Reykjavík-based Treble raises nearly €15 million to bring sound and voice to physical AI

Treble, a Reykjavík-based company specializing in sound simulation and synthetic acoustic data, has secured nearly €15 million ($18 million) in a Series A-2 funding round. The investment was spearheaded by Paladin Capital Group, with support from KOMPAS VC, Frumtak Ventures, and the European Innovation Council (EIC) Fund. This brings Treble's cumulative funding raised to €36 million since its inception in 2020.

Paladin Capital Group's Vice President, François Ruether, highlighted the potential of robotics to revolutionize automation in the physical world. However, he noted that current robots primarily rely on visual perception and struggle with interpreting critical sounds in real-world, acoustically complex environments. These challenges include detecting subtle sounds like someone falling or a crash, and accurately understanding human speech.

Founded by acoustic engineers Dr. Finnur Pind and Jesper Pedersen, Treble has developed a cloud-based platform that models sound behavior in real-world spaces. This physics-based simulation enables the generation of synthetic audio datasets and digital twins of various environments, offering a faster, more cost-effective alternative to physical audio recording.

Treble's technology helps engineering teams evaluate audio performance and generate labeled training data before constructing prototypes, eliminating the need for teams to travel to field locations for similar data collection and analysis.

The company's platform allows customers to create audio digital twins, akin to how asset-intensive industries create digital twins of their physical assets. These virtual replicas exhibit high physical accuracy in replicating sound wave travel, reflection, and decay within a space, bypassing the necessity for teams to gather data in physical locations.

Treble currently collaborates with major consumer and industrial technology firms such as Amazon and Logitech. For instance, Amazon's Senior Audio Research Manager, Wontak Kim, expressed reliance on virtual acoustic environments during Alexa's development, crediting Treble for providing an accurate and scalable method to model these environments.

With the latest funding, Treble intends to further expand its focus on physical AI and world models, systems designed to perceive and interact within physical environments, heavily reliant on vision and touch-based sensing. As voice interfaces become increasingly prominent across devices like wearables, humanoid robots, and autonomous vehicles, Treble asserts that products will need to be designed to distinguish speech from background noise in varying acoustic conditions. The capital will also aid in Treble's expansion into the US market.