IBM has taken a significant step in improving its cloud computing capabilities by announcing the integration of Nvidia H100 GPUs into its IBM Cloud platform. This upgrade will allow customers to access superior performance for their artificial intelligence (AI) workloads, both in training and inference tasks.
Significant performance improvement
The introduction of H100 GPUs represents a significant advancement over the previous Nvidia A100 GPUs that IBM had implemented last year. According to information provided by Nvidia, these new graphics processors offer up to 30 times faster inference performance than their predecessors.
Variety of options for different needs
In addition to the powerful H100 GPUs, IBM Cloud continues to offer a diverse range of GPUs to suit different scales of AI projects:
– Nvidia A100 GPUs: For high-performance AI workloads.
– Nvidia L40S and L4 GPUs: Designed for smaller-scale AI tasks, such as training small models or deploying chatbots.
Global availability and future plans
The new H100 instances are available in IBM’s multi-zone regions in North America, Latin America, Europe, Japan, and Australia. Additionally, the company has announced plans to offer Intel Gaudi 3 AI processors through IBM Cloud in early 2025, further expanding its options for specialized hardware.
Market positioning
Despite this advancement, it is important to note that IBM Cloud lags behind its main competitors in the adoption of H100 GPUs. Companies like Amazon Web Services (AWS), Google, Microsoft, and Oracle launched instances with these GPUs over a year ago. Even some startups specializing in cloud-based GPUs, such as CoreWeave and Cirrascale, had early access to this technology.
The future of GPUs in the cloud
The AI GPU market is evolving rapidly. Major cloud service providers, including Oracle, Amazon, Microsoft, and Google, have already announced plans to deploy the next generation of Nvidia GPUs, known as the Blackwell line. Nvidia expects to ship “several billion dollars” worth of next-generation GPUs in the fourth quarter of 2024.
Restructuring in IBM Cloud
Amid these technological advances, significant changes have been reported in the workforce structure of IBM Cloud. According to recent reports, the division is undergoing a large-scale layoff program, albeit discreetly. IBM has responded to these claims by indicating that workforce adjustments would represent a very small percentage of its global workforce, aiming to maintain employment levels similar to those at the beginning of 2024.
The incorporation of Nvidia H100 GPUs into IBM Cloud marks a significant milestone in the company’s strategy to compete in the cloud AI services market, although the actual impact of this upgrade on its competitive position is yet to be seen.
Source: DCD (https://www.datacenterdynamics.com/en/news/ibm-cloud-adds-nvidia-h100-gpus/)