Exclusive: Startup combines physics with AI to discover new green materials ’10x faster’

Exclusive: Startup combines physics with AI to discover new green materials ’10x faster’

AI tools like ChatGPT, Gemini, or DeepSeek have radically disrupted the way we access and generate information. However, these systems are general — they are jacks of all trades but masters of none. 

Increasingly, though, scientists are training AI to solve very specific problems and fast-track everything from creating new drugs to designing fusion reactors. One area where they’re gaining traction is enhanced material discovery. Their advances are creating a growing range of promising startups. PhaseTree is one of them.

Spun-off from the Technical University of Denmark in 2021, PhaseTree has developed a platform that integrates computer simulations, lab automation, and AI to develop new materials for clean technologies like batteries, solar panels, and wind turbines. While it’s not unveiling its secret sauce, the startup said the tech enables it to find new materials 10 times faster than traditional methods. 

“Our approach accelerates discovery by combining physics-based modelling with AI, allowing us to rapidly identify and refine promising candidates that would traditionally take decades to develop,” Jin Hyun Chang, PhaseTree’s co-founder and CTO, told TNW.  

PhaseTree is in an increasingly crowded sector. AI-powered material discovery startups raised over $260mn last year, according to Dealroom data. Ten out of the 17 funding rounds in 2024 went to startups in Europe, including ExoMatter, Dunia, and Orbital. 

PhaseTree promotes a different approach to its rivals. Unlike many other AI-powered material discovery tools, Chang said the startup prioritises using existing scientific methods first, and then applies machine learning in an approach it calls “physics-first, AI-on-top.”  

“Many AI-driven tools rely heavily on data correlations, but we emphasise the physical principles governing materials such as composition, atomic structure, defects, and microstructures, to ensure a reliable prediction of the materials properties,” he explained.

Today, PhaseTree announced that it has secured €3mn in funding from Denmark-based early-stage VC Heartcore Capital. It will use the fresh funds to boost R&D and expand its team, as it looks to deploy its science to solve real-world problems.

Amit Luthra, co-founder and CEO of PhaseTree, emphasised the company’s focus on practical applications.

“From the outset, we design materials with manufacturability in mind, ensuring they can be synthesised at scale rather than remaining a theoretical concept or lab-scale prototypes,” Luthra told TNW. “By prioritising low-cost, abundant materials with straightforward synthesis routes, we maximise the likelihood of adoption by key industrial players.”

PhaseTree said it is currently working with some of the world’s largest battery, automobile, and steel producers, and is already optimising high-performance battery electrodes and advanced alloys. 

“The Next in Tech” is one of three key themes at TNW Conference, which takes place on June 19-20 in Amsterdam. Tickets for the event are now on sale. To get 30% off, use the code TNWXMEDIA2025 at the check-out.

March 7, 2025 at 02:45PM
https://thenextweb.com/news/phastree-combines-physics-ai-to-discover-green-materials?utm_source=flipboard&utm_content=topic%2Fartificialintelligence
Siôn Geschwindt

Lascia un commento

Il tuo indirizzo email non sarà pubblicato. I campi obbligatori sono contrassegnati *