MacroLab AI
Economic analysis with artificial intelligence
Numbers alone miss the story.
Economic outcomes are shaped as much by real-time narrative flows as by hard numerical metrics. Traditional economic tools often isolate high-dimensional contextual data from quantitative time series, relying on simplified assumptions that leave critical insights behind.
MacroLab AI unifies multimodal data streams with machine learning, transforming unstructured news flows into structured, real-time economic intelligence.
Integrating economics and AI.
We build research-backed products that deliver automated narrative attribution, real-time forecasting, and structural risk decomposition.
From tracking broader macroeconomic cycles to monitoring commodity dynamics and market sentiment, our products explain what is moving in the economy, and why.
We are moving toward agentic systems that can run entire research workflows in parallel. That direction is resource-intensive — home-grown deep neural networks, frontier models, tokens, and compute. But because economics is a social science, full automation is not the goal: humans stay in the loop, machines do the scaling.
Our approach combines domain-informed economic principles with modern artificial intelligence, leveraging the information content in multimodal data. Beyond economics, the tools we build are potentially useful in any setting where unstructured data-streams co-evolve with numeric time series.
Driven by curiosity and dedication, we push the research frontier with new methods while remaining firmly grounded in scientific rigor, transparency, and empirical thoroughness. We are committed to translating scientific innovation into clear, relevant insights with real impact.
Live dashboards.
Our research ships as live dashboards, and the work falls into three groups.
Together they cover three questions: where the economy stands right now, read from the news as it breaks; where it is heading, with forecasts and the scenarios that test them; and what the long text record says about sentiment, uncertainty and the themes that drive both.
Explore the projects →MacroLab AI was started by Vegard H. Larsen and Leif A. Thorsrud. Read more about us →
To get where we want to go, we need a larger team and collaborators. So if this kind of work excites you, we want to hear from you. We are especially seeking talented PhDs eager to tackle frontier challenges with scientific rigor, as well as industry partners who share our ambition to build impactful real-world solutions.