Biomatter raises $18 million to speed up AI enzyme design
Lithuanian biotechnology startup Biomatter has raised $18 million to grow its commercial team, expand its US presence and build a design and testing facility in Vilnius. By bringing more laboratory work in-house, the company aims to cut the time between designing an enzyme with AI and getting test results back from weeks to days.
The 2,000-square-metre facility is scheduled to open in 2027. Astanor led the Series A, with existing investors UVC Partners, Inventure and Metaplanet also taking part. Biomatter says it has delivered more than 50 custom enzymes to more than 25 enterprise customers across sectors including diagnostics, agriculture and food production. About half of its customers are based in the US.
How Biomatter designs custom enzymes
Biomatter grew out of work at Vilnius University’s Institute of Biotechnology, where Laurynas Karpus, Vykintas Jauniškis and Irmantas Rokaitis began exploring AI-generated enzymes in 2017. They founded the company in 2018 with Donatas Repečka and Rolandas Meškys. Their early research, published in Nature Machine Intelligence in 2021, tested whether a generative model called ProteinGAN could produce variants of malate dehydrogenase, an enzyme involved in cell metabolism, that retained its catalytic activity. Thirteen of the 55 designs the researchers tested were soluble and showed the expected activity under the conditions of the experiment.
That work demonstrated that a computer-generated design could become a functioning enzyme. For an industrial customer, the next question is whether the enzyme can do a specific job under the conditions of a production process. Enzymes speed up chemical reactions, but one that performs well in the lab may lose activity at the temperature a factory uses or work too slowly for the time available on a production line.
An established way to solve that problem is to adapt an enzyme that already performs a similar reaction. Biomatter offers that service through LYNX Enzyme Optimization. Its Intelligent Architecture platform can also start with the reaction the customer needs and use generative AI and physics-based modelling to propose candidates. In this kind of design, the production requirements are part of the brief from the outset. Biomatter says it builds a proprietary dataset from experiments on synthetic proteins, so the measurements from one project can inform the selection of candidates in future work.
What companies use custom enzymes for
Inventure and UVC Partners co-led Biomatter’s €6.5 million seed round in 2024. The following year, the European Patent Office named Karpus, Jauniškis and Rokaitis among the ten finalists for its Young Inventors Prize. The office highlighted a project with Japan’s Kirin to develop a scalable way to make human milk oligosaccharides, complex sugars naturally present in human milk. Here, the enzyme is a tool for making the ingredient, and the production process is what the two companies set out to develop.
Neogen, a US food-safety company, confirmed in August 2025 that it was working with Biomatter to develop enzyme-based products, combining Biomatter’s design capabilities with Neogen’s expertise in analytical testing. The following January, Biomatter joined AstraZeneca’s BioVentureHub, an industrial research community in Gothenburg, Sweden. AstraZeneca confirmed the membership in June, linking it to an agreement between the hub and Innovation Agency Lithuania.
The Series A announcement also names a product already on sale. Biomatter says it contributed to Invitrogen SuperScript V, a reagent for RNA research that Thermo Fisher Scientific lists in its catalogue.
How the Vilnius lab could speed up enzyme testing
In a design-and-test round, the team selects candidates, has them made and tests how they perform. The results determine whether a candidate merits further work. If an enzyme shows enough activity but loses it too quickly, for example, the team can use that result to decide whether to refine the candidate or try another. An entire customer project may require several such rounds and further testing in the customer’s own process.
The new facility is intended to give Biomatter more control over when candidates are made and tested and how quickly the results reach the design team. Chief executive Laurynas Karpus told Lithuanian public broadcaster LRT that the centre is a step towards a more autonomous laboratory. Biomatter has yet to disclose how much time it expects to save at each stage. Scheduling lab work alongside design could shorten the wait between stages, for instance by letting the lab plan its next batch around when the design team expects to select candidates.
The investment in Vilnius is part of a wider shift in protein engineering. In an interview with MTN, AI Proteins founder Chris Bahl described how his team starts with what a medicine needs to do and works backwards to design a miniprotein from scratch. Dutch–Swiss startup Cradle, also covered by MTN, puts AI design tools in the hands of researchers who run their own protein-engineering programmes. Biomatter brings the same emphasis on designing for a specific function to the enzymes manufacturers commission.
For those manufacturers, the value of AI design comes down to whether an enzyme can do its job under the conditions of a production line. The new Vilnius lab is where Biomatter plans to get to that answer faster.
Liked this article? You can support our independent journalism via our page on Buy Me a Coffee. It helps keep MoveTheNeedle.news focused on depth, not clicks.