Low Size Weight and Power Edge AI
The iX programme is supporting Tekh to identify innovative hardware partners capable of developing low Size, Weight, Power and Cost (SWaP-C) computing hardware for Edge AI. The challenge will investigate how far AI compute can be reduced while retaining sufficient inference performance to deliver useful capability in constrained operational environments.
Opportunity
Challenge opens
05/10/2026
Challenge closes
06/11/2026
Benefit
Tekh is seeking innovative hardware partners to develop a new class of low Size, Weight, Power and Cost (SWaP-C) computing platform capable of supporting useful AI inference at the edge. The challenge will explore how far AI hardware can be reduced in complexity, power consumption and cost while still delivering sufficient performance for real world applications. The successful solution will support Tekh's research into affordable, scalable Edge AI and has the potential to enable new applications where the cost, power requirements or computational overhead of existing AI hardware currently limits deployment. Selected solution provider(s) will work with Tekh to develop and evaluate prototype hardware, with the potential for further development and commercial exploitation following successful trials.
Background
Tekh is a UK based research and development company specialising in Artificial Intelligence, Machine Learning and data driven technologies, with significant experience developing AI capabilities for constrained and operational environments.
Increasingly capable AI models can provide valuable functionality at the edge, but the hardware required to run them can introduce significant size, weight, power and cost (SWaP-C) constraints. This can limit where AI can be deployed and can make widespread deployment economically or operationally impractical.
This is particularly relevant to Defence, where there is growing value in deploying capability across larger numbers of smaller, lower cost and potentially unmanned platforms. In these applications, the objective is not necessarily to maximise AI performance. Instead, sufficient intelligence needs to be delivered using the minimum practical compute resource, power and cost.
Tekh is therefore seeking to investigate how far Edge AI hardware can be reduced while retaining useful inference capability. This requires specialist hardware design and manufacturing expertise alongside Tekh's existing AI research capability, enabling the relationship between model performance, compute, power, physical constraints and cost to be explored experimentally.
This challenge is brought to you by:


