Site Logo

Nvidia GTC 2024: Blackwell GPU, Robotics & The Future of AI

Marc-Antoine LebrunEditor in chief
Updated at: 3/16/2026 11:06:22 PM

GTC 2024: Nvidia Unveils the Future of AI with Blackwell, Robotics, and More

Nvidia's annual GPU Technology Conference (GTC) has become a landmark event in the tech world, and GTC 2024 was no exception. In a highly anticipated keynote, CEO Jensen Huang took to the stage to unveil a new wave of technologies poised to redefine the landscape of artificial intelligence, high-performance computing, and robotics. The announcements sent ripples through the industry and further fueled the momentum of Nvidia's stock, which has been on a meteoric rise. This article delves into the key reveals from GTC 2024 and explores their far-reaching implications.

The Dawn of a New GPU Era: The Blackwell Platform

The centerpiece of the GTC 2024 keynote was undoubtedly the announcement of the Blackwell architecture, the successor to the incredibly successful Hopper. The new flagship GPU, the B200, is an engineering marvel, boasting specifications that dwarf its predecessors and promise to unlock new frontiers in AI development.

Blackwell B200: A Leap in Performance

The Blackwell B200 is not just an incremental upgrade; it's a monumental leap forward. Here are some of its key features:

  • 208 Billion Transistors : The B200 is built with a custom TSMC 4NP process and features two dies connected by a 10 TB/s chip-to-chip link, effectively functioning as a single, unified GPU.
  • Up to 20 Petaflops of FP4 Compute : This immense processing power is designed to handle the most demanding AI workloads, including the training of trillion-parameter models.
  • Second-Generation Transformer Engine : This new engine doubles the compute, bandwidth, and model size capabilities, significantly accelerating inference for large language models (LLMs).
  • Fifth-Generation NVLink : The new NVLink switch allows for seamless, high-bandwidth communication between up to 576 GPUs, enabling the creation of massive, supercomputer-scale AI systems.

To put the B200's power into perspective, here's a comparison with its predecessor, the H100:

FeatureHopper H100Blackwell B200
:---:---:---
Transistors80 Billion208 Billion
AI Performance4 PetaflopsUp to 20 Petaflops
Interconnect900 GB/s1.8 TB/s
Memory Bandwidth3.35 TB/s8 TB/s
Key Takeaway

The Blackwell B200 is designed to make training and running generative AI models, especially those with trillions of parameters, more efficient and cost-effective. Nvidia claims up to a 25x reduction in cost and energy consumption for LLM inference compared to the H100.

The Next Wave of Robotics: Project GR00T

Nvidia is extending its AI dominance into the physical world with a major new robotics initiative. Huang introduced Project GR00T , a general-purpose foundation model for humanoid robots.

Enabling Intelligent Robots

The goal of Project GR00T is to create a model that allows humanoid robots to understand natural language and learn skills by observing human actions. This "embodied AI" could revolutionize industries from manufacturing to healthcare. Key components of this initiative include:

  • Jetson Thor : A new computer specifically designed for humanoid robots, based on the Thor system-on-a-chip (SoC).
  • Isaac Robotics Platform : A comprehensive set of tools for simulating, training, and deploying AI-powered robots.

Nvidia is collaborating with leading robotics companies to develop this technology, aiming to create a future where robots can safely and effectively perform a wide range of tasks in human environments.

Omniverse and the Rise of Digital Twins

Nvidia's Omniverse platform, a real-time 3D collaboration and simulation environment, also received significant updates. The focus is on creating "digital twins"—virtual replicas of real-world objects and systems. These digital twins can be used for everything from designing and testing new products to optimizing factory layouts and supply chains.

New Omniverse Cloud APIs will make it easier for developers to integrate Omniverse capabilities into their applications. In a notable collaboration, Nvidia is working with Apple to bring Omniverse to the Vision Pro, enabling immersive, high-fidelity simulations in mixed reality.

A Word of Caution

While the advancements announced at GTC 2024 are groundbreaking, they also raise important considerations. The immense power of the new Blackwell systems comes with significant energy demands, and the concentration of AI power in the hands of a few large companies could have long-term societal implications.

The Impact on Nvidia and the AI Industry

The announcements at GTC 2024 solidify Nvidia's position as the undisputed leader in the AI hardware market. The company is not just selling chips; it's providing an entire ecosystem of hardware, software, and services that are becoming the backbone of the AI revolution.

The new Blackwell platform is expected to be adopted by all major cloud providers and AI companies, including Amazon Web Services, Google Cloud, Microsoft Azure, and OpenAI. This will likely translate into continued strong revenue growth for Nvidia and further gains for its stock (NVDA).

For the broader AI industry, these new tools will accelerate the pace of innovation, making it possible to build larger, more capable AI models and deploy them in a wider range of applications. The future of AI, as envisioned by Nvidia, is one of massive, interconnected systems that will transform every aspect of our lives.

FAQ

On the same topic

Marc-Antoine Lebrun
Editor in chief
Passionate about finance and new technologies for many years, I love exploring and delving deeper into these fascinating fields to better understand them. Curious and always eager to learn, I’m particularly interested in cryptocurrencies, blockchain, and artificial intelligence. My goal: to understand and share the innovations that are shaping our future.