At the Vision event, Intel unveiled the Gaudi 3, a new artificial intelligence chip designed to challenge Nvidia’s market dominance in this area. Compared to Nvidia’s H100 processor, the new chip can compute generative AI responses 50% faster and train large language models 50% faster.
Tech companies are rushing to find alternate sources of the rare chips needed for artificial intelligence (AI), which is why this announcement is made. Part of Intel’s portfolio of next-generation products, open scalable systems, and strategic partnerships, the Gaudi 3 was unveiled at the Intel Vision 2024 customer and partner conference with the goal of propelling the use of generative AI as quickly as possible.
According to Intel CEO Pat Gelsinger, every business is rapidly turning into an AI company, and Intel is implementing AI throughout the entire enterprise. It is anticipated that OEMs such as Dell Technologies, HPE, Lenovo, and Supermicro will have access to the Gaudi 3 by the second quarter of 2024.
Additionally, Intel announced partnerships with Ola/Krutrim, Infosys, and Bharti Airtel to implement the Gaudi accelerator solutions.
Bharti Airtel: In order to improve its AI capabilities and speed up customer experiences, Airtel intends to use telephone data. For the second-largest telecom company in the nation, the deployments will help open up new revenue streams.
Infosys: The company announced a strategic partnership to support Infosys Topaz, an AI-first suite of services, solutions, and platforms that leverage generative AI to accelerate business value. This partnership will bring Intel technologies, such as 4th and 5th Gen Intel Xeon processors, Intel Gaudi 2 AI accelerators, and Intel Core Ultra, to Infosys.
Ola/Krutrim: First India foundational model with generative capabilities in 10 languages will be pre-trained and refined by the Indian AI model using Intel’s new technology. Using an Intel Gaudi 2 cluster, Krutrim is currently pre-training a larger foundational model.
Additionally, Intel disclosed partnerships with Thales, Cohesity, and Google Cloud to take advantage of Intel’s private computing power in their cloud instances. This covers Intel’s attestation service, Intel Software Guard Extensions (Intel SGX), and Intel Trust Domain Extensions (Intel TDX). Clients may use trusted execution environments (TEEs) to test their AI models and algorithms, and they can use Intel’s trust services to independently confirm these TEEs’ reliability.
Nvidia, which held about 83% of the market for data center chips in 2023, has a competitive advantage over Intel and Advanced Micro Devices in terms of producing a compelling combination of chips and the software needed to develop AI applications. In the second quarter of this year, server manufacturers like Supermicro and Hewlett Packard Enterprise will be able to purchase the Gaudi 3 chip, which is more than twice as fast as its predecessor and was created using Taiwan Semiconductor Manufacturing Co.’s 5nm process.