Nvidia Reveals Next-Gen Rubin AI Platform

Nvidia Corporation, led by CEO Jensen Huang, has unveiled the Blackwell Ultra chip slated for 2025 and introduced the development of the next-generation Rubin platform scheduled for 2026, the firm announced ahead of this week’s COMPUTEX technology conference in Taipei.^[1]  

Nvidia aims to diversify its customer base beyond major cloud-computing entities.

To help software developers, Nvidia offers tools and pre-trained AI models to facilitate the integration of AI functionalities into PC applications. Additionally, Nvidia's MGX program for server design enables expedited product launches by partners like Hewlett Packard Enterprise Co. and Dell Technologies Inc.

“The next wave of AI is physical AI,” Huang said ahead of COMPUTEX. “AI that understands the laws of physics, AI that can work among us.”^[2]  

“Robotics is here. Physical AI is here. This is not science fiction, and it’s being used all over Taiwan. It’s just really, really exciting,” Huang said.

According to Nvidia, physical AI refers to "models that can understand instructions and autonomously perform complex tasks in the real world." Nvidia's Isaac AI robot development platform is already being utilized by Siemens and BYD Electronics as well as other factories and warehouses to increase productivity. Isaac is being utilized for the development, research, and manufacturing of "tens of millions" of AI-powered autonomous devices and robots, including industrial arms and humanoids.

Furthermore, Nvidia's previously announced products, such as Spectrum X for networking and Nvidia Inference Microservices (NIM), are now widely available, with NIM offerings accessible for free.

There are presently more than five million preprogrammed robots being tested and approved globally for use in factories, including through digital twin simulations in Nvidia's Omniverse, according to Nvidia.

“All the factories will be robotic,” Huang said. “The factories will orchestrate robots, and those robots will be building products that are robotic.”

Previously, CEO Huang emphasized the importance of countries owning the production of their own intelligence. In a fireside chat at the World Governments Summit in Dubai with the UAE’s Minister of AI Omar Al Olama, Huang described sovereign AI as a vast opportunity for global leaders.

In April, Meta Platforms Inc. announced the deployment of a new chip designed to bolster its artificial intelligence (AI) capabilities, a move to lessen dependence on semiconductor suppliers like Nvidia Corp. and other external entities.

The unveiling of the chip represents the latest iteration of the Meta Training and Inference Accelerator (MTIA). This technology plays a role in powering algorithms responsible for content ranking and recommendation algorithms across Meta's platforms, including Facebook and Instagram.

COMPUTEX 2024 will open tomorrow, June 4th, at the Tiwani Taipei Nangang Exhibition Center, Halls 1 and 2.