Top five things we learned from day one of Nvidia’s GTC conference

Yesterday was a day to remember in San Jose and around the world, as the Nvidia GTC 2024 conference kicked off.

The highlight of the day was undoubtedly the spectacular, two hour-long keynote address by Nvidia founder and CEO Jensen Huang.

Those who came for computing chips got so much more, like revolutionary insights into how GPUs are changing everything from robotics to manual labour, healthcare and weather systems.

Here, the most impressive takeaways we noticed from day one of #GTC24:

1. The new AI superchip is here
2. So is the new AI supercomputer
3. Introducing NIM
4. Robotics is about to get revolutionary  
5. Nvidia is creating a whole digital planet Earth: here’s how it wants to use it

1. The new AI superchip is here

Easily the biggest announcement yesterday was Jensen Huang’s introduction of the new Blackwell chip – the latest, more powerful GPU successor to Nvidia’s Hopper chip.

It’s twice as fast as Hopper and the Blackwell platform can scale up to 576 GPUs. It also manages to do this in a smaller space than previous computing systems.

This is what Nvidia had to say on the launch:

2. So is the new AI supercomputer

In order to handle all of this AI heavy lifting, Huang also announced a newer, more powerful Nvidia AI supercomputer: the new DGX SuperPOD.

According to Nvidia, it’s designed to run on the company’s powerful Grace Blackwell superchips, so as to be able to process the enormous amount of data required to turn a computing system into an ‘AI learning factory’.

Specifically, that means “11.5 exaflops of AI supercomputing at FP4 precision and 240 terabytes of fast memory” for the “processing trillion-parameter models with constant uptime for superscale generative AI training and inference workloads” that serious AI work requires, according to Nvidia's blog.

3. Introducing NIM

To compliment these new releases, Nvidia also launched a new software service called Nvidia Inference Microservice (NIM).

According to Huang, NIMs will revolutionise the way software is built and coded in future, assembling Ais to essentially write and create it for you without having to write a single line of code.

NIMs will be cloud-based, meaning they can be downloaded from an Nvidia hub online, to work with anywhere you are in the world.

4. Robotics is about to get revolutionary  

Another announcement for the day was ‘Project GROOT’: a new foray into robotics by Nvidia, which will enable various types of robots to learn from and mimic human movements and speech with a powerful sensory robotics processing chip.

Several mind-bending applications were introduced for these robotics breakthroughs, including self-driving cars, robot factory workers, talking food kiosks and more.

5. Nvidia is creating a whole digital planet Earth: here’s how it wants to use it

With its newly improved ability to collate and process massive amounts of data, Nvidia is able to now make identical ‘digital twins’ online that exactly mirror their real-life counterparts.

And, in a scene that seemed straight out of The Hitchhiker’s Guide to the Galaxy, Huang announced that one of those twins will be its next new big project: ‘Earth-2’, a full remodel of our blue and green planet.

E-2 will, hopefully, be used to more precisely determine and predict weather patterns to lessen the devastating effects of natural disasters.