RESONIKS, the AI acoustic testing startup headquartered in The Hague, today launched the MPK801 – an Ethernet-based industrial measurement microphone built for industrial AI pipelines. The MPK801 is priced at €1,170 per unit. It captures 20 Hz to 80 kHz bandwidth and data integrity monitoring. Multiple microphones are synchronized to sub-100 nanoseconds via the ARK141 Sensor Hub.
Machinists hear a worn tool before any gauge shows it. Maintenance crews hear a bearing begin to fail weeks before it stops. Humans have always listened to what cannot be seen. The machines around them mostly cannot. “Factories are full of cameras and yet are still almost entirely deaf. The microphones available to them were designed for recording studios, not production lines,” says Fabian Oberndorfer, CTO and co-founder of RESONIKS. “Customers have been asking to buy our microphones on their own since we first built them for QCFlex. The MPK801 gives any team the same calibration and synchronization we developed for our own use.”
According to Fortune Business Insights, the predictive maintenance market is expected to grow from roughly USD 11 billion in 2024 to more than USD 70 billion by 2032. According to Mordor Intelligence, ultrasound is the fastest-growing sensing technique inside it. Yet most of the AI behind these projects fails on data, not on algorithms. A model trained on one microphone in the lab does not behave the same once deployed on different hardware in the field.
The MPK801 digitizes sound inside an IP54-rated housing built to sit on a robot arm or machine frame and sends it over a single screw-lock M12 cable – available in drag-chain-compatible variants – to a standard Ethernet network. No preamplifier, no external converter. Units are calibrated to within ±0.5 dB of each other, and calibration data lives on each microphone, so a replacement can be hot swapped in with no recalibration and no model retraining. The ARK141 Sensor Hub synchronizes up to 14 microphones to within 100 nanoseconds and monitors the health of every channel. Data arrives in Python, MATLAB or any Linux toolchain; Windows support is available by the end of 2026.
Applications for the MPK801 include hearing cracks forming inside a part as it cools after welding or additive manufacturing, before they are visible on the surface; monitoring tool and drill-bit wear in machining centers; and locating gas and compressed-air leaks.
RESONIKS’ products are available via its online shop: www.resoniks.com/mpk801, shipping Europe-wide. Customers not based in Europe are encouraged to reach out for an individual quote. RESONIKS is onboarding distribution partners in Germany, France and Spain ahead of 2027.
“The market for acoustic AI is growing faster than the sensors that feed it,” says Felix Wassmann, CEO and co-founder of RESONIKS. “We benchmarked 27 commercial microphones before we built our own, and the trade-off was always the same: pay thousands per channel for bandwidth and synchronization or give up the bandwidth and hope your model does not need it. With the MPK801 that trade-off is gone. More than 20 companies and research institutes approached us before this launch to explore taking acoustic research into production.”
About RESONIKS
Founded in 2022 and headquartered in The Hague, Netherlands with a Testing Centre in Tampere, Finland, RESONIKS is bringing Acoustic AI to the factory floor. Winner of the VDMA Startup Award 2025, RESONIKS has pioneered a solution that detects anomalies in test parts by activating, capturing, and analyzing their acoustic fingerprint, combining a patented sensor technology with continuously evolving machine learning in a production-ready system. Because a suitable AI-native Precision Measurement Microphone did not exist, RESONIKS built its own: plug into Ethernet and get calibrated reliable 20 Hz to 80 kHz data.