At NVIDIA’s GTC 2024, founder and CEO Jensen Huang revealed groundbreaking advances in AI software and hardware with the potential to revolutionize industries from healthcare to manufacturing.
An overflow crowd of nearly 8,000 filled the keynote address, which highlighted NVIDIA’s latest AI platform powered by the new Hopper Super Chip.
“Hopper is a quantum leap for AI computing.”
Jensen Huang, Founder and CEO, NVIDIA
“Its efficiency, performance, and scalability will help companies of all sizes use AI to drive innovations that can change the world.”
Unmatched Speeds via Novel Chip Architecture
The highlight of GTC was the unveiling of NVIDIA’s Hopper GPU architecture, which achieves massive speedups for AI training and inference workloads.
The Hopper Super Chip integrates over 80 billion transistors and is manufactured using the latest 3nm process technology.
NVIDIA researchers optimized Hopper’s architecture specifically for AI, incorporating model and data parallelism techniques along with new Transformer Engine blocks.
Together, these advances enable Hopper-powered systems to train transformer-based AI models up to 15x faster compared to NVIDIA’s previous generation architecture.
Hopper Super Chips will be available in NVIDIA DGX and HGX AI supercomputers as well as cloud platforms in 2024, enabling more organizations to achieve state-of-the-art results with large language, recommendation, and computer vision models.
NVIDIA GTC 2024 – AI Innovation with Omniverse and BioNeMo
In addition to new hardware, NVIDIA announced major updates to its AI software stack spanning tools like Omniverse, a simulated 3D design collaboration platform, and BioNeMo, a toolkit for protein sequencing and folding.
Omniverse now features physically-based rendering support along with AI-assisted design plugins to automate asset generation. Early adopters demonstrated using Omniverse to create interactive digital twins of retail stores, factories, and cities to optimize operations.
Meanwhile, BioNeMo’s new synthetic data generation capabilities allow researchers to create massive simulated protein folding datasets. By training models on this data, scientists can discover new drug candidates faster.
“We’re committed to putting simplified, extremely powerful AI tools into the hands of all developers and engineers.”
Jensen Huang, Founder and CEO, NVIDIA
Pioneering Responsible AI for Social Good
GTC sessions went beyond product announcements to explore bleeding-edge AI research and development focused on social responsibility.
Scientists from NVIDIA and partner institutions delved into work on trustworthy AI, algorithmic fairness, differential privacy, and tools to detect AI-generated synthetic media.
“As AI grows more powerful, ensuring it is safe, ethical, and benefits humanity is our highest priority,” Huang emphasized.
“We are investing tremendous resources toward this goal and calling on the brightest minds to bring thoughtful leadership to AI development.”
Researchers also presented real-world case studies on using AI in healthcare, accessibility tech, sustainability initiatives, and humanitarian relief.
Attendees responded strongly to sessions focused on developing AI to expand opportunity, protect civil rights, and democratize access to services.
Upgrade Now to Unlock AI’s Potential
NVIDIA GTC 2024 made clear that AI solutions are rapidly moving from research to reality.
With the Hopper Super Chip powering the next generation of transformative AI applications, business leaders need to act now to upgrade infrastructure and build internal skills.
By leveraging GTC’s latest innovations, companies can gain a significant competitive edge.
Visit nvidia.com/gtc for full conference details, session recordings, and AI adoption resources tailored to your unique needs.
The future of AI is here—tap into its nearly limitless potential today.
