Treble Raises $18 Million to Expand Virtual Sound Testing for Voice AI

Treble has raised $18 million to expand a platform meant to test voice AI and audio devices against difficult sound conditions before they reach the physical world. The company says its simulations can reduce the need for long recording and prototype programs, but the announcement does not provide comparative accuracy results showing how well those virtual tests predict real-world performance.

Paladin Capital Group led the Series A-2 financing. Existing investors KOMPAS VC, Frumtak Ventures and the European Innovation Council Fund participated; TechCrunch described the deal as an extension of Treble’s Series A and also named Omega ehf. among the backers.

A digital version of a listening environment

Treble says this matters when a product leaves a controlled setting. Sound can be changed by room shape, background noise and microphone placement. Moving products add another variable: the company identifies head-worn devices, robots and autonomous vehicles as examples that must work while navigating changing acoustic environments.

Generating test cases instead of collecting them

What Treble says it can model

• Different rooms, materials, speakers, devices and ambient sound conditions.
• Product configurations before physical prototypes are built.
• Acoustic edge cases that are difficult to recreate at scale.

Treble says the approach can cut prototype testing and data-collection work from months to days. That is a company performance claim, not a published comparison across customers or product categories. Whether simulation speeds development without missing important real-world failures is therefore the key practical question behind the funding.

Customers offer a real use case, not a verdict

Treble says the new capital will support U.S. expansion and growth in consumer, enterprise, industrial and physical AI applications. It also plans to extend its simulation and testing work into robotics, automotive and drone companies. Founded in 2020 by acoustic engineers Finnur Pind and Jesper Pedersen, the company is now funding a larger attempt to make machines better listeners before they operate around people.