Treble Raises $18 Million to Expand Voice AI Simulation Platform
Iceland-based Treble has raised an $18 million extension to its Series A round, led by Paladin Capital Group. Existing investors KOMPAS VC, Frumtak Ventures, EIC and Omega ehf also participated. The acoustic simulation startup, founded in 2020 by engineers Finnur Pind and Jesper Pedersen, has now raised more than $40 million in total after receiving $12 million in 2024.
Treble is building a platform for companies developing voice AI, consumer hardware and physical AI systems. Its customers include Amazon and Logitech. The company combines synthetic data generation, model evaluation and acoustic simulation to help teams assess how sound-related products and models perform in varied conditions.
Simulation for voice models and devices
For voice AI developers, Treble offers synthetic data generation that can support speech enhancement, noise suppression and model training. It also evaluates voice models in different acoustic settings, giving developers feedback on performance beyond the recordings and internet-scraped data traditionally used in much sound-related AI work.
Pind described audio AI as a data challenge and said accurate physics simulation can offer an alternative way to create sound data. Treble partnered with Hugging Face earlier this year to launch a benchmark for speech-recognition models across realistic conditions. That work sits alongside a wider voice AI market in which model providers and hardware companies need testing loops to improve products.
From virtual prototyping to physical AI
Treble also applies its technology to hardware design. Headphone and speaker makers can use virtual prototyping to understand how a prospective product may sound, while smart-speaker teams can test how device placement affects command recognition. The company has recently begun simulation testing for smart glasses and other AI devices.
The product direction overlaps with the growing market for speech systems, where Gradium investment in voice-model capabilities reflects investment in voice-model capabilities while Treble concentrates on the testing, data and simulation layer around such products.
Pind highlighted wearable uses such as improving hearing in difficult acoustic environments, including isolating nearby conversation in a restaurant or muting surrounding people at a seminar. Treble also plans to increase its focus on robotics, automotive and drone companies seeking to enable sound-based functions through testing and simulation.
What the funding supports
Paladin Capital Group vice-president Francois Ruether said the platform’s ability to simulate different models and devices makes it relevant across voice AI, wearables, robotics and physical AI. He added that customers retain ownership of their models, products and development workflows while using a shared simulation-native acoustic infrastructure layer.
For businesses building voice-enabled models or hardware, the funding underlines the value of testing sound performance in realistic environments before product decisions are locked into training pipelines, device designs or deployment workflows.