Google said Tuesday that one of the few alternatives to Nvidia’s popular AI chips is now available to developers — and took a shot at Microsoft and Amazon, too.
The tech giant’s latest AI chip, the Cloud TPU v5p, was first announced in December, the same day as its chatbot Gemini. The new TPU, or tensor processing unit, can train large language models almost three times faster than its predecessor, Google’s TPU v4, the company said. Large language models (LLMs) power AI chatbots like ChatGPT.
“Now in their fifth generation, these advancements [to Google’s TPUs] have helped customers train and serve cutting edge language models,” Google CEO Sundar Pichai said at the company’s annual Google Cloud Next conference on Tuesday in Las Vegas.
Google’s announcement was another milestone in Big Tech’s AI arms race. Nvidia is the main supplier of the AI chips known as GPUs, or graphics processing units. And Google parent Alphabet is one of Nvidia’s biggest customers, behind Microsoft and Facebook parent Meta.
“[Google’s] investments [to develop new AI hardware] put us at the forefront of the AI platform shift,” Pichai said.
Google’s rivals Microsoft, Amazon, and Meta have also developed their own AI chips.
Still, it’s clear that Nvidia is very important to Google. In the same blog post announcing its latest AI chip, Google mentioned Nvidia 20 times. Right under its detailing of the TPU v5p, the company said it’s updating its A3 supercomputer, which runs on Nvidia GPUs. And Google reminded customers that it’s using Nvidia’s latest chip, the Blackwell, in its AI Hypercomputer.
Read more: Intel says its new AI hardware is ‘way ahead of Nvidia’
After discussing Google’s new AI chip, Google Cloud CEO Thomas Kurian gave a flashier presentation of its new, Arm-based central processing unit called Google Axion. In a game show-esque reveal, Kurian walked the stage during his keynote address with the Axion in hand, flashing the chip to applause.
Google Axion is a new rival to the CPUs of Microsoft and Amazon, which have already made their own Arm-based computing chips. The British tech company Arm licenses its chip infrastructure design for actual chip manufacturers to build upon. Google’s release of Axion marks the first time the company has used Arm’s chip infrastructure for a CPU.
Google said Axion provides “up to 30% better performance than the fastest general-purpose Arm-based instances available in the cloud today” and “up to 50% better performance and up to 60% better energy-efficiency” than other general purpose Arm chips.
Google customers can use Axion on its cloud services, which basically means those users would be opting to run their cloud services on a more efficient computer processor in Google’s physical data centers. Google also told Reuters that “customers using Arm anywhere can easily adopt Axion without re-architecting or re-writing their apps.”