Nvidia’s annual GTC event returned to an in-person format this year, attracting a large crowd due to its expanded presence in various markets and its leadership in AI computing for the cloud. The company announced the Blackwell family of products, featuring GPUs with impressive specs like 8 TB/s of memory bandwidth and peak FP4 Dense Tensor performance of up to 9 PFLOPS. These GPUs are designed using TSMC’s 4NP process node and offer significant performance improvements over previous models.
In addition to the Blackwell GPUs, Nvidia also introduced NIM microservices for generative AI, aimed at optimizing AI models for Nvidia hardware and improving time-to-market and quality of results. The company announced a partnership with MediaTek to develop the Dimensity Auto Cockpit lineup, combining Arm v9-A CPUs with Nvidia GPUs to enable AI and RTX graphics in vehicles. These chipsets also support QNX, Linux, and Android Automotive OS, making them versatile options for OEMs.
Nvidia’s Omniverse platform was highlighted at GTC, with announcements regarding partnerships with Siemens and Cadence to enhance data center planning and simulation capabilities. The Earth-2 climate digital twin cloud platform was introduced for simulating weather and climate, leveraging generative AI models and Omniverse APIs. Nvidia also unveiled the 6G Research Cloud, aimed at using AI in the RAN to plan and build 6G solutions, with industry leaders like Ansys and Nokia already onboard.
Overall, GTC 2024 was a significant event for Nvidia, reaffirming its dominance in the AI space with a wide range of announcements across GPU, automotive, and Omniverse segments. The company showcased its commitment to advancing AI technology and supporting its partners through innovative solutions like the Blackwell GPUs, NIM microservices, and collaborations with industry leaders. Despite supply constraints, Nvidia’s leadership in AI computing is evident, positioning the company as a key player in the evolving landscape of AI-powered technologies.
