In case you missed it – Industry analyst Dean Bubley, founder of Disruptive Analysis, recently published an article in Light Reading challenging CTIA’s claim that artificial intelligence (AI) growth requires more exclusive-licensed spectrum for high-power national wireless networks. Bubley argues that many AI use cases instead point toward shared, local and enterprise-controlled connectivity models like Citizens Broadband Radio Service (CBRS) and Wi-Fi.
Bubley notes that AI-driven applications such as robotics, sensors, industrial systems and autonomous devices are often deployed in controlled environments like factories, hospitals, airports and farms. Those use cases align closely with CBRS’s strengths, including flexible, locally controlled connectivity that can support specific traffic patterns, latency needs and uplink-heavy applications without depending on nationwide macro-cellular networks.
“CBRS is almost purpose-built for that,” Bubley writes. “It is precisely the sort of spectrum model AI infrastructure needs: local, fast to deploy, enterprise-controllable, compatible with private LTE/5G, and configurable for very different traffic profiles.”
Rather than using AI as a rationale for more exclusive-licensed spectrum, Bubley points to CBRS as a proven model for supporting the next generation of connected devices, private networks and localized AI infrastructure. He also warns that overlooking the role of CBRS and unlicensed Wi-Fi risks targeting the very spectrum frameworks likely to support much of the AI-related traffic in homes, offices, campuses and enterprise settings.
Read an excerpt from the article below.
###
CTIA’s AI Report Makes An Accidental Case For Wi-Fi And CBRS
Light Reading
July 14, 2026
The CTIA’s May report on Wireless & AI makes a familiar move. It starts with some general observations about the evolution of AI, robotics, agents and sensors, but then jumps to a much bigger conclusion, wildly exaggerating the role of national high-power 5G/6G networks in driving AI innovation and adoption, as well as AI-related wireless traffic – and demanding more exclusive-licensed spectrum.
That chain of logic is much weaker than CTIA implies. And to the extent AI does expand wireless traffic, CTIA’s proposed solution – taking 275MHz of spectrum away from the Wi-Fi bands that are likely to carry the vast majority of this data – makes little sense.
Some of the technology trends are real. AI is indeed moving beyond chatbots. Humanoid robots, cameras, drones, vehicles and industrial systems will use more cloud and local intelligence and have different network demand profiles. AI agents may become ubiquitous, running on user devices such as smartphones and home gateways. They will use network connections too, perhaps more than their human owners. Some applications will be bursty, uplink-oriented and operationally critical.
But CTIA repeatedly treats “wireless” as if it mostly means high-power public cellular networks run by national operators. It does not. Exactly the same AI story also supports Wi-Fi, CBRS, private 5G, defense systems, satellite and aviation communications. And many new AI systems and applications actually use wired connections rather than wireless ones.
Check out the full article here.