Challenge concluded

Cambridge Edge AI Innovation for Sustainability Challenge 2026

The 2026 Challenge has concluded. HiveMind takes first place.

The final was held at the University of Cambridge on 29 September 2026. Three teams presented to the judging panel, and HiveMind from the University of Bristol was named the overall winner. This platform, which hosted registration and submissions as the Sustainable Edge AI Innovation Challenge 2026, is now closed.

1 July – 24 Augsubmission window
8 → 3semi-final to final at Cambridge
29 Sept 2026final and awards

Final results

  1. 1st

    HiveMind University of Bristol

    £6,000 first prize

    A collaborative crop-disease detection system for farms. Edge devices learn from one another without sharing raw images, so that disease can be detected earlier while data movement, energy use and communication demands are kept down.

    Wong E Chern, Yap Kar Boon, Yung Lynn Leo, Dr Chang Siow Wee, Aiman Rosli

    “Our idea is to allow AI systems to learn from each other without sharing the underlying data, while reducing unnecessary communication and energy use.” Wong E Chern, on behalf of HiveMind
  2. 2nd

    PyroWatch University of Manchester

    £2,000 second prize

    A low-cost, solar-powered network of devices that run small AI models locally to predict wildfire risk and support earlier intervention, reducing the need for continuous communication with central servers.

    Marwan Barakat Abdelfattah

    “It was the first time I had produced an independent research document of this kind, and I learned so much through the process.” Marwan Barakat Abdelfattah
  3. 3rd

    SustainGRF University of Hull

    Third place

    Uses smartwatch data to assess whether everyday activity supports bone health, with federated learning so that models learn from local devices without transferring raw sensor data.

    Parvin Ghaffarzadeh, Debarati Chakraborty

    “I wanted to bring this project closer to real life, and this competition was the perfect opportunity.” Parvin Ghaffarzadeh

About the Challenge

The Challenge asked a single question: when every node must keep its own raw data, and compute and energy are limited, can a system still produce collective intelligence that surpasses any single node, by exchanging the knowledge it has learned rather than the data itself? Teams designed multi-node Edge AI learning schemes and were judged on four criteria, each weighted equally: sustainability impact, ease of integration and feasibility, technical maturity, and creativity.

Submissions opened on 1 July 2026 and closed on 24 August 2026. After preliminary judging, eight teams were selected for the semi-final at the University of Cambridge on 28 September, and three finalists presented again on 29 September, when the results were announced.

“The future of AI will not just be defined by bigger models and larger data centres. It will also depend on Edge AI systems that can learn locally, protect data and use resources far more efficiently.” James Yang, CEO and Co-founder, MICROIP Inc. and Arculus System
“The real innovation challenge is how we create greater intelligence within real-world constraints.” Viola Jardon, Head of Innovation Programmes, Cambridge Institute for Sustainability Leadership

The Challenge was hosted by MICROIP Inc. and Arculus System in partnership with the University of Cambridge Institute for Sustainability Leadership (CISL). Our thanks go to every team that entered, to the judging panel in Cambridge and in Taiwan, and to CISL for convening the final.

Read more: CISL news release (1 October 2026) · University of Manchester news (2 October 2026) · Challenge page at CISL

Platform closure and your data

Registration, sign-in and submissions on this platform closed when the Challenge ended. No new entries are accepted and existing accounts can no longer be accessed here.

Personal data collected through the platform is handled as set out in the Privacy Notice. In line with that notice, data of teams that did not advance is retained for up to 90 days after the Challenge, so until 28 December 2026, and records relating to advancing teams and prizes for up to 365 days, so until 29 September 2027. Minimal audit records may be kept where needed for legal or dispute-resolution reasons.

To request access to, correction of, or deletion of your data, or to withdraw optional future-use consent, email challenge@arculussystem.com.