Building AI-ready Telecom Networks: Insights from Industry Leaders

AI innovation is shaking up the telecom industry. Are telecom networks equipped to handle it? Get insights from Pure Storage experts and other industry leaders.

AI-ready Telecom Networks

Summary

Two new reports and a webinar explored the impact of AI in the telecom industry and how telecom companies can be better prepared to build AI-ready networks.

image_pdfimage_print

AI-led innovation is sweeping across the telecom sector. Enhanced customer experience, network security, and new revenue opportunities are just a few of the focus areas. The ultimate telecom goal is a truly autonomous, self-healing network. But a lot of work remains to be done. 

To help carriers better understand what telecom AI means at the network level, Pure Storage partnered with Fierce Network to create a new research report, “Beyond bandwidth: Building AI-ready telco networks for the next generation of data demands.” The report includes interviews with representatives from MetTel, Telus, and Verizon, along with Pure Storage Telecom Field CTO Patrick Lopez and Chris Penrose, VP, Head of Telecoms Business Development at NVIDIA. 

As a companion event to the report, Lopez was joined by Penrose for a webinar discussion moderated by Mitch Wagner, Chief Analyst at Fierce Network. 

The conversation was wide-ranging and insightful. Wagner opened up by asking, “How is AI changing the way telcos think about data?”

Lopez responded with a true industry perspective. “AI is a strange thing for a telco network. They haven’t been designed to handle AI or to use AI.” A telco network is essentially a greenfield AI environment for telcos, even though they’ve been using aspects of AI and machine learning for many years. 

Lopez provided both a history lesson and insight into the current state of things: 

“Historically, network functions had their own data storage based on spinning disk. Some of the collected data would go into a data lake to be formatted and prepared, and then moved to a data warehouse to apply analytics. And that’s OK to analyze up to half a day or an hour ago.” 

– Patrick Lopez, Pure Storage Telecom Field CTO

“Where it becomes difficult is when you want to automate. When you want to have the AI not only observe and detect, but actually predict and suggest. Or with agentic AI, take decisions and make operations. To do that, you need to reduce the time to insight…  

The old data architecture of cascading data and multiple protocols isn’t good for AI, because it starves the GPU, and you end up with late decisions.”

Penrose agreed and added, “You have to have data to make AI work.” Speaking from NVIDIA’s perspective, he noted: “Telcos have an incredible amount of data. How do we help them with their entire data pipeline?”

This is one of the key challenges as telcos strive to turn their massive amounts of internal data into actionable, useful AI outputs. Ultimately, it means re-examining the entire network. “You’re only as strong as your weakest link,” Penrose said. “So you have to look at the whole pipeline. What does the networking look like, what does the storage look like, and you need optimized software to get the most out of the GPU.”

That was just the opening exchange. The conversation continued, discussing autonomous networks, telco AI factories, the importance of sovereign AI, technology procurement models, power and space concerns, and much more. Tune in for yourself to take advantage of the insights provided.   

The Advent of AI-RAN 

Another area where AI is making an impact is in the radio access network (RAN). It’s early days, but advances are happening. Pure Storage announced in March that we had joined the AI-RAN Alliance and are working with many carriers and vendors to move the technology forward. 

In conjunction with The Mobile Network, we recently co-sponsored (with Fujitsu, Nokia, and Viavi) a very informative 36-page report, “AI-RAN Market Update.” As the report notes, the most advanced carrier to date is Japan’s SoftBank. They’re rebuilding their entire network with support for AI-RAN. As the report notes, “SoftBank’s AITRAS, which stands for AI & Telecom Radio Access Solution, incorporates the vRAN and AI on the same GPU server, which is the Nvidia Grace Hopper 200. Red Hat OpenShift handles the platform apps and cloud environment, whether they are AI apps or the vRAN running concurrently.”  

While SoftBank is the current leader, T-Mobile is committed to AI-RAN and has established an AI-RAN Innovation Center based in Bellevue, Washington. 

There’s a great deal more information in the report, including an article from Pure Storage on “Why storage will play a crucial role in the AI-RAN vision” (see page 19). AI continues to create massive impact in telecom, and Pure Storage is working alongside our many telecom customers to drive innovation and change. 

Explore telecom and AI solutions from Pure Storage.