Enterprises across industries are constantly exploring new ways that artificial intelligence (AI) can support business innovation and growth. According to research from IBM, the majority of executives say their organization needs to quickly adopt generative AI (Gen AI) to accelerate innovation. However, only 39 percent of organizations are currently implementing or operating Gen AI for innovation and research. Many organizations lack the infrastructure and skills to successfully manage the data and deployment challenges that come with these compute intensive workloads.

IBM Cloud is helping our clients boost competitiveness with the combined power of our hybrid cloud and AI stack. Our enterprise cloud platform is built for even the most regulated industries, and coupled with our heritage in high performance computing (HPC), we are uniquely positioned to support the AI infrastructure needed for performance intensive workloads. We’re working on several new solutions and tailor-made collaborations to help enable a seamless and secure integration of AI.

David Tan, CTO, CrushBank:

“For CrushBank, it's critical that we deliver on the promise of excellent customer experience. The combined power of cloud and AI has been revolutionary in this area. By leveraging watsonx on IBM Cloud for our AI Knowledge Management platform, our agents can serve our customers faster and more effectively - decreasing total time to resolution by 45 percent. Working with IBM has also helped us reduce IT complexity, while helping to ensure we are meeting our security and compliance goals."

Performance Intensive Computing Momentum: Training and Executing AI Models

Over the past couple of years, IBM has announced several offerings to support training and inferencing generative AI models. IBM Research introduced Vela, IBM’s first AI-optimized, cloud-native supercomputer hosted on IBM Cloud, using NVIDIA A100 Tensor Core GPUs. Vela was designed to scale up on-demand and readily deploy similar infrastructure into IBM Cloud data centers, allowing IBM to collaborate with partners to train generative AI models. Over the last year Vela’s primary mission was to train IBM’s Granite models and drive AI development. Last year, IBM also announced the availability of NVIDIA GPU offerings (NVIDIA A100 Tensor Core) on IBM Cloud which clients are using for inferencing on enterprise-class foundation models via watsonx services or as GPUaaS for their specific needs.

We’re also working with IBM Research and Red Hat on several initiatives that will support generative AI implementation. The IBM Cloud’s GPU Infrastructure is being used to train watsonx Granite models and support InstructLab, an open source AI project that facilitates contributions to Large Language Models (LLMs). Granite is IBM's flagship series of LLM foundation models that are trained on an enterprise’s trusted data spanning internet, academic, code, legal and finance. IBM Cloud will also be adding support for Red Hat Enterprise Linux AI (RHEL AI) and Red Hat OpenShift AI platforms that enable users to develop, test and deploy Gen AI models more seamlessly.

Joe Fernandes, vice president and general manager, GenAI Foundation Model Platforms, Red Hat:

“More users in more organizations need to be able to access the innovation of GenAI, which means lowering the barriers to contributing and aligning LLMs. RHEL AI and the InstructLab project, launched in collaboration with IBM, do exactly this by bringing more accessible AI models to a broader community of users, including domain and subject matter experts, wherever they are across the hybrid cloud.”

Building on our GPU Momentum: Inferencing with NVIDIA L4 and L40S GPUs

IBM is expanding its NVIDIA GPU offerings in IBM Cloud and is among the first cloud providers to offer NVIDIA L40S. IBM Cloud also offers NVIDIA L4 Tensor Core GPUs to support both AI and other accelerated workloads including 3D graphics and video applications. This will enable our clients to deploy these GPU instances on-demand in an enterprise cloud platform with NVIDIA AI Enterprise software that is optimized to help address the needs of mission-critical workloads across resiliency, performance, security, compliance, and total cost of ownership.

The direct access to NVIDIA GPUs on IBM Cloud in VPC and managed Red Hat OpenShift environments gives clients the flexibility to choose fit-for-purpose accelerators for AI training, fine-tuning, and inferencing. This can help lower the costs of running AI-powered applications. NVIDIA GPUs can integrate with data and AI platforms and services on IBM Cloud, including watsonx to help you build, scale and govern AI. IBM Cloud anticipates that it will also enable the latest NVIDIA Blackwell platform in 1H 2025.

Later this year, IBM Cloud will be adding NVIDIA H100 GPUs, based on the NVIDIA Hopper architecture, to nine regions with the ability to cluster multiple systems in Washington, D.C. and Frankfurt, Germany. We also plan on introducing NVIDIA H200 Tensor Core GPUs in select regions early next year.

Accelerating Automation with Deployable Architectures

In addition to our work with partners, IBM is bringing our clients an integrated, enterprise platform experience with a set of pre-configured patterns to help with automating our solution offerings. Deployable Architectures (DAs) can offer quick, easy, and secured pre-configured deployment automation for IBM offerings on IBM Cloud. Our mission with these pre-configured offerings is to make it easier for businesses and developers to drive outcomes in hours, while also supporting them in their security and compliance goals. By addressing the risk and burden of deployment, clients will be able to redirect resources back into their business.

IBM Cloud is on a mission to drive AI and hybrid cloud innovation for our clients. We have the infrastructure to train our own models and we have the leading skills from IBM Consulting to help clients define and execute hybrid cloud strategies that can drive tangible business value and help harness the potential of generative AI. Our solutions are designed to help transform how enterprises, developers and the open-source community build and leverage generative AI.

Learn More

For more information and to leverage the watsonx portfolio on IBM Cloud, visit HERE for a free trial.

To learn about addressing risk and regulatory
requirements for your AI workloads,join our webinar HERE.

Statements regarding IBM’s future direction and intent are subject to change or withdrawal without notice and represent goals and objectives only.