Dutch AI Venture PolyScout Accelerates Plastic Materials Design

2026-10-21 data

The Hague, Wednesday, 21 October 2026.
Spun out from Dutch research institute TNO in October 2026, AI startup polySCOUT uses machine learning to design custom, sustainable polymers in weeks rather than years.

Flipping the R&D Paradigm Under Regulatory Deadlines

Conventional material discovery in the chemical sector has historically suffered from low efficiency, with traditional trial-and-error development yielding success in only 5% to 10% of cases [1][2][3][5]. The core driver of this low yield is the daunting mathematical complexity of polymer science; the search space encompasses more than 10^60 potential polymer structures [1][2][3][5]. To address this challenge, the Netherlands Organisation for Applied Scientific Research (TNO)—an independent research institution in the Netherlands founded in 1932—has spun off polySCOUT B.V., an independent Dutch artificial intelligence company based in the Netherlands [2][3][5]. Formally established in August 2026 and officially launched on October 9, 2026, with a gong ceremony at Euronext Amsterdam, polySCOUT operates via an exclusive technology license while TNO maintains a shareholder position [1][2][3][5][8].

Addressing Regulatory Pressure

The commercial launch comes at a critical juncture for chemical manufacturers facing tightening environmental mandates across Europe. Global plastic production stood at 431 million tonnes in 2024, but regulatory frameworks are forcing material substitutions on tight schedules [2][3]. The European Union’s ban on bisphenol A in food contact materials took effect on 20 July 2026, alongside ongoing restrictions phasing out synthetic polymer particles by 2035 and pending rules targeting per- and polyfluoroalkyl substances (PFAS) [1][2][3][7]. As polySCOUT chief executive officer Michaël van der Jagt observed, forced substitution represents a performance problem accompanied by strict deadlines [1][2][3][7]. Instead of browsing catalogued materials and adapting product designs, polySCOUT reverses the process by using target physical properties and regulatory constraints as inputs to generate bespoke polymer candidates [2][3][7].

Polymer-Native AI Architecture and High-Throughput Validation

Engineered over four years as a dedicated research program within TNO beginning in 2022, polySCOUT’s technology differentiates itself from general-purpose AI platforms by adopting a ‘polymer-native’ approach [2][3][5][6]. According to co-founder and chief scientific officer Milad Golkaram—a University of Groningen PhD graduate and former TNO program lead—polymers are defined not as single molecules but as structural distributions comprising chain length, sequence, branching, and crystallinity [1][2][5]. General material models designed for crystals or small molecules fail to capture these distributional traits [1][2][3][5].

Accelerating Timelines from Years to Weeks

To turn algorithmic predictions into verified materials, polySCOUT’s machine learning models evaluate tens of thousands of copolymer combinations made exclusively from building blocks registered under the EU’s REACH regulation [3][5][7]. The software weighs key criteria including material safety, manufacturing cost, and end-of-life processing [7]. These predictions are fed directly into polySCOUT’s high-throughput laboratory, which synthesizes and tests 25 physical candidate materials per week [1][2][3][5]. By bridging computational modeling with rapid physical validation, the company compresses the traditional design and selection phase of new polymers from years down to weeks [1][2][3][5].

Commercial Expansion and Industrial Strategy

Led by CEO Michaël van der Jagt, CSO Milad Golkaram, and chief commercial officer Ellen de Ruiter—a chemical engineer and former McKinsey consultant who joined the project at TNO in 2023—polySCOUT is currently in active discussions with material producers and product manufacturers for initial design projects [1][2][5]. To support platform scaling and workforce expansion, the startup is initiating its first external financing round in autumn 2026 [1][2][3][5]. Commenting on the broader economic implications, TNO chief executive officer Tjark Tjin-A-Tsoi highlighted that the advanced materials intended to power European industry in 2050 are being designed today, making sovereign design capability pivotal to the continent’s industrial future [2][3].

Bronnen

Artificial Intelligence Polymer Design