UK Firm Transforms Street Lights into Distributed AI Data Centres

A Warwickshire-based tech company has introduced an innovative method to distributed computing by converting street lights into solar-powered AI data centres. Conflow Power Group Limited (CPG) has entered into a formal contract with a Nigerian state to install 50,000 of its connected iLamp units, which combine street lighting functionality with low-powered computing capabilities. The solar-charged lampposts are engineered to work collectively, providing the processing power of a traditional data centre whilst drawing no energy from the grid. The company argues the innovation represents a sustainable solution for artificial intelligence processing, though industry experts have cautioned that the technology is unsuitable for demanding computational tasks and better suited to lighter workloads.

The Innovation Behind Smart Lampposts

Each iLamp unit constitutes a precisely crafted combination of clean energy systems and processing equipment. The lampposts are outfitted with curved solar arrays that charge integrated batteries during daylight hours, which then power a compact low-energy computer contained in the structure. The advancement came by working together with chipmaker NVIDIA, which developed a chip designed to perform machine learning functions whilst consuming just 15 watts of power—a threshold reduced sufficiently to be reliably supplied by solar energy alone. This performance allows CPG to roll out installations without needing attachment to the power network, making them viable for deployment in isolated or disadvantaged areas.

According to CPG chairman Edward Fitzpatrick, the real power resides in deploying these systems across thousands of interconnected street lights. When interconnected, the distributed network establishes a shared computing platform that rivals conventional data center performance. The company’s outlook surpasses simple computing services; the lampposts can function as public lighting, CCTV infrastructure, and air quality sensors. This multi-functional approach maximises the value obtained from each installation, repurposing metropolitan assets into smart connection points within a larger urban intelligence network. The green advantages are substantial, as the system eliminates the high energy demands associated with traditional server facilities.

• Solar-powered units eliminate grid dependency and reduce carbon footprint
• NVIDIA 15-watt chip enables eco-friendly artificial intelligence capabilities
• Networked lampposts create distributed computing infrastructure
• Multi-functional design combines lighting, computing, and surveillance

Deployment and Real-World Applications

Conflow Power Group has started demonstrating the practical viability of its iLamp technology in real-world settings. The lampposts are now in use in the car park at Warwick Hospital, where they function as intelligent surveillance systems capable of CCTV monitoring and number plate recognition. These deployments serve as proof-of-concept installations, showcasing how the technology fits smoothly into existing infrastructure whilst providing tangible security and operational benefits. The company indicates positive results from these early implementations, which have shaped the design and functionality of units destined for expanded global deployment.

Beyond basic lighting and computing functions, the iLamps incorporate sophisticated artificial intelligence-enabled surveillance capabilities that enhance their utility significantly. The cameras can spot parking violations, detect speeding vehicles, and track seatbelt compliance—transforming ordinary street furniture into smart enforcement systems. CPG is also exploring facial recognition technology to identify wanted or missing persons, though such deployments would necessitate direct collaborations with competent bodies and full compliance with privacy legislation. Final-stage negotiations are underway with state schools and municipal bodies in Florida to deploy the entire set of these features in North American markets.

Expansion in Nigeria and Revenue Model

The company has established a formal agreement with a Nigerian state to implement 50,000 iLamp units, representing the largest commitment to the technology to date. This deployment will integrate artificial intelligence-enabled imaging systems capable of detect unauthorised parking, vehicles exceeding speed limits, and failure to wear seatbelts across the region. The scope of this implementation demonstrates considerable faith in the technology’s reliability and real-world effectiveness within developing markets where infrastructure investment remains a priority. Nigeria’s selection underscores both the technology’s suitability for the climate and the state’s dedication to upgrading urban infrastructure.

The Nigerian deployment exemplifies CPG’s business model, which surpasses upfront equipment purchases to cover ongoing data processing services and monitoring features. By treating the lampposts as decentralised computing hubs, the company derives earnings from computational services whilst concurrently delivering municipalities better traffic coordination and security features. This dual-revenue approach—merging infrastructure delivery with ongoing service provision—creates sustainable business opportunities in markets pursuing affordable smart city technologies. The model shows considerable promise in regions where standard computing infrastructure proves scarce or financially unviable.

• 50,000 units installed throughout Nigerian state for traffic and safety monitoring
• Revenue generated through computing services and surveillance capabilities
• Budget-friendly substitute for traditional data centre infrastructure deployment

Security Concerns and Technical Constraints

Whilst the notion of distributed AI data centers delivers economic and environmental advantages, industry experts have voiced substantial worries about the technology’s real-world feasibility and security implications. Data centre veteran Professor Ian Bitterlin warned the BBC that physical security represents a significant weakness, especially considering that each iLamp unit houses components valued at approximately £2,000. The streetlights’ exposed positions leave them as likely targets for larceny, a danger that cannot be entirely mitigated by design alone. Moreover, experts have questioned whether the technology can truly substitute for traditional data centres when handling demanding AI tasks, suggesting instead that iLamps may work well just for lighter computational tasks.

The technical limitations stem partly from the energy limitations inherent to solar-powered street lighting systems. Each unit relies on a cylindrical solar panel to power batteries that power a low-wattage computer, restricting the processing capabilities available for artificial intelligence tasks. Whilst NVIDIA has developed chips consuming just 15 watts—small enough to fit within street lights and powered entirely by solar energy—such limited processing power cannot replicate the capabilities of large-scale data centers. This core limitation means iLamps function best as secondary processing units rather than primary infrastructure, limiting their applicability to specific, less demanding AI tasks such as edge processing and localised data analysis.

Physical Security Measures

Conflow Power Group accepts the risk of theft and has introduced safeguards created to make stolen components unusable. The company claims that the internal chip would be “fried”—irreversibly damaged—if removed from its casing, thereby destroying its appeal to criminal elements. However, this protection addresses only the immediate problem rather than the fundamental weakness of having valuable electronics distributed across many publicly accessible locations, where motivated offenders might continue to attempt removal in spite of the security measures in place.

The Larger Context of AI Energy Use

The development of distributed AI data centres via street lighting reflects mounting apprehension about the ecological consequences of centralised computing infrastructure. Traditional hyperscale data centres require substantial volumes of electricity, with major facilities requiring hundreds of megawatts of continuous power to operate cooling systems and processing equipment. The environmental burden has faced increasing examination as artificial intelligence applications spread across the globe, driving demand for computational resources at unparalleled levels. Conflow Power Group’s proposition addresses this challenge by utilising established urban infrastructure—street lighting networks already embedded throughout towns and cities—to generate processing capacity without pulling extra power from the grid, theoretically decreasing the carbon footprint associated with AI deployment.

Solar-powered distributed systems offer theoretical advantages beyond mere energy conservation. By distributing processing tasks across thousands of linked nodes, iLamps could theoretically reduce transmission losses built into centralised data centre models, where power travels considerable distances through infrastructure. The approach aligns with broader industry trends toward edge computing, where processing occurs closer to data sources rather than in remote facilities. However, this vision must be balanced against practical realities: solar panels in Britain’s climate generate inconsistent power, battery storage stays limited, and the aggregate processing capacity of thousands of low-power units cannot match the sheer processing power required for training large language models or running complex AI inference tasks at scale.

Expert Assessment of Viability

Industry specialists remain somewhat doubtful about iLamps’ capacity to transform AI infrastructure. Whilst acknowledging the innovation’s value in particular applications, experts stress that decentralised street lighting systems cannot replace purpose-built data centres for tasks requiring significant computational power. The technology’s viability depends entirely on practical implementation expectations: iLamps perform best for edge computing applications where computational capacity stays limited and geographically distributed. For organisations requiring significant artificial intelligence capacity—whether developing neural networks or running inference at scale—conventional data centre systems remains essential, regardless of sustainability considerations.

Conflow Power Group’s agreement with Nigerian authorities represents a significant real-world test case, though successful implementation will ultimately establish whether the approach proves economically sustainable beyond initial trials. The company’s assertions about ecological advantages and decentralised computational capacity need verification through real-world performance metrics rather than theoretical projections. Success hinges upon proving that vast networks of iLamps can reliably deliver expected results whilst resisting security vulnerabilities and weather-related challenges. Until comprehensive deployment data becomes available, industry agreement suggests treating iLamps as a supporting solution rather than a revolutionary approach to data centre energy demands.

Data Protection, Monitoring and Moral Considerations

The integration of AI-powered surveillance cameras into street light systems raises substantial concerns about personal privacy and individual freedoms. Conflow Power Group’s proposal to equip iLamps with facial recognition capabilities, capable of identifying wanted individuals or missing people, constitutes a major extension of surveillance systems in public spaces. Critics argue that extensive rollout of such technology could fundamentally alter the connection between people and their cities, establishing an ever-present monitoring system that tracks movement and behaviour without explicit consent. The potential for misuse, function creep, and biased use of facial recognition systems remains a pressing concern for privacy advocates and civil rights organisations.

The company asserts it will only implement surveillance features in conjunction with relevant authorities and in full compliance with relevant legal requirements. However, this pledge provides little reassurance to those doubtful about existing protections surrounding surveillance technology. Face recognition technology have revealed significant bias against members of ethnic minority groups, raising questions about equitable application and potential discrimination. The absence of comprehensive legal structures governing such technology in many jurisdictions means rollout could advance with limited scrutiny. Without thorough independent assessment, transparent governance structures, and genuine stakeholder dialogue, iLamp surveillance capabilities risk entrenching institutional disparities whilst eroding fundamental privacy protections.

• Facial recognition bias disproportionately impacts ethnic minorities and vulnerable populations
• Lack of transparent governance and independent oversight of surveillance operations
• Function creep poses a risk of expanding monitoring capabilities past the initial intended use
• Insufficient legal protections do not adequately safeguard citizens from discriminatory technology misuse