Amazon Web Services (AWS) remains focused on cloudifying telecom operations – work that has been amplified by AI – but knows it must be patient in that long-gestating and complex process.

Amir Rao, global director for telecom solutions at AWS, recently explained to SDxCentral that the pattern operators are taking on their cloudification journey remain diverse, and that “there’s no pattern, I’ll be honest.”

“There are few who are much more open to having this because in a way, looking at it from a top-down perspective, a CEO coming in or a board coming in and saying, ‘Okay, Jensen Huang says this, [AWS CEO Andy Jassy] says he's spending $200 billion on AI, and I'm not seeing even a 1% cost eradication. What are we doing?’ For those guys, this level of thinking resonates really well,” Rao said. “There are, of course, other operators who think that they would want to focus on the cost control and really need to analyze which use cases need to be automated, identified first, they may have a different trajectory and probably a slower pace of identification.”

These diverse paths have been a challenge for cloud providers to navigate, but that doesn’t mean they have stopped trying. Many continue to tout significant telecom cloud wins, but as Rao noted, the scale of those opportunities has required hyperscalers to adjust their message to those “other operators.”

Rao said that AWS has not “taken the pressure off the pedal any single bit in the context of cloudification” and instead has adjusted that pressure to more accurately take into account legacy systems – like mainframes – that might be hindering modernization efforts.

“A lot of these operators have homegrown OSS/BSS [operating support systems/business support systems],” Rao said. “We've seen this in Korea. We've seen this even in the U.S. They have homegrown systems on BSS and billing and mediation, which are still running on mainframes for the last 30 years. Now, with the whole agentic era, another opportunity is not only the agentification of the processes itself, but also leverage whether you, in case of Amazon, leverage Amazon Kiro or Bedrock and Bedrock Agent Core in the context of reinventing those applications.”

Rao noted that AWS has used these paths with U.S. operators like AT&T and with operators in South Korea and Japan, but that many other operators are in a technology bind.

“Not everybody has this capacity to build their own homegrown system,” Rao said, adding that many might have traversed this challenge by going to a centralized platform vendor that “are actually in better shape. The people who had the wherewithal to do their own development, they got, unfortunately, larger tech debt because the fact that they were builders as well as buyers.”

Using AI to helping those in tech need

Hyperscalers have traditionally focused on using the cloud as a foundation for this help, tapping that base as a way to introduce autonomous network operations. But Rao explained that this thinking can only go so far.

“For a network to be autonomous, the applications have to be autonomous,” Rao said, noting that this has been “mostly achieved” for software “but the cloud infrastructure, especially if it's [on-premises] and an appliance-based system, it can't breathe. It can't exhale or inhale or fly on demand.”

Putting more of that system in the cloud can help, but AWS is now seeing how AI can help assist in that process.

“So, the key focus from our perspective is to help the telco service providers realize the agentic value across the value chain of the businesses, across all of their knowledge processes,” Rao said. “I don't want us to sound too full of ourselves, but our thesis is that if the AI era is going to allow us to completely reimagine the business, which includes the business processes and the way we've historically designed and implemented networks from a technology stack perspective, we first have to meet the industry where it is today, which is based on their existing knowledge processes.”

This can help those operators identify where they are in terms of “existing knowledge processes,” which Rao said AWS has learned in working with operators on their cloudification journeys.

“And I believe that the agentic era, the foundational premise, will not change,” Rao added. “We'll have to meet the industry where it is first before we can embark upon some of the extremely disruptive thinking about completely revamping or redesigning the entire business, which may present, of course, productivity gains as well as the top-line enhancements, but it is going to be a journey.”