Treble raises $18M for its voice AI simulation platform
Treble, an Iceland-based startup founded in 2020 by acoustic engineers Finnur Pind and Jesper Pedersen, has raised $18 million in an extension of its Series A round. The round was led by Paladin Capital Group, with participation from existing investors KOMPAS VC, Frumtak Ventures, EIC, and Omega ehf. Treble received a $12 million investment in 2024, and the new round brings its total raised to date to over $40 million.
Treble builds a simulation platform aimed at model makers, robotics companies, and consumer hardware makers who need to test and improve voice and sound-based AI. For voice AI companies, it runs a synthetic data generation platform used for speech enhancement, noise suppression, and model training, and it evaluates voice AI models under different conditions to give labs feedback. Earlier this year it partnered with Hugging Face to launch a benchmark for speech recognition models across realistic conditions. The startup counts Amazon and Logitech as customers.
Treble also works with headphone and speaker makers on virtual prototyping, letting them hear how a product will sound before it is built, and it can test how well a smart speaker understands commands depending on where the speaker sits in a room. It has recently extended this simulation testing to smart glasses and other AI devices. Founder Finnur Pind said the company wants to work on wearables that enhance hearing, describing devices that could let a user hear only people within a set range in a noisy restaurant or mute everyone but a speaker at a seminar.
Pind said the strategy behind the platform is to replace reliance on recorded and scraped audio data with physics-based simulation: "Audio AI is really a data challenge, and this is where the most opportunities to enable next-generation models and hardware lie. To date, pretty much all sound-related AI has been made from recordings and data scraped from the internet. We believe that accurate physics simulation can be an alternative way to create data for sound." Treble also intends to increase its focus on physical AI, including robotics, automotive, and drone companies, using testing and simulation to enable sound-based functions in those products.
Francois Ruether, VP of Paladin Capital Group, told TechCrunch that Treble's platform stands out because it simulates different models and devices, and that its importance will grow as it is applied to more areas. He framed the investment thesis around infrastructure: as more products depend on understanding sound, a shared simulation-native acoustic infrastructure layer becomes more valuable across voice AI, wearables, robotics, and physical AI, while customers keep ownership of their own models, products, and workflows.
Key facts
- Treble raised $18 million in a Series A extension led by Paladin Capital Group, with KOMPAS VC, Frumtak Ventures, EIC, and Omega ehf also participating.
- The round follows a $12 million investment in 2024 and brings Treble's total funding to date to over $40 million.
- Founded in 2020 by acoustic engineers Finnur Pind and Jesper Pedersen, Treble builds a physics-based simulation platform for testing voice AI models, hardware, and, increasingly, robotics and other physical AI products.
- Amazon and Logitech are named customers, and Treble partnered with Hugging Face earlier this year on a speech-recognition benchmark.
- The company plans to expand further into physical AI, including robotics, automotive, and drone applications.
Why it matters
Voice AI has become one of the most heavily funded areas of AI, covering customer support and sales automation, meeting notetakers, and voice-driven smart glasses. All of that requires testing infrastructure, and Treble is positioning its simulation platform as a shared layer that model makers, robotics companies, and hardware makers can all build on rather than each building their own testing pipeline.
Who it affects
The round involves lead investor Paladin Capital Group and existing backers KOMPAS VC, Frumtak Ventures, EIC, and Omega ehf. On the customer side, Treble names Amazon and Logitech, plus unnamed headphone and speaker makers it works with on virtual prototyping. Its stated expansion targets are robotics, automotive, and drone companies working on sound-based functions.
How to use it
Treble offers a synthetic data generation platform for speech enhancement, noise suppression, and model training, plus evaluation of voice AI models under different conditions. It also does virtual prototyping for hardware makers, testing how a product will sound or how well a smart speaker parses commands based on speaker placement, and it has extended this to smart glasses and AI devices. It also runs a Hugging Face speech-recognition benchmark launched earlier this year. No pricing or commercial terms for any of this are disclosed.
How solid is it
The story is a direct TechCrunch report carrying on-record quotes from co-founder Finnur Pind and Paladin Capital Group VP Francois Ruether, with a named funding history: $12 million in 2024, over $40 million raised in total, and this $18 million extension led by a named investor. The specific investors and prior round give the funding claim solid grounding.
Risks and caveats
No valuation is disclosed for Treble, before or after this round, and no headcount is given. There is no timeline for when the physical AI expansion into robotics, automotive, and drones will materialize, no detail on the terms of the Hugging Face benchmark partnership beyond its existence, and no breakdown of how much each participating investor contributed to the $18 million round.
“Audio AI is really a data challenge, and this is where the most opportunities to enable next-generation models and hardware lie. To date, pretty much all sound-related AI has been made from recordings and data scraped from the internet. We believe that accurate physics simulation can be an alternative way to create data for sound.”
— Finnur Pind, co-founder of Treble