Intel unveils new AI chip Gaudi 3 to compete with Nvidia, partners with Infosys, Bharti Airtel, Ola in India

Intel has introduced a new version of its artificial intelligence chip, Gaudi 3, at its Vision event. The chip is said to train large language models 50% faster than Nvidia’s H100 processor, marking a significant leap in AI training and inference.

Intel has revealed details of its new artificial intelligence chip, the Gaudi 3, at the Vision event, in an attempt to challenge Nvidia’s dominance in AI semiconductors. The new chip is capable of training large language models 50% faster than Nvidia’s H100 processor and can compute generative AI responses more quickly.

This announcement comes amid a scramble by tech companies to find alternative sources of the scarce chips required for AI. The Gaudi 3, introduced at the Intel Vision 2024 customer and partner conference, is part of Intel’s suite of new open scalable systems, next-gen products and strategic collaborations aimed at accelerating the adoption of generative AI.

Intel CEO Pat Gelsinger said that every company is quickly becoming an AI company, and Intel is bringing AI everywhere across the enterprise. The Gaudi 3 is expected to be available to OEMs, including Dell Technologies, HPE, Lenovo, and Supermicro, in the second quarter of 2024.

Intel also announced collaborations with Bharti Airtel, Infosys, Ola/Krutrim to deploy the Gaudi accelerator solutions.

Bharti Airtel: Airtel plans to leverage its telecom data to enhance its AI capabilities and expedite the experiences of its customers. The deployments will help drive new revenue streams for the second biggest telecom company in the country.

Infosys: The company announced a strategic collaboration to bring Intel technologies including 4th and 5th Gen Intel Xeon processors, Intel Gaudi 2 AI accelerators and Intel Core Ultra to Infosys Topaz – an AI-first set of services, solutions and platforms that accelerate business value using generative AI technologies.

Ola/Krutrim: The Indian AI model will use Intel’s new technology to pre-train and fine-tune its first India foundational model with generative capabilities in 10 languages. Krutrim is now pre-training a larger foundational model on an Intel Gaudi 2 cluster.

Intel also announced collaborations with Google Cloud, Thales and Cohesity to leverage Intel’s confidential computing capabilities in their cloud instances. This includes Intel Trust Domain Extensions (Intel TDX), Intel Software Guard Extensions (Intel SGX) and Intel’s attestation service. Customers can run their AI models and algorithms in a trusted execution environment (TEE) and leverage Intel’s trust services for independently verifying the trust worthiness of these TEEs.

Despite the advancements, Intel and Advanced Micro Devices have struggled to produce a compelling bundle of chips and the necessary software to build AI applications that can rival Nvidia, which controlled roughly 83 per cent of the data center chip market in 2023. The Gaudi 3 chip, more than twice as fast as its predecessor, was built using Taiwan Semiconductor Manufacturing Co’s 5nm process and will be available to server builders such as Supermicro and Hewlett Packard Enterprise in the second quarter of this year.