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

Treble, a Reykjavík-based sound simulation and synthetic acoustic data company, has raised approximately €15 million ($18 million) in a Series A-2 round.

The round was led by Paladin Capital Group, with participation from existing backers KOMPAS VC, Frumtak Ventures, and the European Innovation Council (EIC) Fund. In September 2024, the company announced its €11 million Series A round, led by KOMPAS VC. The fresh capital brings Treble’s total funding raised to date to €36 million.

“Paladin predicts that robotics will drive a profound wave of automation across the physical world. Yet today’s robots are largely limited to visual perception: they may detect what is in front of them but cannot interpret critical sounds – such as someone falling in another room or a crash around a corner. Even accurately understanding human speech remains difficult in real-world, acoustically complex environments,” said François Ruether, Vice President at Paladin Capital Group.

Founded in early 2020 by acoustic engineers Dr. Finnur Pind and Jesper Pedersen, Treble has built a cloud-based platform that models how sound behaves physically in real-world spaces, then uses that physics-based simulation to generate synthetic audio datasets and digital twins of rooms, vehicles, devices, and acoustic environments.

This simulation platform, Treble states, is a faster, cheaper alternative to gathering physical audio recordings, letting engineering teams evaluate a product’s audio performance and generate labelled training data before building prototypes.

Treble allows its customers to create an audio digital twin, similar to how asset-intensive, data-driven industries create digital twins of their physical assets. This virtual replica of a room, vehicle cabin, or device is recreated with high physical accuracy, emphasising how sound waves travel, reflect, and decay within it.

This eliminates the need for teams to travel to a field or physical location to collect and analyse similar data. Because the simulations are physics-based rather than approximated, Treble says the resulting datasets hold up to the same engineering scrutiny as real-world measurements.

The company already works with major consumer and industrial technology firms, including Amazon and Logitech. “In many stages of developing Alexa as an AI assistant, we rely on virtual acoustic environments to accelerate product design and model development. Treble provides an accurate and scalable way to model those environments,” said Wontak Kim, Senior Audio Research Manager at Amazon.

Treble says it will use the funding to deepen its push into physical AI and world models, systems designed to perceive and act within physical environments and dependent on vision- and touch-based sensing. 

The company argues that as voice interfaces become dominant across devices such as wearables, humanoid robots, and autonomous vehicles, products will need to be designed to distinguish speech from background noise amid shifting acoustic conditions. The capital will also support Treble’s expansion into the US market.

Dr. Pind, co-founder and CEO of Treble, said, “AI models and devices that perform well in a laboratory can struggle when they encounter reverberation, background noise, or an unfamiliar physical environment. Treble gives development teams at the world’s leading device makers a faster and more scalable way to model those conditions, generate the data they need, and test products to ensure they work reliably in the environments where machines operate, and people interact with them.”

The funding announced today will also support Treble’s expansion into the US market. The Iceland-based startup aims to turn its simulation and synthetic data platform into core infrastructure for the emerging physical AI and robotics industry, helping machines combine visual and auditory perception to operate more safely and effectively alongside people.

The post Reykjavík-based Treble raises nearly €15 million to bring sound and voice to physical AI appeared first on EU-Startups.