Intel launches Gaudi 3 AI chip in India and partners with Infosys, Airtel, Ola
During the Vision event, Intel unveiled its latest artificial intelligence chip, the Gaudi 3, aiming to disrupt Nvidia's stronghold in AI semiconductor technology. This new chip boasts the capability to train extensive language models with a 50% speed advantage over Nvidia's H100 processor and excels in rapidly computing generative AI responses.
Against the backdrop of a chip shortage driving tech firms to seek alternatives, Intel's unveiling of the Gaudi 3 at the Intel Vision 2024 conference signals a strategic move to address the demand for AI chips. This introduction aligns with Intel's broader initiative, featuring new scalable systems, next-gen products, and partnerships, all geared towards expediting the integration of generative AI technologies.
Intel's CEO, Pat Gelsinger, emphasized the ubiquity of AI across industries, affirming Intel's commitment to democratizing AI deployment. Scheduled for release to OEMs like Dell Technologies, HPE, Lenovo, and Supermicro in Q2 2024, the Gaudi 3 chip is poised to extend Intel's AI footprint. Additionally, Intel revealed partnerships with Bharti Airtel, Infosys, and Ola/Krutrim for the deployment of Gaudi accelerator solutions.
Airtel aims to utilize its telecom data to bolster its AI capabilities, accelerating customer experiences and opening avenues for new revenue streams, positioning itself as the second-largest telecom company in the country. Infosys has revealed a key partnership aimed at integrating Intel technologies such as 4th and 5th Gen Intel Xeon processors, Intel Gaudi 2 AI accelerators, and Intel Core Ultra into Infosys Topaz. This collaboration seeks to enhance the AI-driven services, solutions, and platforms offered by Infosys, thereby accelerating business value through generative AI technologies.
Ola's Krutrim initiative is set to employ Intel's latest technology for pre-training and refining its inaugural Indian foundational model equipped with generative capabilities across 10 languages. Krutrim is currently in the process of pre-training an expanded foundational model using an Intel Gaudi 2 cluster.
Intel has revealed partnerships with Google Cloud, Thales, and Cohesity to utilize Intel's confidential computing capabilities within their cloud instances, incorporating features such as Intel Trust Domain Extensions (Intel TDX), Intel Software Guard Extensions (Intel SGX), and Intel’s attestation service. This collaboration enables customers to execute their AI models and algorithms within a trusted execution environment (TEE), supported by Intel's trust services for independent verification of TEE trustworthiness.
Despite their progress, Intel and Advanced Micro Devices have faced challenges in developing a competitive combination of chips and accompanying software for creating AI applications capable of matching Nvidia's dominance, which held approximately 83 percent of the data center chip market in 2023. The Gaudi 3 chip, boasting more than double the speed of its predecessor, was fabricated using Taiwan Semiconductor Manufacturing Co's 5nm process and is scheduled for release to server manufacturers like Supermicro and Hewlett Packard Enterprise in the second quarter of this year.