This summer has brought another stark reminder of the threat posed by wildfires, with extreme heat and major fires affecting communities across Europe and, increasingly, closer to home in the UK.
For co-founder and CEO Dr Savvas Gkantonas, the subject is both scientific and personal.
In 2018, while completing his PhD at the University of Cambridge, a wildfire in Mati, Greece, destroyed his family’s home. At the time, Savvas was a combustion scientist studying the behaviour of flames in engineered systems.
That experience prompted Savvas, his supervisor Prof. Mastorakos, who was also Greek and had grown up in Mati, and Dr Andrea Giusti to start looking at the problem in their own time, spending afternoons and evenings exploring whether their understanding of combustion could be used to predict how wildfires behave.
The research that followed ultimately led to Pinepeak, a University of Cambridge spin-out and winner of the £10,000 Angel Prize at last year’s Trinity Bradfield Prize, held at Cambridge Science Park.
Its FLAMESIGHT technology combines physics-based modelling with AI to help predict where and when fires may start and how they spread through natural landscapes and towards towns, villages and other built-up areas.
Savvas has a simple way of describing it: “Think of it like weather forecasting, but for fires.”
Research into the model began in 2019, with academic papers following in 2022 and 2023. Before Pinepeak was spun out in 2023, Dr Daniel Fredrich joined the team and is now co-founder and COO. The company secured investment in late 2024, allowing the founders to build a multidisciplinary team around the technology.
Today, Pinepeak’s primary commercial focus is the insurance industry. It works with partners, including the London market, to help them better understand and quantify wildfire risk.
Much of that work is centred on the US and California, where wildfire risk is high and there is a significant concentration of valuable property exposed to it.
Pinepeak’s models look at where fires could start, how they might spread and the probability of individual properties or other assets being damaged. The resulting data can help insurers assess risk.
The company can also use data from existing satellites to locate fires as they happen and model where they may travel next.
There could be wider applications for that technology in the future, including supporting civil protection services and firefighters, and expects to expand into these areas as the technology continues to mature and the right opportunities emerge. But Savvas is careful not to get ahead of where the company is today.
Wildfires are complex and unpredictable, and applying forecasting technology in situations where lives may depend on the information brings a very different level of responsibility.
“When you work in something like wildfires that destroy lives, making claims is hard,” he says.
That caution is reflected in Pinepeak’s approach to growth. Technologies designed to model wildfires have existed for many years, but Savvas says adoption can be difficult depending on the end user and the level of maturity required. Pinepeak made a conscious decision to focus first on the insurance industry, where there is an established need to understand and quantify the risk, as well as demand for more differentiated and granular solutions.
For now, its attention is firmly on delivering for select customers, continuing to develop and test its technology, and building on its growing commercial traction.
As Savvas puts it, the priority is “results, results, results.”