Startup Mattiq aims to synthesize and analyze over 1 trillion new material combinations by 2024.
Mattiq, a Skokie, Illinois-based startup, is using artificial intelligence (AI) to help energy producers, chemicals manufacturers, and auto manufacturers find new, sustainable chemicals and fuels. The company announced its launch this week with $15 million in seed funding led by Boston venture capital firm Material Impact.
Mattiq, formerly known as Stoicheia, says it pairs AI with nanotechnology to synthesize and analyze novel material combinations in the context of real-world industrial systems. Its “Megalibrary” technology encodes over 200 million nanomaterials with complex compositions, structures, sizes, and morphologies on a two-by-two centimeter chip, and uses the resulting data to train machine learning (ML) algorithms.
“It is critical to understand that AI relies on vast amounts of underlying data to be properly trained. A lack of consistent, high-quality data has historically been one of the major limitations of successfully applying AI in the field of materials science,” Jeff Erhardt, CEO of Mattiq, said to engineering.com.
Erhardt says that Mattiq’s unprecedented dataset allows the company to leverage AI to address three broad classes of problems: designing many-parameter experiments that are likely to have a positive outcome; interpreting the results of those experiments amidst a mountain of data; and synthetic experimentation, or inferring the behavior of new classes of problems from ones already understood.
By 2024, Mattiq expects to have synthesized over a trillion novel combinations. The company says its technique is a huge improvement upon traditional series of experiments which are expensive, slow and inefficient. Mattiq says its technology could help find solutions for several current challenges, including the sustainable production of fuel cells, green hydrogen, hydrocarbon commodities, biomass fuels, and hydrogen peroxide.
The Deep Blue of Materials Engineering
Mattiq’s work shows that AI is changing the role of materials engineers by helping automate mundane, repeatable tasks common to many endeavors. AI also assists with identifying subtle patterns in complex and high-dimensional data that would otherwise be imperceptible to people.
“We are taking an existing process that is fragmented, linear, and slow and creating an integrated, closed loop platform that mimics and miniaturizes real-world systems, digitizes the results, and accelerates the whole process with AI,” Erhardt says. “We are using this capability to tackle some of the most pressing challenges of our era.”
Erhardt adds Mattiq’s work with AI could be comparable to competitive chess after the computer program Deep Blue beat grandmaster Gary Kasparov in 1997.
“That space evolved not into a world of either computers or people, but of a new genre of computers and people collaborating to play at a higher level than either could do alone. We envision the future of materials science playing out in much the same way,” Erhardt says.