IBM brings Watson AI to AWS, Microsoft Azure and the Google Cloud

Watson AI can now be run across all public, private and hybrid clouds using Kubernetes

IBM has revealed that its Watson artificial intelligence (AI) service will now work across Amazon Web Services (AWS), Microsoft Azure and the Google Cloud.

In a bid to win over customers demanding greater flexibility in how they store and analyse data, IBM revealed the development at its Think conference in San Francisco, California on Tuesday, stating that the Watson intelligence will soon be available not just on AWS, Azure and the Google Cloud, but on pretty much any cloud platform, including hybrid, public, and private.

IBM CEO Ginni Rometty said on stage during a keynote address that the announcement represents "a new chapter" in the journey from AI experimentation to widescale deployment and industry transformation. "It's the most scalable AI in the world," she added.

Customers interested in taking advantage of the Watson "Anywhere" service will be able to connect data stored in any of those clouds or in private data centres to Watson through IBM Cloud Private for Data, which already runs on multiple clouds.

Essentially, the move means that companies will be able to use Watson AI tools, such as Watson Assistant, which can help them develop conversational services, such as a virtual customer service agent, in mobile apps hosted on Amazon.com and Microsoft, as well as IBM servers.

"Businesses have largely been limited to experimenting with AI in silos due to the limitations caused by cloud provider lock-in of their data," said general manager of IBM Data and AI, Rob Thomas. "With most large organisations storing data across hybrid cloud environments, they need the freedom and choice to apply AI to their data wherever it is stored."

To make the links to rival clouds, IBM will rely on open-source software package Kubernetes. However, it's not the first major tech company to do this. In recent months, Google adopted Kubernetes for its AI solutions by launching Kubeflow Pipelines. Then there's the Linux Foundation for Deep Learning, which introduced its Acumos platform late last year, followed more recently by Intel, who debuted its Nauta platform last month, also with Kubernetes.

The Watson on any cloud announcement also builds on IBM's attempt to position its services as compatible with nearly any form of computer infrastructure a customer chooses. This also includes its pending acquisition of open-source software vendor Red Hat for an eye-watering $34 billion (£26bn).

IBM said its new AI service will be offered through its Automation Platform for Digital Business, and made generally available later this year.