- Cloud-based radio now matches traditional hardware performance, removing a key barrier to open and cloud RAN adoption, said Dell's Sandro Tavares
- Telcos face a strategic fork: become an edge AI and sovereign cloud provider or optimize ruthlessly as a connectivity-only business
- Delivering edge AI to small and midsize businesses is an underappreciated telco opportunity
DELL TECHNOLOGIES WORLD, LAS VEGAS — The telecom industry has been on the verge of transformation for more than a decade, always nearly unlocking new revenue streams and business models but never quite getting there. That's finally changing, according to a Dell executive.
Open RAN is key to the transition, said Sandro Tavares, who until recently led Dell's telecom marketing and now heads enterprise server marketing (including telco). For years, telecom operators resisted cloud-based radio solutions because proprietary hardware simply outperformed it. But cloud-based RAN has finally caught up with proprietary performance.
"When you get to the radio, there was still a pretty significant performance gap between traditional solutions and cloud-based solutions," Tavares told Fierce. With Dell's XR8720 and now the XR 9700 — a ruggedized, IP66-rated unit designed for outdoor far-edge deployment — "that performance gap is pretty much gone."
The industry, however, hasn't caught up with the technology. Open RAN still represents less than 25% of actual deployments globally, even though 74% of tier-1 operators have open network architecture plans, according to the 2025 Open Network Index compiled by Dell and Analysys Mason.
AT&T is an early adopter of open RAN at scale, with Ericsson providing software and Dell providing hardware, Tavares said.
And Dell launched an on-premises sovereign AI stack earlier this week, in part to serve operators looking to build sovereign cloud infrastructure for regulated enterprise customers and governments.
Where telcos can compete with hyperscalers
Open infrastructure is key to unlocking new business models for operators, Tavares said.
Telcos can compete at the edge, offering latency-sensitive AI inferencing, running workloads that generate massive data volumes at the edge and providing sovereign AI deployments that cannot legally route data through U.S.-based cloud providers, Tavares said.
"If you can actually be running [workloads] locally, or as close as possible to where data is being created, there is a benefit there," Tavares said. Telcos still control the last mile and reach places hyperscalers have no structural interest in building toward. AWS Wavelength, Azure Edge Zones and Google Distributed Cloud are all pressing toward the edge, but telcos retain an inherent advantage in physical proximity and regulatory relationships.
But telcos should not try to take take general-purpose compute from AWS, Azure or Google, Tavares said. Hyperscalers are better able by far to serve centralized cloud applications, such as CRM and other traditional enterprise software.
The SMB telco opportunity
Small and medium-sized businesses present an opportunity for telcos to deliver AI bundled with connectivity, Tavares said.
"There is a kind of common mindset that AI is for huge corporations," he said. Smaller businesses can benefit enormously from AI, but typically lack the in-house expertise to deploy AI systems on their own.
Dell participated in outfitting a small brewery with computer vision monitoring for fermentation tanks and canning lines, eliminating overnight manual checks and driving enough efficiency improvement that the company moved to a larger facility. T-Mobile provides connectivity for that deployment, Tavares noted.
The SMB advantage also applies to local government. Dell worked with the city of Bellevue, Washington, alongside AT&T and Nvidia, on a computer vision system for traffic management — pedestrian detection feeding directly into signal control, with measurable impact on accident prevention.
Two roads, one choice
But it's OK for operators to remain as pure connectivity players, Tavares said.
Operators who make that choice will need to become great at optimization, transforming their network infrastructure. Demand for capacity will keep growing while revenue per bit declines. Operators clinging to traditional, vertically integrated architectures will be squeezed.
"Even if they decide to be a connectivity player and just a connectivity player, they're going to have to be extremely optimized," Tavares said.
For those choosing the AI path, technology is no longer the primary obstacle. The organization is, Tavares said.
Telcos that want to become digital service providers must retrain sales teams to sell beyond connectivity, retool operational KPIs beyond network uptime metrics and build the agility to launch new services on a faster cadence. "It is a real transformation, basically, of who they are as a company," Tavares said.
Then again, telcos have tried this kind of pivot before — Verizon's public cloud play ended with a write-down and a fire sale of its data centers in 2016. The current enthusiasm around AI factories carries some of the same risks.