We are building a hybrid AI + physics platform that predicts localized weather at the 10–100 metre scale, providing actionable insights and enabling industries like agriculture and aerospace to make better operational decisions.

Micro-scale weather prediction

Weather forecasting is highly advanced at regional and global scales. However microclimates vary dramatically over short distances due to terrain, vegetation, soil moisture, and wind behaviour. A forecast that is correct at the regional level is often wrong at the paddock, vineyard block, power station or launch site — exactly where real economic decisions are made.

In simple terms: we aim to predict what the weather is actually going to do at this location, not the average of what might happen in the region. From there we calculate actionable insights for operational decisions. Our vision is to create a foundational environmental simulation platform that can ingest domain-specific datasets across industries requiring precise micro-scale forecasts — agriculture first, then aerospace and infrastructure.

Farmers, aerospace operators, and infrastructure managers frequently make high-stakes decisions based on forecasts never designed for site-level accuracy. This gap between regional weather prediction and operational decision-making creates billions of dollars of avoidable inefficiency, risk, and missed opportunities. And with climate variability placing mounting pressure on all industries, the gap between “good enough for the evening news” and “good enough for operational decisions” represents a structural opportunity.

technical details

Micro-scale environmental prediction engine

A hyper-local simulation platform that combines physics-based modelling with machine learning acceleration.

Built GPU-native from the beginning. The physics layer models how weather interacts with terrain, vegetation, and surface conditions at 10–100 metre resolution, while the AI layer accelerates simulations and learns local corrections over time. This hybrid approach lets us predict microclimate behaviour that regional forecasts miss.

Competitors in this field are using AI-only models. But AI models struggle with rare events, don’t enforce the laws of physics and rely on decades of training data that simply doesn’t exist for most locations at sub 100metre resolution.

What makes our system different - hybrid physics and AI, GPU-native design and domain data integration - isn’t theoretical. It’s the same approach that made our previous work industry-leading.

The global investment signals are clear: this is foundational infrastructure with global economic impact - and we have the rare team to build it.

WHO WE ARE

Dr Millie Maier

PhD Astrophysics, Oxford

Millie did her PhD in Astrophysics at Oxford, worked around the world in Astronomy, then moved into leading teams at places like Atlassian. For the last 6 years, she was Head of the Research Lab at Wētā FX, leading teams developing advanced simulation and AI technologies used in major films including Avatar: The Way of Water and Avatar: Fire & Ash.

Inge Rademeyer

Operations & Partnerships

Inge started at Wētā FX 22 years ago in operations and was a senior member of the Marketing team for over a decade. Most recently the Creative Content Director & Manager, she was instrumental in Wētā’s Oscar, SciTech, BAFTA and Emmy campaigns. Her career includes producing film & commercial projects and serving as Head of Operations for an Ed Tech in Los Angeles.

The Research & Engineering Team

We have assembled a world-class team with specialized expertise in numerical simulation, GPU engineering, and machine learning. Our researchers bring deep and proven expertise in complex data simulation, physically-based modeling, and commercial deployment.

Numerical simulation specialists with extensive experience in real-world physics
GPU engineers capable of running large-scale simulations efficiently
Machine learning/AI researchers experienced in hybrid approaches