Extreme heat is reshaping how cities operate. It strains health systems, drives up energy demand and exposes vulnerable residents to conditions their buildings and infrastructure were never designed for. At the same time, more granular temperature data is creating new possibilities for how cities, infrastructure operators and other decision-makers anticipate and respond to temperature risk.

FortyGuard Hackathon'26 brought together nearly 4,000 participants across 842 teams from more than 50 countries and six continents. Thirteen projects were shortlisted and three winners selected. Judges assessed submissions on impact and relevance, technical execution, innovation and communication.

The winning projects apply temperature intelligence to very different challenges, from prioritising urban cooling investments to reducing mechanical cooling in data centres and protecting crops from frost. The results are prototypes and modelled applications rather than demonstrated impacts at scale. But they illustrate how temperature data can increasingly be translated into practical decisions.

The Urban Canopy maps Midtown Manhattan building by building

The winning project, developed by the Ministry of Temperature team from New York University, takes street-level heat data and turns it into a costed, prioritised action plan for an urban district.

Image
Hackathon

The Urban Canopy models 29,415 façade panels across 5,329 buildings in Midtown Manhattan, hour by hour for a full year, using measured FortyGuard air-temperature data. A simulated US$57 million allocation selects 91 buildings and 292 interventions, with the model estimating a reduction of 1.88 million person-hours of exposure above 35°C.

The tool is designed to help translate heat exposure into investment priorities, identifying where interventions could have the greatest effect within a defined budget. An in-app analyst built on the Claude Agent SDK can rerun the underlying physics and show the reasoning behind its recommendations.

For cities developing heat action plans or directing limited adaptation funding, the prototype illustrates how heat data, building information and cost analysis could be brought together to support more targeted decisions.

Agentic-Arbiter cuts data-centre cooling when outside air can do the work

The second-place project addresses another growing urban infrastructure challenge: data-centre cooling.

Agentic-Arbiter, developed by Bismah Javed, uses FortyGuard's two-metre air-temperature forecast to generate hour-by-hour schedules identifying when a data centre could safely use outside air instead of relying entirely on mechanical cooling.

The system adjusts its safety margin based on previous forecast errors and accounts for site geometry. Where conditions make free cooling unsuitable, it declines the hour rather than recommend an unsafe operating window. The tool was tested across 238 real data-centre campuses in 36 US states and reports its own gaps in coverage.

As cities host growing digital infrastructure, data centres can add significant electricity and water demand. Reducing unnecessary mechanical cooling could help ease grid pressure during periods of high demand while improving infrastructure efficiency.

FrostLine extends temperature intelligence beyond urban heat

The third-place project looks at another dimension of temperature risk.

FrostLine, developed by Darin Levesque, uses multi-year temperature data to map frost risk in vineyards and support decisions about the placement of wind machines used to protect crops during vulnerable growth stages.

The tool estimates the area that could be protected and potential crop savings against installation costs, with its assumptions made visible. When its model ran warm against an actual April 2026 frost event in Virginia, the team developed an independent cross-check rather than conceal the discrepancy.

Its connection to urban heat is indirect, but the project demonstrates the wider potential of temperature intelligence across climate-sensitive systems, including agriculture and food supply.

From temperature data to local action

The three winning projects address different problems but follow a similar logic: combine granular temperature information with the physical and economic characteristics of a particular system, and turn that information into something decision-makers can act on.

That transition, from understanding temperature risk to acting on it, is where the gap lies for many cities today. Heat action plans exist and temperature data is increasingly available. What is often missing is the layer in between, with tools specific enough to tell a planner which building to prioritise, an operator when to switch systems, or a procurement officer where to direct limited funding.

It is also the transition at the heart of Beat the Heat, the flagship implementation drive of the COP30 Presidency and the UNEP Cool Coalition. Launched in 2025, Beat the Heat works to accelerate local action on extreme heat and sustainable cooling and to translate the Global Cooling Pledge into implementation at city and regional level. Its areas of action include urban heat assessments and planning, nature-based and passive cooling, sustainable procurement and the integration of cooling measures into building codes and urban policy. Tools such as those developed through the hackathon could help strengthen that ecosystem by giving cities more precise ways to identify risks, prioritise interventions and direct investment.

"Every building, farm and city should be able to anticipate temperature and act on it," says Jay Sadiq, Founder and CEO of FortyGuard. "This hackathon showed how many different things people can build with the same intelligence. Our ambition is to make temperature part of how the world designs infrastructure, operates machines and invests in its future."