistician David Blackwell. Blackwell was one of the first African Americans admitted to the National Academy of Sciences in the U.S. He made significant contributions to game theory and statistics, which have been essential in the development of advanced AI and machine learning technologies.


2. Rubin AI Chip Architecture

After Blackwell, Rubin AI is set to be NVIDIA’s next-generation AI chip architecture (it’s planned to be released in 2026). NVIDIA currently dominates the AI chip market with an 80 percent share. News about Rubin AI and the plan to have annual chip releases shows NVIDIA’s commitment to innovation and to remaining the market leader. 

Why is it important? Rubin AI aims to meet the growing demand for AI hardware in the coming years. It will comprise new GPUs, CPUs and networks designed to power AI applications. 

3. Digital Humans 

Digital Humans are virtual representations of people through AI-powered text, speech, animations and graphics. These digital humans can interact naturally with users, offering personalized, realistic and conversational responses. We can think of them as very advanced virtual assistants that can understand and respond in a more human-like way.

Why is it important? In a wide variety of industries, a digital human can help users find the right products, care, answers and entertainment, enhancing their overall digital experience. This would allow companies and providers to provide better service while reducing costs.

4. Earth-2

Earth-2 is a “digital twin” of Earth. This full-stack, open platform allows for detailed weather simulations using AI. This innovation has the potential to transform our understanding of weather and the environment. 

Why is it important? With Earth-2, we could simulate and predict how extreme events like hurricanes, droughts or floods (as well as climate change) could affect different regions of the world, helping us to prepare better and respond.

Fun Fact: NASA employee John Vickers coined the term “digital twin” in 2010 in a report on the space agency’s technology roadmap. Digital twins can create virtual replicas of physical objects, which allows for more precise testing and simulations without the real, physical risks.


5. NIMs

NIMs (NVIDIA Inference Microservices) are virtualized containers that can host pre-trained proprietary and open-source large language models (LLMs). They allow developers to create custom generative AI applications more easily and efficiently.

Why is it important? NIMs aim to create a pathway for an AI app marketplace where creators could set up their own AI models that could blend into other existing models. This would simplify the development of customized AI solutions for individuals and companies.

6. Spectrum-X networking platform and DGX SuperPOD supercomputers

Another important innovation that NVIDIA presented was the company’s advancements in networks and supercomputers, like the Spectrum-X platform and DGX SuperPOD supercomputers. Meant to manage AI applications and generative AI workloads, these computer systems can deliver high-performance AI, machine learning and natural language processing.

Why is it important? These improvements in networking and supercomputers will help push the boundaries of AI and high-performance computing with faster processing, greater scalability and more cohesive operations for large companies, research institutions, governments and more.

Watch NVIDIA’s CEO Jensen Huang full keynote at COMPUTEX 2024.

Making AI and data processing more accessible

Through NVIDIA’s advancements, we can envision a future where AI and data processing are more powerful, efficient, and accessible to everyone. If you’re looking for the critical computing resources needed to succeed in an AI landscape, get started with our cloud GPU service.