Nvidia GTC 2024 key announcements: From AI Chip to Humanoid Robot

Nvidia's CEO, Jensen Huang, kicked off the company's annual GPU Technology Conference (GTC) on Monday by unveiling a range of exciting products and advancements, from the Blackwell platform to the foundational model for humanoid robots. With a focus on generative AI, GTC 2024 highlighted not only innovative technologies but also the partnerships with industry leaders, demonstrating the platform's remarkable capabilities. 

Let's take a look at some of the key announcements from the event: 

Blackwell 

Nvidia has introduced the Blackwell platform, the latest advancement in its AI technology. Dubbed the "world's most powerful chip," the Blackwell GPU is poised to usher in a new era of computing, according to the company. 

Compared to its predecessor, Hopper, the Blackwell system boasts enhanced capabilities and efficiency. Nvidia highlights that the Blackwell architecture can operate a large language model for generative AI while consuming approximately 25 times less cost and energy than Hopper. 

In addition, Nvidia unveiled the GB200 Grace Blackwell Superchip, a significant innovation that links two NVIDIA B200 Tensor Core GPUs to the NVIDIA Grace CPU. This super chip, known as GB200, will be available on NVIDIA DGX Cloud, an AI platform developed in collaboration with leading cloud service providers. It offers enterprise developers dedicated access to the necessary infrastructure and software for building and deploying advanced generative AI models. 

AI Humanoid Robot 

Nvidia revealed "Project GR00T," a foundation model tailored for humanoid robots. CEO Huang highlighted it as the forefront of AI and robotics advancement. 

This initiative aims to equip robots with GR00T, short for Generalist Robot 00 Technology, enabling them to comprehend natural language and mimic human movements through observation. Additionally, Nvidia introduced "Jetson Thor," a cutting-edge computer designed specifically for humanoid robots. It integrates GenAI capabilities as part of Nvidia's AI-powered robotics agenda. 

NIM concept  

Nvidia unveiled an intriguing concept called Inference Microservices (NIM). NIM simplifies the deployment of generative AI applications, ensuring security, stability, and scalability. It's a vital component of Nvidia AI Enterprise, offering pre-packaged AI models, integration code, and a ready-to-use Kubernetes Helm chart for deployment. 

With NIM, developers can significantly reduce the time-to-market for AI deployments, launching models in just 10 minutes. 

6G Cloud Platform AI 

The company also introduced a 6G research platform designed to boost AI advancements in radio access network (RAN) technology. This platform enables organizations to fast-track the development of 6G technologies, facilitating the connection of trillions of devices to cloud infrastructures. It sets the stage for a hyper-intelligent world, fostering autonomous vehicles, smart spaces, extended reality, immersive education experiences, and collaborative robots. 

Nvidia – Google Partnership 

Google Cloud and Nvidia have strengthened their partnership to empower the machine learning (ML) community by providing technology that speeds up the creation, expansion, and management of generative AI applications. 

In pursuit of advancing AI capabilities in its products and for developers, Google is embracing the latest Nvidia Grace Blackwell AI computing platform. Furthermore, Google Cloud will offer the Nvidia DGX Cloud service, alongside making the Nvidia H100-powered DGX Cloud platform readily accessible. 

Snowflake Partnership 

Snowflake, partnered with Nvidia, to boost enterprise capabilities with an AI platform. This collaboration integrates Nvidia’s accelerated platform with Snowflake's trusted data foundation and secure AI, offering customers a robust infrastructure for AI-driven transformations. Together, both the companies will provide a secure and potent combination of infrastructure and compute power, aimed at boosting AI productivity and driving business innovation across all sectors. 

CrowdStrike  

CrowdStrike collaborated with Nvidia to integrate its AI computing services into CrowdStrike Falcon XDR platform. This partnership combines Falcon platform data with Nvidia’s GPU-optimized AI pipelines and software, including new Nvidia NIM microservices, empowering both CrowdStrike and Nvidia customers to create custom and secure generative AI models.