- A British startup wants to bring computing even closer by integrating micro data centers inside streetlights
- Self-destructing Nvidia chips will ensure sensitive information remains private
- Some challenges need to be addressed before this becomes a reality.
British startup Conflow Power Group Limited (CPG) has proposed a major restructuring of the world’s data centers by integrating micro units directly into urban street infrastructure, such as streetlights, rather than concentrating computing within hyperscale facilities as we know them today.
In addition to distributing computing into cheaper, more manageable micro-locations, the plan also focuses on local solar power generation and battery backup systems to address one of the biggest criticisms facing data center campuses: sustainability and environmental impacts.
Under the new proposals, CPG aims to bring AI computing closer to users and devices, which would reduce latency and relieve pressure on the national telecommunications infrastructure.
Future data centers could be located right outside your front door
He bbc reports that $2,000 Nvidia AI accelerators could be used instead of high-end, flagship GPUs like the H100 and B200 systems, which cost tens of thousands of dollars per unit.
The use of self-destructing chips is also a notable addition to the scheme, as Nvidia includes firmware locking, encryption, and other anti-tamper protections that can effectively disable hardware if it is compromised, relocated, or accessed by unauthorized methods or persons.
These anti-tamper technologies already exist for export compliance and are typically seen when AI accelerators are sold in restricted or cutting-edge implementations to ensure maximum security.
Tens of thousands of micro data centers could spread computing across cities, each handling localized AI workloads that could span applications such as traffic monitoring, CCTV, autonomous vehicle coordination, telecommunications, environmental sensing, and more.
While the idea of integrating data centers into streetlights may be novel, pushing AI to the edge is a trend that is accelerating as AI workloads become more sensitive to bandwidth and latency. Placing computing physically closer to where it is applied would also likely reduce costs associated with data transfers.
The energy implications are also compelling, as hyperscale data centers face limitations due to delays in network connections and unsustainable supplies. Geographically distributing energy consumption certainly solves current bottlenecks, as some of the largest campuses consume as much electricity as small cities.
Streetlights in particular are attractive due to their existing electrical connections, their dense distribution in urban environments such as towns and cities, and the fact that many are also already connected to fiber networks.
Many have already been converted into 5G small cells, traffic camera mounts, Wi-Fi hotspots and electric vehicle chargers.
Interestingly, there are also important geopolitical benefits to locating micro data centers on existing road infrastructure networks. Growing concerns about sovereign computing in Europe and the UK are causing governments to push for schemes that support more local processing.
However, for them to work at scale, CPG would have to address the many challenges that come with integrating computing into streetlights. Existing infrastructure may need to be upgraded to ensure protection from weather and vandalism. There are also concerns around thermal management, as large-scale data centers are scheduled for excessive water consumption for cooling.
While improving existing infrastructure may not be economically viable, building multipurpose networks in the future could allow a single element, such as a streetlight, to serve many more functions than we had previously imagined.
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds.
