AI data – Misryoum reports a UK firm plans 50,000 solar smart lamp posts in Nigeria to power low-energy AI compute without grid electricity.
A plan to turn street lamps into distributed AI power is gaining ground, and it could change how we think about where computing happens.
Misryoum reports that a UK-based company. Conflow Power Group. has agreed with Katsina State’s government to deploy 50. 000 solar-powered smart lamp posts known as iLamps across the region.. The idea is to use each iLamp as a small, networked compute node, reducing reliance on conventional, grid-fed data centers.
Each iLamp is built around a cylindrical solar panel and a battery, designed to run low-power hardware.. The company says the system can power a low-energy Nvidia chip consuming about 15 watts per unit. while the distributed network is intended to deliver substantial combined computing capacity without drawing electricity from the grid.
This kind of “edge-first” approach matters because AI infrastructure is increasingly constrained by power availability and the environmental footprint of large facilities. Using existing urban infrastructure, even at small scales, is one way developers are trying to make deployment more practical.
Beyond computing, the iLamps are positioned as multifunctional street systems.. Misryoum notes that the iLamps are intended to support traffic enforcement cameras for identifying speeding. parking violations. and seatbelt non-compliance. with facial recognition described as a future capability tied to identifying wanted or missing people.
The network also aims to provide public WiFi and Bluetooth connectivity.. Financially. the state is expected to generate revenue from traffic fines collected through the camera system. with Conflow taking a share after an initial period. while income tied to compute rentals is directed toward a green bond intended to cover installation and maintenance.
Still, lamp posts are unlikely to fully replace traditional data centers for the heaviest AI workloads.. Misryoum says experts generally expect limitations from how far nodes are spaced and the resulting communication delays. which can make demanding training and latency-sensitive tasks harder to support in a purely distributed street setup.
Even so. distributed networks like this may carve out a role for lighter AI use cases. functioning more like scaled-out access points rather than standalone replacements for massive server halls.. If talks expand across additional Nigerian states and institutions. Misryoum reports the broader deployment could grow into one of the largest distributed compute networks on the continent.
At the same time. Misryoum highlights a bigger backdrop: the global push for AI capacity is straining resources and raising concerns about the long-term waste burden of rapidly expanding hardware.. Solutions that emphasize lower power draw and staged deployment may help. but they also shift the challenge toward responsible scaling and end-of-life planning.