Intel’s Gaudi 3 AI chips are now available via IBM Cloud.
Announced by IBM at the Intel Vision 2025 event, the chips are available in IBM’s Frankfurt (eu-de) and Washington DC (us-east) cloud regions. In Q2 2025, this should extend to the Dallas (us-south) region in Texas.
The Intel Gaudi 3 accelerator was revealed by Intel in April 2024. The company claims that the 5nm Gaudi 3 accelerator offers 2x more AI FP8 compute power, 4x more AI compute for BF16, 2x network bandwidth, and a 1.5x increase in memory bandwidth over its predecessor, the Gaudi 2, and when compared to the Nvidia H100, it delivers 50 percent better inference on average across Meta’s Llama2 models with 7B and 13B parameters, and OpenAI’s GPT-3 175B parameter model while being a “fraction of the cost.”
IBM hopes that the Intel Gaudi 3 AI chip will help its customers test, innovate, and deploy generative AI solutions more effectively.
“By bringing Intel Gaudi 3 AI accelerators to IBM Cloud, we’re enabling businesses to help scale generative AI workloads with optimized performance for inferencing and fine-tuning. This collaboration underscores our shared commitment to making AI more accessible and cost-effective for enterprises worldwide,” said Saurabh Kulkarni, vice president of data center AI strategy and product management at Intel.
The Intel chips can be deployed through IBM Cloud Virtual Servers for VPC cloud instances and from Q2 will be available as a containerized worker node for Red Hat OpenShift AI clusters and Red Hat OpenShift on IBM Cloud.
“Bringing together the power of Red Hat AI with Intel Gaudi 3 AI accelerators on IBM Cloud offers organizations a clear path to accelerate their AI roadmaps. Red Hat is pleased to build upon our long-standing collaborations with IBM and Intel to deliver end-to-end open-source AI solutions to help organizations realize the value of AI and generative AI faster,” said Steven Huels, vice president of AI Engineering, Red Hat.
IBM first revealed its intention to bring the Intel AI chips to its cloud in August 2024. The company also offers access to the Nvidia H100s, and has said it will add AMD Instinct M1300X GPUs in 2025.
