We Build the Physics Engine for the Global Economy
Every day, trillion-dollar decisions in aviation, energy, and logistics are made with statistical guesswork. Nubio builds causal world models that simulate how these systems actually work — learning the hidden mechanics that drive markets, operations, and risk. We don’t forecast from patterns. We reason from first principles.
The World Runs on Invisible Causal Wires
A storm in the South China Sea delays container ships. Three weeks later, a factory in Stuttgart halts production. Six weeks later, an airline in Singapore adjusts fuel hedging. These events are causally connected — but no system sees the wires.
Modern AI learns correlations from historical data. It tells you what happened before, and assumes it will happen again. But the physical economy doesn’t repeat — it evolves through causal mechanisms that statistical models cannot represent.
Nubio was founded on a single conviction: the next generation of AI must understand why things happen, not just what happened before. We build world models that simulate the causal physics of complex systems — and then reason forward to engineer better outcomes.
Causal World Models for the Physical Economy
We develop the foundational AI systems that enable industries to move from pattern-matching to causal reasoning. Each research domain feeds into a unified platform that simulates complex systems as they actually work.
Aviation Intelligence
Our first production vertical. Archimedes unifies revenue, network, scheduling, cargo, and operations into a single causal world model — replacing 6+ siloed vendor systems with one reasoning engine.
Energy Systems
Grid balancing, demand response, and renewable intermittency require understanding causal chains across weather, consumption patterns, and market dynamics. Coming 2026.
Supply Chain
Global supply chains are vast causal networks. We model how disruptions propagate through interconnected logistics, manufacturing, and trade systems. In development.
Financial Markets
Causal reasoning for risk modelling, scenario analysis, and portfolio construction. Understanding why markets move, not just predicting that they will. Early research.
The Values Behind Our Platform
Principles that guide how we build, deploy, and advance causal AI for industrial applications.
Causation Over Correlation
We reject blackbox prediction. Every model must explain the causal mechanism behind its output.
Privacy by Design
Federated architectures ensure competitive entities learn shared patterns without exposing proprietary data.
Research-First
Production code grows from peer-reviewed methods. We publish openly and build on proven science.
Speed
Millisecond inference for real-time decisions. Our architecture separates slow learning from fast acting.
Symbolic Reasoning
Neural networks augmented with explicit causal structure. Graphs and equations alongside gradients.
Scale
From single routes to global networks. Our models compose hierarchically across geographies and timescales.
Federated Learning
Distributed training across organisational boundaries. Airlines, grids, and banks collaborate without sharing data.
Human Feedback
Five levels of AI autonomy. Humans set the dial. The system earns trust incrementally.
Built by Domain Experts
Nubio’s team combines deep expertise in causal inference, production ML systems, aviation technology, and quantitative finance. We’ve built pricing engines at Sabre, demand systems at Amadeus, and real-time platforms at Grab — and we know where traditional approaches break.
Aditya Jalisatgi
Founded Nubio to bring causal reasoning to industrial AI. 15+ years building pricing, demand, and optimisation systems across aviation and mobility. Previously Sabre, Amadeus, Grab.
