Doug Recker, CEO of Duos Technologies, is pushing edge infrastructure into smaller markets. His strategy focuses on bringing GPU power closer to local customers.
The approach targets Tier 3 and Tier 4 communities. These areas often lack large data centers and strong compute access. Yet they increasingly need local processing for modern digital services.
Recker discussed the strategy on the Let’s Get Digital podcast. His background spans the U.S. Marine Corps, telecom sales, fiber, and data centers. That experience shaped his modular view of infrastructure growth.
At the center of the plan is Duos Edge. The unit deploys modular data center clusters near customers. These facilities support schools, healthcare providers, governments, carriers, and businesses.
The company first deployed 300-kilowatt pods. Demand quickly shifted toward larger sites. Customers now seek 1- to 15-megawatt facilities for GPU clusters and AI workloads.
This shift reflects a wider market change. AI applications need faster responses and nearby processing. Sending every task to distant cloud regions can increase delays.
Edge computing helps solve that issue. It places compute closer to where data starts. This matters for inference, autonomous vehicles, agriculture, and industrial systems.
For telecom operators, the model creates new partnership options. Carriers can link local compute with fiber and wireless networks. That may improve service quality in underserved areas.
However, the model also brings execution challenges. Smaller communities may have limited power, fiber routes, and skilled labor. Each site needs careful planning and local cooperation.
Duos aims to solve part of this through stranded power. These are power resources that remain underused. The company sees them as useful anchors for distributed data centers.
The economic angle also matters. Local data centers can generate tax revenue and technology jobs. They can also attract new enterprise services into rural regions.
Still, modular sites must prove long-term reliability. Customers running GPU workloads expect high uptime and strong equipment protection. Any local disruption can affect performance immediately.
Duos says it addresses this with patented facility designs. One feature includes a clean-room entry system. It helps reduce dust, pollen, and other contaminants.
Cooling also remains a major concern for GPU-heavy deployments. High-performance chips generate significant heat. Duos uses closed-loop cooling systems that require little added water.
The company is deploying more than 2,000 GPUs in modular environments. That scale shows growing confidence in distributed compute. It also signals rising demand outside hyperscale campuses.
Rather than chase gigawatt-scale projects, Duos targets repeatable mid-sized deployments. This approach may move faster than traditional data center development. It also fits markets that large providers often overlook.
For telecom and IT leaders, the message is clear. AI infrastructure will not live only in major hubs. More compute will move toward users, devices, and local networks.
If Duos executes well, smaller markets could gain new digital capacity. The model could support better connectivity and faster AI services. It may also redefine where next-generation infrastructure gets built.
