April 5 (UPI) — Google has released information on its new AI supercomputer, saying it is more powerful than Nvidia, its main competitor.
The Google supercomputer is based on Tensor Processing Units, which work together to develop the AI system.
Google has announced that a new supercomputer it is developing for AI systems is faster and more efficient than its main competitor, Nvidia. File Photo by John Angelillo/UPI | License Photo
According to Google researchers, the TPU v4 supercomputer, is “1.2x-1.7x faster and uses 1.3x-1.9x less powerful than the Nvidia A100.”
While Google creates, and conducts research on, AI systems, Nvidia dominates about 90% of the market.
At a recent company meeting, Google employees raised concerns about falling behind competitors in the AI market, despite the fact that Google has been working on AI systems for years.
AI systems such as ChatGPT require thousands of chips to complete the machine learning required to prep the systems over extended periods of time.
The sheer volume of computers needed to operate AI systems is a major challenge for the industry, expending a great deal of power.
Google hopes their new product will be far more efficient, saying their system has more than 4,000 TPUs to train AI systems.
Google currently is testing an AI system called Google Bard, known alternately as Apprentice Bard.