Thank you so much, Mr. Chair, for the invitation to join you today.
My name is Donald Killorn. I'm the executive director of the P.E.I. Federation of Agriculture. I'm here today to discuss food security in the face of global instability.
As I'm sure you know, 24% of people in Canada live in food-insecure households. That's about 10 million people, with 2.5 million being children. Those percentages are very similar to what we see on Prince Edward Island, which is remarkable since we grow so much food on Prince Edward Island.
Food security is not simply the ability to grow food. It is the outcome of a complex system that includes agricultural production, labour availability, transportation, processing, trade, land use, environmental stewardship and public policy. Disruptions in any of these areas can ripple through the entire food system, affecting affordability, availability and the resilience of that food system.
Our expertise at the P.E.I. Federation of Agriculture is at the grassroots. It's in the farmers' fields. In those fields, across Canada, there is a bottleneck. Information struggles to move between the field and the larger food system. Farmers are wary of how their data is handled. As Canada considers both its AI strategy and its food security strategy, we must take this opportunity in the space where these overlap to build intelligence for farmers that connects the food system to the farmers' field.
At the P.E.I. Federation of Agriculture, we developed a sovereign agricultural compute strategy that will strengthen food security by improving Canada's ability to understand, manage and respond to the complex system that produces, processes, transports and distributes food. Our efforts began by building AgIntel, a precision agriculture platform that runs on artificial intelligence.
We quickly learned that the future of precision agriculture is not the development of artificial intelligence that understands farming. Rather, it's the implementation of a national sovereign compute strategy that reduces the resources required to equip our Canadian farmers with the technology they need to drive improvements to our food security. By keeping agricultural data, models and computational capacity under Canadian control, a Canadian agricultural compute strategy would enable better decisions across the entire food system, from farm-level production to national food security planning. This would transform our sovereign compute from a technology initiative into a foundational piece of our food security infrastructure.
As agriculture becomes increasingly data-driven, decision-making capacity becomes infrastructure. The ability to model risk, optimize production, forecast disruption and support policy decisions depends on access to trusted compute infrastructure and sovereign agricultural data. Farmers demand it.
A Canadian agriculture compute strategy would provide the digital foundation required to strengthen national food security while maintaining ownership and control of agricultural data. Farmers grow the food, but many of the inputs required to support modern Canadian agriculture are sourced from other countries. Agricultural data faces the same challenge. While Canadian farms generate the data, the platforms and models that transform it into actionable intelligence are often foreign-owned. As a result, both the data and the value it creates are processed outside Canada's jurisdiction. This represents a growing sovereignty gap in a sector critical to national food security and economic resilience.
A $150-million investment in sovereign agricultural data infrastructure would turn Canada's farm data into a national strategic asset, strengthening food security, keeping the intelligence layer Canadian and returning multiples of the spend in farm-level value. By adopting a sovereign compute strategy, Canada would lower the electricity required to implement artificial intelligence in agriculture from 200 gigawatt hours to less than two. The system would spend full compute only when a farm record teaches the system something new. It would interpolate the rest.
Expanded nationally, this system would reduce the national compute load from 150 petaflops to 15 teraflops. This sounds like science fiction, but these are quickly becoming critical resources we must use efficiently in order to be competitive. On top of a significant reduction in resources, the system would also reward the farmer when they provide this novel data, giving them ownership over the platform and credit to use on the insights derived from their data.
In the near future, our farmers will utilize robust models to maximize their agronomic and economic performance. They will access most of this modelling locally. However, when something novel takes place, that information—