IBM's Watson Anywhere lets customers run AI on any cloud they want, and it's a sign that IBM is pulling back from plugging its own cloud (IBM)

ginni rometty

On Feb. 12, IBM announced Watson Anywhere, which allows customers to run IBM's flagship artificial intelligence service on any cloud they want or even on their own equipment.
Analysts believe this is an effective strategy and shows that IBM is focusing on its customers who want to use multiple clouds or a hybrid cloud, rather than pushing on its own cloud.
But analysts are also skeptical about whether Watson has significant advantages over the artificial intelligence services provided by bigger clouds, like Amazon Web Services or Microsoft Azure.

Instead of losing its breath trying to keep up with the top three leaders in the cloud race, IBM is now taking a more flexible approach, analysts say.

On Feb. 12, IBM announced Watson Anywhere, which allows people to use the company's Watson artificial intelligence on any cloud they want, whether it's a public cloud, a private cloud, or hybrid cloud -- a combination of cloud and data centers.

Watson Anywhere is optimized for IBM's cloud, but the fact that it can also run on any other cloud is a sign that IBM is looking to capitalize on important market trends, such as customers who want to use multiple clouds, says Sid Nag, research director at Gartner.

"IBM is saying, we're not going to compete with the usual suspects," Nag told Business Insider. "We're doing one better where we're going to take our technologies and overlay that not just on IBM cloud, but also the other cloud providers like Amazon, Google, Azure and others. That's their strategy creating more velocity around IBM cloud ecosystem."

Since analysts say it's unlikely that IBM's cloud will reach the scale of Amazon Web Services, Microsoft or Google anytime soon, they see it as an effective strategy that shows IBM is responding to what customers need.

A "very good step"

John Roy, lead analyst at UBS, says allowing customers to use Watson wherever they want is a "wise decision." Otherwise, if IBM kept requiring people to use Watson on IBM's cloud, it would not be sustainable.

"Watson Anywhere is certainly a very good step," Roy told Business Insider. "I think making it available in whatever platform the end user wants to use it on is a very good strategy...You want your core software products used in as many places as possible."

An AI service like Watson that only works on IBM's cloud would be a useful strategy if IBM's cloud had reached sufficient scale, like Amazon or Microsoft's clouds. Currently, AWS and Microsoft Azure offer services that work exclusively on their clouds.

Dave Bartoletti, vice president and principal analyst at Forrester, says Watson Anywhere is somewhat similar to Google's approach of making its AI services, like TensorFlow, available to run anywhere.

"The IBM public cloud has never reached the scale of AWS or Azure, so IBM can’t afford to limit the potential of Watson to its own cloud," Bartoletti told Business Insider. "IBM’s betting that Watson can compete with native public cloud AI services well enough to generate revenue, and that it doesn’t make sense anymore to tie Watson to IBM public cloud."

On the downside, Nag questions whether customers will choose to use Watson, instead of artificial intelligence services that are already provided by the cloud they're using, such as Amazon Rekognition.

The question clients may have, Nag says, is "Why would I use IBM's AI functionality over the major functionalities around AI that my provider already has?"

"IBM's Watson functionalities works on multiple clouds, so that's definitely an advantage, but it's going to be a decision making process on behalf of the buyer," Nag said.

One thing Watson has going for it is its ability to understand natural languages, but some analysts are skeptical about whether it signals progress.

"Watson Anywhere isn't a competitive advantage yet, even though it holds some promise in the future," Clement Thibault, senior analyst at Investing.com, told Business Insider. "I believe this isn't enough to tip the scales in IBM's favor when it comes to cloud providers at the moment."

A hybrid cloud strategy

The fact that Watson can run on hybrid cloud is an advantage, Roy says. Right now, many companies still have to keep some workloads in their in-house data centers due to regulations, and the only top 3 cloud provider that has a generally available hybrid cloud service is Microsoft. This means hybrid cloud customers can use Watson instead of AWS or Google's AI services.

As for its own cloud, IBM will focus its energies on its upcoming acquisition of Red Hat to enhance its hybrid cloud.

In the near term, analysts sees IBM focusing on its software and consulting services that help customers manage different clouds, rather than pushing its own cloud forward. And company trust is on IBM's side — customers still see IBM as a strong player in the enterprise.

"IBM is saying, 'We're going to meet the customer where they are, give them choices and gain more revenue to the service rather than build a public cloud,'" Nag said. "That's their strategy."

Join the conversation about this story »

NOW WATCH: Meet the three women who married Donald Trump