The archive · Climate & Energy · Technical decision · 2024–2026
Entalpic bets generative AI can redesign industrial catalysts; €8.5M seed
Paris 2024 startup inverts materials discovery — design the catalyst for ammonia first, validate later; €8.5M seed from Breega, Cathay and Felicis.
Entalpic
What the business is
A Paris-based AI lab building a generative platform that designs and tests catalysts for industrial chemistry, starting with ammonia production for fertilisers.
Starting capital：€8.5M seed (Sep 2024, led by Breega with Cathay Innovation and Felicis)
How it started
Founded in Paris in 2024 by Mathieu Galtier (CEO, former Chief Data & Platform Officer at Owkin) and Mila researchers Victor Schmidt (CTO) and Alexandre Duval (CSO), Entalpic applies generative AI, graph neural networks and LLMs to materials discovery for the carbon-intensive chemicals industry, with operations in France, Germany and Canada.
What happened
In September 2024 Entalpic raised an €8.5M seed round co-led by Breega, Cathay Innovation and Felicis. The platform generates candidate catalysts, evaluates them with simulation methods such as DFT, and tests hypotheses through automated experiments; the first target is ammonia production, a process behind more than 1% of global CO2 emissions.
No ending yet — it is still running.
Background
Entalpic's bet is that the same generative-AI playbook that accelerated drug discovery can be applied to industrial chemistry: instead of starting from known crystals and testing forward, design the catalyst from the desired reaction and validate later. Founded in Paris in 2024 by Mathieu Galtier (CEO, ex-Owkin), Victor Schmidt (CTO) and Alexandre Duval (CSO) — two of them researchers from Mila — the startup targets catalysts for carbon-intensive processes, starting with ammonia production for fertilisers, a market its investor Cathay Innovation pegs at $100 billion.
In September 2024 the company closed an €8.5 million seed round co-led by Breega, Cathay Innovation and Felicis. Its platform combines large language models, graph neural networks and active learning to generate and evaluate new materials, then runs automated experiments, blending open research with proprietary datasets — the Owkin-style data-partnership model Galtier helped build in drug discovery.
Entalpic operates in France, Germany and Canada and is expanding toward the US and Asia. Its first wedge — catalysts for ammonia, a century-old process responsible for more than 1% of global CO2 emissions — gives it a concrete, high-impact target while the platform generalises to polymers, semiconductors, hydrogen and carbon capture.
What has to be true
- Materials discovery has been slow trial-and-error; generative AI can invert it by starting from desired properties and working backward.
- Ammonia production runs on a century-old process responsible for more than 1% of global CO2 emissions — a big, measurable target.
- The founders brought the drug-discovery playbook: data partnerships, proprietary datasets and open research, proven at Owkin and Mila.
- Catalysts underpin most carbon-intensive industries, so a platform that improves them compounds across chemicals, energy and transport.
What can be applied
Entalpic imported the AI-for-science playbook from drug discovery — data partnerships, proprietary datasets, open research — into chemistry; the moat is the data edge, not the model alone.
Aftermath
As of September 2026, Entalpic is a seed-stage lab still building: after the September 2024 raise it is assembling its scientific team and experimental lab, working with industrial partners in France, Germany and Canada, with expansion plans toward the US and Asia. Its platform designs catalysts for ammonia and other carbon-intensive processes; no further funding round has been announced.
Sources
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card