Intel Introduces Gaudi 3 AI Chip, Teams Up with Indian Firms
Intel has unveiled its latest artificial intelligence chip, Gaudi 3, at the Vision event, claiming it to be a game-changer in AI training and inference. The chip is reported to train large language models at a rate 50% faster than Nvidia’s H100 processor, representing a significant advancement in AI processing.
This announcement comes as Intel aims to challenge Nvidia’s dominance in the AI semiconductor market. The Gaudi 3 chip is positioned to outperform Nvidia’s offerings, particularly in tasks related to training large language models and generating AI responses.
Intel showcased the Gaudi 3 at the Intel Vision 2024 customer and partner conference, emphasizing its commitment to accelerating the adoption of generative AI. Pat Gelsinger, Intel’s CEO, highlighted the widespread integration of AI across various industries and expressed Intel’s dedication to making AI accessible everywhere in the enterprise.
The Gaudi 3 chip is slated to be available to original equipment manufacturers (OEMs) like Dell Technologies, HPE, Lenovo, and Supermicro in the second quarter of 2024, promising enhanced AI capabilities in their products and solutions.
In addition to unveiling the Gaudi 3 chip, Intel announced strategic collaborations with prominent Indian companies to deploy its AI accelerator solutions:
Bharti Airtel: Airtel aims to leverage its telecom data to bolster its AI capabilities, enhancing customer experiences and exploring new revenue streams in the highly competitive telecom market.
Infosys: Intel is collaborating with Infosys to integrate its advanced technologies, including 4th and 5th Gen Intel Xeon processors and Intel Gaudi 2 AI accelerators, into Infosys Topaz – an AI-centric suite of services, solutions, and platforms designed to accelerate business value using generative AI technologies.
Ola/Krutrim: Intel’s new technology will enable Ola/Krutrim to pre-train and fine-tune its foundational AI model with generative capabilities in multiple languages, catering to the diverse linguistic landscape of India.
Furthermore, Intel announced partnerships with Google Cloud, Thales, and Cohesity to leverage its confidential computing capabilities, including Intel Trust Domain Extensions (Intel TDX) and Intel Software Guard Extensions (Intel SGX), for enhanced cloud security and data protection.
Despite these advancements, Intel and Advanced Micro Devices (AMD) have faced challenges in developing a competitive suite of chips and software for AI applications compared to Nvidia, which currently dominates the data center chip market. However, with the introduction of the Gaudi 3 chip, Intel aims to strengthen its position in the AI semiconductor industry and offer compelling solutions to its customers.
Built using Taiwan Semiconductor Manufacturing Co’s 5nm process, the Gaudi 3 chip is expected to be available to server builders in the second quarter of the year, promising improved performance and efficiency for AI-driven applications and services.
