Intel reveals details of new AI chip to take on Nvidia

  • Intel detailed a new version of its artificial intelligence chip at its Vision event on Tuesday.
  • Intel said that its new Gaudi 3 chip was capable of training specific large language models 50% more quickly than Nvidia’s prior generation H100 processor.
  • Gaudi3 aims at Nvidia and AMD, opening up a new dominance in the AI semiconductor space.

Intel released the latest generation of AI chips, Gaudi3 at Intel Vision 2024 on Tuesday.

The new Gaudi 3 chip

Intel says its new Gaudi 3 chip can train specific large-scale language models 50% faster than NVIDIA’s previous-generation H100 processor. It is also capable of computing generative AI responses, a process called inference, more quickly than the H100 chips for some of the models Intel tested.

Gaudi 3 is expected to ship to the original equipment manufacturer (OEM) in the second quarter of 2024.

Also read: Intel discloses US$7 billion operating loss for chipmaking unit

Intel versus Nvidia and AMD

The competition on the AI circuit continues.

Intel’s old rival AMD launched its MI300 series of products last year, with the MI300X chip boasting more than 150 billion transistors, 2.4 times the memory density of NVIDIA’s H100, and 1.6 times the memory bandwidth.

According to TechInsights, AMD is implementing AI acceleration at the system level has an edge in inference workloads, and is expected to gain market share from Nvidia.

Competition will intensify in 2024 when Nvidia will launch its next-generation products, which will surely overtake AMD’s offerings.

In March of this year, Nvidia refreshed its product line by releasing its latest GPU architecture platform, Blackwell, a new generation of AI product, B200, which has strong performance and contains 208 billion transistors.

“Our customers, first and foremost, are asking for choice in the industry,” said Intel vice president, of strategy and product management Jeni Barovian.    

Intel and AMD have struggled to produce a compelling bundle of chips and the software necessary to build AI applications that can become a viable alternative to Nvidia.

Jennifer-Yu

Jennifer Yu

Jennifer Yu is an intern reporter at BTW Media covering artificial intelligence and products. She graduated from The University of Hong Kong. Send tips to j.yu@btw.media.

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *