The Rise of Large Quantitative Models (LQMs)
How AI Models That Understand Numbers Are Powering the Next Wave of Innovation"
Welcome to this edition of my ‘Dissected’ dispatch.
If you’ve been following my work, you know I’m endlessly fascinated by science, data, and technology. Years ago, I coined the term "deep technology" to describe innovations built on scientific discovery and engineering breakthroughs. Today, I want to talk about a new and innovative concept I believe will be just as transformative: Large Quantitative Models, or LQMs.
What Are Large Quantitative Models?
We’ve all seen how Large Language Models (LLMs) like GPT have revolutionized communication, coding, education, and more. LQMs do something similar, not with words, but with numbers.
While LLMs treat numbers as symbols in text, LQMs understand numbers as data points with meaning. They ingest vast volumes of structured and unstructured numerical data, learning from complex relationships and patterns that arise across time, space, and scale.
Quantitative modeling is not new. What is new is the breadth and richness of data we now have access to, from real-time sensor data in industrial environments, to molecular simulations in biotech, to massive datasets streaming in from energy systems, financial markets, and climate sensors. LQMs can now analyze, synthesize, and predict from this deluge of data in a way that was simply not possible before.
Why LQMs Matter
LQMs represent a powerful convergence of AI, data science, and domain-specific expertise. They’re not just tools for better analytics, they’re strategic assets.
Unlike LLMs, which are typically trained on public web data, LQMs often require proprietary data: internal enterprise systems, sensor streams, government databases, scientific research, and more. This creates a natural moat for those who build and train them, with strong potential for defensible business models and high-value applications.
Where LQMs Are Already Making an Impact
LQMs are reshaping some of the most complex and critical sectors of our economy. Here are a few areas where their impact is already being felt:
Life Sciences & Healthcare: From accelerating drug discovery to modeling disease progression and genomics, LQMs are helping researchers explore previously intractable biological questions.
Climate & Weather Modeling: Improved accuracy in climate forecasts and extreme weather prediction can save lives and guide more sustainable policies.
Fusion Energy & Physics: Atomic-scale simulations are unlocking new insights into materials science, energy generation, and quantum interactions.
Financial Markets: LQMs can model complex market dynamics in real time, offering deeper insights into risk, opportunity, and macroeconomic shifts.
Industrial Optimization: In manufacturing, energy, and logistics, LQMs power smarter systems that think predictive maintenance, dynamic supply chain management, and grid resilience.
Companies Leading the Charge
A few companies that are breaking ground with LQMs include:
SandboxAQ – I had a great conversation with CEO Jack Hidary at the FII conference. Sandbox AQ was spun out from Alphabet. They’re innovating at the intersection of quantum tech and AI, leveraging LQMs.
Applied Intuition - They build software for autonomous vehicle simulation and testing. (In fact, Autonomous Driving is an excellent example of the marriage of LLMs and LQMs - where LQMs are used for path planning, and controls, while LLMs are used for communication)
Precision Neuroscience - Building brain-computer interfaces. Their potential is to enable disabled patients to operate devices using thoughts alone (similar to Neuralink).
Deep Genomics – Using highly structured genomic and RNA data to train predictive models for drug discovery. Their AI-driven approach is accelerating the development of novel genetic medicines.
The Climate Service (‘TCS’, now S&P) - Featured on Propel(x) in 2019, and acquired by S&P in 2022, the TCS’s Climanomics platform is a climate model used to assess climate risks of various kinds, including in the Insurance industry.
(The Climate Service and HealX (discussed in an earlier newsletter) - are examples of how Propel(x) has been ahead of the curve for years.)
These examples only scratch the surface. We’re at the dawn of a new era where deep numerical understanding will reshape our approach to science, engineering, economics, and health.
Why Now?
The conditions are ripe. Computing is cheaper and more scalable than ever. Specialized hardware for training numerical models continues to improve. And crucially, we now have access to massive volumes of high-fidelity, real-world data.
If LLMs helped us talk to machines and each other, LQMs will help us understand the world.
I believe the next wave of scientific and industrial breakthroughs will be powered by these models. If you’re building in this space, collaborating with researchers, or simply curious about the future of quantitative AI, I’d love to hear from you.