IBM Corp. is teasing a number of details about how it’s expanding its partnership with Amazon Web Services Inc. to advance responsible artificial intelligence for mutual enterprise customers.

The company said it will showcase a number of key collaborations at AWS re:Invent next week, including the launch of its most powerful Granite models on the Amazon Bedrock and Amazon SageMaker JumpStart services. It’s also planning a new integration with the IBM watsonx.governance platform and Amazon SageMaker, plus various full-stack observability tools for Amazon’s leading AI platforms.

According to IBM, it’s becoming more imperative than ever for enterprises to choose the most appropriate large language model to use in each new AI application. With its latest Granite 3.0 models, available now in Amazon Bedrock and Amazon SageMaker, the company says it’s offering access to some of the industry’s most powerful, trusted and cost-effective models in the business, directly within AWS’ powerful AI development environments.

Developers will be able to use the Granite models to power various kinds of AI applications they build on those platforms, and they’ll also benefit from additional integrations. For instance, the Granite models will gain support for AWS Neuron, so developers can run deep learning workloads on Amazon’s AWS Inferentia and AWS Trainium instances. It’s also working on integrating watsonx with Trainium2, which is Amazon’s most advanced AI chip of all, designed specifically for high-performance training and inference workloads.

The partnership extends beyond just models and processors, though, as IBM is also putting a lot of emphasis on “responsible AI” to ensure effective governance of AI applications. IBM watsonx.governance provides users with access to a comprehensive framework for governing the entire AI model lifecycle. It will soon be integrated directly with the SageMaker platform, providing developers with integrated, customizable risk assessment and model approval workflows, among other capabilities.

Responsible AI is not just about governance, as it also requires comprehensive security tools. IBM is making these available too, through the IBM Guardium AI Security software-as-a-service platform, which is also being expanded to Amazon’s cloud. It’s being made available on the AWS Marketplace starting today, giving users the tools they need to discover AI deployments, find and mitigate vulnerabilities, and protect their most sensitive data pipelines, IBM said.

Other features coming to AWS soon include a preview of IBM Instana Generative AI Observability for optimizing the performance of AI models on SageMaker and Bedrock, and previews of new capabilities in IBM Concert and IBM Instana, which offer predictive intelligence for AI application environments.

For some customers, the comprehensive selection of tools may not be enough to satisfy their desire for responsible AI, so IBM is offering to go even further with its expanded consultancy offerings.

Holger Mueller of Constellation Research Inc. said it’s no surprise to see IBM flexing its AI muscles at AWS re:Invent, which is by now the biggest party of the year in the technology industry.

“These days it’s key for companies like IBM to be at re:Invent in order to carve out a substantial portion of enterprise’s 2025 IT spending budgets, and that’s why it’s showcasing one of its strongest AI assets, namely the Granite models that can now be found in Amazon Bedrock and Amazon SageMaker,” the analyst said.

According to Mueller, the Granite models should be be attractive to the developer-centric core AWS customer base, so the collaboration is a smart move, putting them in the hands of thousands of exciting innovators using Amazon’s AI tooling.

“Developers will also be happy to see IBM extend the watsonx.governance platform and other tools to Amazon’s cloud,” Mueller said. “This is good news for enterprises because 2025 will be the year that they need to get going with their AI efforts. By combining the infrastructure services of AWS with the professional capabilities of IBM, many should be able to do so.”

In addition to its various AI tools, IBM will also use re:Invent to showcase its consulting services and how these are helping enterprises develop various customized generative AI applications and services. For instance, it plans to demonstrate a number of examples of custom solutions developed using the Amazon Q platform to meet specific business needs.

Along with this initiative, IBM is making a number of its consulting services, such as IBM Consulting Contact Center Intelligence, IBM Consulting AIOps on AWS, IBM Consulting Digital KYC and IBM Consulting Supply Chain Ensemble, available on the AWS Marketplace.

AWS Managing Director of Technology Partnerships Chris Grusz said IBM is one of the company’s largest partners of all. As such, the collaboration goes way beyond just tooling.

“IBM has gone far beyond making their technology available on our platforms,” he said. “It’s prioritizing integration with AWS at each layer of the AI stack, making it easier for our mutual customers to choose the technology that best fits their needs and benefit from responsible AI in new ways.”

Photo: IBM