IPT Leads R$20 Million MaterI.A. Centre Using AI to Speed Up Advanced Materials
São Paulo’s Instituto de Pesquisas Tecnológicas (IPT) is leading the creation of MaterI.A., a centre that uses artificial intelligence to speed up the discovery, development and validation of advanced materials and industrial processes, the institute announced on October 21. Total investment is about R$20 million, of which more than R$14 million comes from Finep, the federal innovation funding agency.
The centre, supported by the IPT’s foundation FIPT, brings together scientific partners including Poli-USP, the São Carlos Institute of Chemistry, the Hospital das Clínicas of the USP medical school, Unesp, Unifesp, UFABC, Ital, CNPEM and Senai. On the business side, the network is backed by companies such as Vale, WEG, CBMM, Renner Herrmann, Marcopolo and Tupy, among others.
IPT researcher André Nunis said conventional materials research is slow and costly, with cycles that can take decades. MaterI.A. aims to shorten them by combining physics-informed machine learning, prediction of physical properties, digital twins and reinforcement learning with laboratory experiments. Work is organised into four research lines: digital materials discovery, an intelligent testbed, extended applications for new uses, and intelligent processes that optimise production and cut input use. The centre will have a dedicated AI computing cluster and multi-user laboratories.
What it means in practice
The institute expects about 80% of the centre’s operation to serve industrial demand, extending work the IPT already does through its Embrapii materials unit. Target sectors include electric mobility, energy, health, food and high-performance industry. Nunis cited estimates that the global advanced materials market was worth US$513.1 billion in 2023 and could reach US$695.5 billion by 2028, and said the goal is for the IPT, with more than 127 years of applied research and more than 25 years in advanced materials, to become a world reference in AI-assisted materials.
The project adds to a wider push to build AI capacity in Brazilian science, from the LNCC’s planned AI supercomputer in Macaíba to USP’s TITAN-LSTM storm-tracking model. Finep is also behind the BNDES and Finep AI fund run by Monashees.
What to watch: which industrial partners bring the first projects, how quickly the computing cluster and laboratories come online, and whether the centre can show measurable cuts in development time rather than promises.