- Cityside Fiber sees fiber-to-the-home as the foundation for 5G, 6G, edge computing and smart-city infrastructure
- As AI moves into the physical world, distributed networks may become as strategically important as centralized data centers
- The emerging AI economy is splitting between America’s hyperscale LLM ambitions and a more distributed machine-learning model taking shape elsewhere
For Cityside Fiber, building fiber-to-the-home was never supposed to end at broadband.
The Southern California fiber operator was founded during Covid after its founders experienced firsthand the limitations of local broadband infrastructure. Today it has active service in 10 communities, another 16 under development and plans to continue expanding across Southern California.
But the more interesting part of the Cityside story is what comes next.
Jonathan Restivo, co-founder and chief development officer of Cityside Fiber, says the company’s founders came from the wireless and tower industries, where shared infrastructure has long been fundamental to the business model. They saw fiber-to-the-home in much the same way: not simply as an access network, but as infrastructure capable of supporting multiple future applications.
“We saw the ability to bring fiber to the home as kind of your foundation,” Restivo told me, enabling additional use cases further out into the network, closer to users.
Fiber finds new use cases
Those use cases potentially include 5G and eventually 6G small cells, edge data centers and smart-city infrastructure.
Some are already emerging.
Cityside works with municipalities on traffic-signal connectivity and city facilities, while exploring applications including earthquake detection, gunshot detection and traffic routing. The economics are compelling: once fiber is already running through a community, applications that could never justify a dedicated fiber build suddenly become viable.
That makes Cityside a useful microcosm of a much larger change taking place across the communications industry.
The world is rapidly separating into two AI domains.
The global AI domain split
The first is the world of probabilistic large language models concentrated inside gigantic data centers at the core of the network. This is where America is overwhelmingly focused: GPUs, AI factories, hyperscale compute and enormous concentrations of power, water and capital.
The second lies beyond the periphery of the traditional telecom edge, where machine-learning AI is moving into the physical world — factories, transportation systems, robots, energy infrastructure, sensors and cities.
Most other leading industrial economies are placing far greater strategic emphasis on this second domain.
I explored that split recently in Ericsson’s AI bet: The intelligent fabric vs the AI factory, and again in Meta’s $100B house of GPU cards.
The distinction matters because physical AI has radically different infrastructure requirements. A language model can sit inside a remote data center and generate probabilistic answers. A factory robot, traffic-management system or autonomous machine has to sense, decide and act in the physical world, often in real time.
Connectivity combines with computation
That makes connectivity increasingly inseparable from computation.
As Ericsson’s Åsa Tamsons recently put it, “AI will meet the physical world.” In my analysis, When the network becomes the operation, I argued that this pushes telecom beyond simply connecting industrial systems. Networks increasingly become part of the operation itself.
Restivo sees a similar architectural shift developing at the local level.
The industry has talked about edge data centers for years without ever quite delivering the promised market. But the extraordinary expansion of hyperscale AI infrastructure may finally change the economics.
“There’s only so much we can build at that scale,” Restivo said of giant centralized data centers. The question, he argues, is whether smaller data centers closer to users eventually become “table stakes.”
Community connection
That may be the real opportunity hiding inside today’s fiber boom.
Fiber-to-the-home gets the network into communities. From there, the same infrastructure can connect radios, sensors, municipal systems, distributed compute and eventually millions of intelligent machines.
Fiber enables the physical AI economy by turning last-mile broadband into a distributed infrastructure layer that supports edge compute, smart-city systems and machine intelligence beyond centralized data centers.
The broadband connection may merely be the beginning.
Stephen M. Saunders MBE is a communications analyst and USPTO-registered inventor examining how digital infrastructure — 5G, cloud and AI — is reshaping industry, power and society, as well as underpinning the emerging, ubiquitous global digital economy. As anchor of FNTV and a longtime industry insider, he focuses less on growth narratives and more on execution, risk and how hyperscale technology is distorting markets, governance and society at scale.
Opinion pieces from industry experts, analysts or our editorial staff do not represent the opinions of Fierce Network.