Chipmaker Nvidia (NVDA) has lifted the curtains on its latest creation: the Nvidia Rubin CPX. This is a new group of graphics processing units (GPUs) specifically designed to handle “million-token” software coding and generative video applications.
Elevate Your Investing Strategy:
- Take advantage of TipRanks Premium at 50% off! Unlock powerful investing tools, advanced data, and expert analyst insights to help you invest with confidence.
Tokens are the basic units of data that AI models use to understand and generate content. Traditional AI models typically handle tens of thousands of tokens at a time. However, Nvidia said its new GPU will be able to handle much more massive loads “with groundbreaking speed and efficiency.”
Nvidia revealed the new product on Tuesday at the AI Infrastructure Summit in California. However, the new GPU will not be ready for sale until the end of 2026, the chipmaker said in a statement.
Nvidia Creates GPU for Massive AI Capacity
Providing details on the new product, Nvidia noted that the new GPU was designed to work with its Vera Rubin NVL 144 CPX, its new AI computing system. The system is said to combine the power of Nvidia’s Vera central processing unit (CPU) and the Rubin GPU. It also comes with 144 GPUs that enable it to deliver 7.5 times more AI performance compared to the company’s older systems.
“The Vera Rubin platform will mark another leap in the frontier of AI computing — introducing both the next-generation Rubin GPU and a new category of processors called CPX,” said Jensen Huang, CEO of Nvidia.
Nvidia Eyes ‘New Class of AI Factories’
Meanwhile, Nvidia at the Summit also disclosed that it had started working on reference designs to transform traditional data centers into fully integrated AI factories. The goal is to create infrastructure enhanced for “the age of AI reasoning.”
Nvidia further noted that it is already in partnership with several companies to achieve this goal. One such key partner is engineering consulting firm Jacobs (J), which is helping the chipmaker to coordinate the integration of the physical and digital layers of the new system.
“It’s a natural evolution for the company, scaling beyond chips and systems into a new class of industrial products – so complex and interconnected that no single player can build them alone,” noted Ian Buck, Nvidia’s VP of Accelerated Computing.
Is NVDA a Strong Buy?
On TipRanks, Nvidia’s shares have a Strong Buy consensus recommendation based on assessments by 39 Wall Street analysts over the past three months. The recommendation is based on 35 Buy, three Hold, and one Sell ratings.
The average NVDA price target is $210.73, which stands for a 25% potential upswing.
