Thesis
DualEnergy is building the foundation model for dual-use land — the strip of earth where solar generation and agriculture have to coexist. Today that interaction is engineered project by project, on local intuition. In a decade it will be the largest constrained optimisation problem in the energy transition.
Two engines
The training data does not exist yet, so we build the apparatus that creates it.
- A software platform every developer is regulated into using. Yield modelling, permitting workflows, financial assumptions — bankable from day one. Each project that runs through it labels a parcel of land with the inputs and outcomes that matter.
- A sensor network deployed on every project we touch. Continuous ground truth on irradiance, soil moisture, micro-climate, crop response, panel performance. Cheap enough to put in the field; instrumented enough to feed a model.
The two engines compound: the software defines the schema, the sensors fill in the values, the model improves on every new project.
Horizon
By 2035, we own the only ground-truth dataset of energy–agriculture interaction on Earth — and the AI that turns it into decisions: where to site, what to plant, how to operate, what to underwrite.
I’m CTO. The job, today, is to make the early data engine tight enough that the model has something to be trained on.