Traditionally, research and development of new materials has been a slow and challenging process. "Historically, over the last 50 years, research and development to find new materials has progressed at a very slow pace," Thibaud Martin, co-founder and CEO of Altrove, told TechCrunch. One of the biggest challenges has been predicting the existence of materials made from various elements. Combining two different chemical elements offers tens of thousands of possibilities, three elements result in hundreds of thousands of combinations, and four elements generate millions of possibilities.
Teams from companies like DeepMind, Microsoft, Meta, and Orbital Materials have developed artificial intelligence models to overcome computational limitations and predict new materials in a stable state. "More stable materials have been predicted in the last nine months than in the previous 49 years," Martin noted.
A Complex Process
Altrove does not invent new materials from scratch. Instead, it selects promising candidates from the newly predicted materials and uses its own AI models to generate possible recipes. Currently, the company tests these recipes one by one and produces a small sample of each material. Altrove has developed a proprietary characterization technology that uses an X-ray diffractometer to check if the material functions as expected.
"It sounds trivial, but it is actually very complicated to check what you have made and understand why. In most cases, what you have made is not exactly what you were looking for in the first place," Martin commented. Altrove's co-founder and CTO, Joonatan Laulainen, is an expert in characterization with a Ph.D. in materials science, and the startup owns the intellectual property related to characterization.
Learning from the characterization process is key to improving recipes for new materials. Therefore, Altrove seeks to automate its laboratory to test more recipes simultaneously and accelerate the feedback cycle. "We want to build the first high-throughput methodology. Pure prediction only takes you 30% of the way to having a material that can be used industrially. The other 70% involves iterating in real life. That's why having an automated lab is so important, because you increase throughput and can parallelize more experiments," Martin stated.
Altrove defines itself as an AI-enabled hardware company. It plans to sell licenses for its newly produced materials or manufacture them in collaboration with external partners. The recent €3.7 million funding round was led by Contrarian Ventures with participation from Emblem and several business angels, including Thomas Clozel (CEO of Owkin), Julien Chaumond (CTO of Hugging Face), and Nikolaj Deichmann (founder of 3Shape).
The startup draws inspiration from biotech companies that use AI to discover new drugs and treatments, applying this approach to the development of new materials. Altrove plans to build its automated laboratory by the end of the year and sell its first asset within 18 months.
