NVIDIA CEO Discusses New GPU Architecture at Computex 2024
At Taiwan’s Computex 2024 event, NVIDIA CEO Jensen Huang used part of his keynote speech to announce the arrival of a new GPU architecture, Rubin.
As the Rubin is designed with an awareness of the limits placed on data centre size and power use restrictions, Huang hopes that this will help data centres become more efficient.
NVIDIA assist data centre technology expansion
NVIDIA is a computer hardware manufacturer, which aims to explore edge solutions for businesses, with its offerings transforming possibility into real-world results. In regards to data centres, NVIDIA assists data centres in many ways, from leveraging GPU technology, to designing software platforms and supporting ecosystem partnerships. However, power consumption remains one of the biggest costs in running a data centre and NVIDIA is focused on how to manage this sustainably.
The Rubin GPU will include eight stacks of HBM4 memory, plus a roadmap to Rubin Ultra, which will feature 12 HBM4 stacks.
Furthermore, NVIDIA aims to introduce a successor to its Grace Blackwell chips. Huang said that the Blackwell Ultra will be next, which will include 12 HBM3E compatibility. The Rubin architecture, which will arrive in 2026, will be followed by an Ultra version. This is keeping on track with NVIDIA’s ‘one year rhythm’ for data centre releases.
"Our company has a one-year rhythm. Our basic philosophy is very simple: build the entire data centre scale, disaggregate and sell to you parts on a one-year rhythm, and we push everything to technology limits," said Huang.
In another move, which underscores the growing convergence of AI and cybersecurity, NVIDIA and Trend Micro have also announced a partnership to develop new AI-driven security tools. Designed to safeguard data centres where AI workloads are processed, the collaboration aims to leverage NVIDIA’s industry-leading AI hardware and software capabilities to bolster Trend Micro's cybersecurity offerings. Trend Micro’s offering, for business and government customers, underscores the growing value organisations are placing on their AI capabilities and the cybersecurity company’s ambition to secure that.
NVIDIA’s expanding AI developments
NVIDIA has surpassed all expectations within the data centre industry, hitting revenues of US$22.10bn, with its data centre revenue increasing by 409% to US$18.4bn.
Analysts expect that the company’s data centre unit will record US$17.06bn in sales for its fiscal fourth quarter at the end of 2024. This is due in part to NVIDIA’s AI developments rapidly increasing profits, with the company expected to reach US$2tn in value over the next five years.
“Our data centre platform is powered by increasingly diverse drivers - demand for data processing, training and inference from large cloud-service providers and GPU-specialised ones, as well as from enterprise software and consumer internet companies,” Huang said. “The year ahead will bring major new product cycles with exceptional innovations to help propel our industry forward.”
NVIDIA utilises an automating intelligence at the point of action to drive decisions in real-time, with its EGX Edge Computing Platform enabling organisations to deploy, manage and scale edge computing solutions across their distributed infrastructure.
NVIDIA’s AI-on-5G is a unified platform that simplifies the deployment of AI applications over 5G edge networks. The platform aims to bring together expansive 5G connectivity, powerful computers and AI to accelerate the digital transformations.
As a result, we featured NVIDIA in our Top 10: Edge Providers in 2024.
******
Make sure you check out the latest edition of Data Centre Magazine and also sign up to our global conference series - Tech & AI LIVE 2024
******
Data Centre Magazine is a BizClik brand