Treble Technologies has raised $18 million in Series A-2 financing to expand its cloud platform for acoustic simulation, synthetic audio data generation, and digital twins. The round, led by Paladin Capital Group, targets US growth and a push into physical AI - the category of intelligent products that must interpret and generate sound reliably outside controlled lab conditions.
Who invested and where the money goes
Paladin Capital Group led the round, with participation from existing backers KOMPAS VC, Frumtak Ventures, and the European Innovation Council Fund. The Reykjavík-based company has now raised approximately €36 million in total. Treble said it will direct the new capital toward building its US presence, with a West Coast expansion already underway under newly appointed US General Manager Vineet Ganju. Ganju previously held executive roles at Synaptics, Conexant, Texas Instruments, and Silicon Labs, with a focus on voice, audio, and wireless product strategy.
Why acoustic simulation matters for voice-enabled products
Voice-controlled devices, wearables, robots, and smart-home products face acoustic conditions that lab testing rarely captures. Reverberation, competing talkers, microphone placement, device orientation, and user movement all degrade performance. Treble's platform lets development teams model those variables digitally, creating labeled acoustic scenes with defined source, receiver, geometry, and material parameters. The output includes device-specific impulse responses and spatial audio data that can feed directly into machine-learning workflows.
"Audio and voice interaction is becoming a fundamental interface for the next generation of intelligent products, but the physical world is acoustically complex," said Finnur Pind, co-founder and CEO of Treble. "AI models and devices that perform well in a laboratory can struggle when they encounter reverberation, background noise, or an unfamiliar physical environment."
Hybrid simulation and the Treble SDK
The platform combines wave-based and geometrical-acoustics simulation methods. The wave-based component captures diffraction, interference, and modal behavior that simplified models miss, while geometrical methods handle higher-frequency propagation efficiently in large or complex spaces. Developers access these capabilities through the Treble SDK, a Python-based interface built for scalable simulation runs and integration into R&D pipelines. Recent SDK updates add reusable simulation-data collections and scene-generation tools, letting teams filter results by source-receiver distance, reverberation, clarity, and other acoustic metadata instead of working with isolated room impulse responses.
Early traction with device makers
Treble disclosed collaborations with Amazon and Logitech, whose teams use virtual acoustic environments to assess audio quality, test difficult scenarios, and draw engineering conclusions without physical prototypes or large-scale recording campaigns. The company claims its approach can compress prototype testing and data collection from months to days. That timeline matters for teams working on AI for Product Development, where iteration speed directly affects time-to-market for voice-enabled hardware.
The physical AI angle
Much of the industry conversation around robots and world models has focused on vision and spatial mapping. Treble's argument is that useful machines also need to interpret speech, alarms, environmental sounds, and activity outside their visual field. Acoustics introduces complexity because the signal reaching a microphone array depends on room materials, source position, device movement, and user behavior - all variables that physics-accurate simulation can make available at development scale. This positions acoustic digital twins as a practical data and validation layer for Speech-To-Text systems and other audio AI models that must generalize across real environments.
Why this matters for product development teams
For teams building voice-enabled products, the bottleneck is often data - specifically, realistic, labeled data that captures the acoustic variability of actual use. Treble's platform shifts that work from physical recording sessions to parameterized simulation runs, which can generate consistent datasets, reproduce edge cases, and compare hardware configurations under controlled conditions. The direct implication is faster validation cycles and fewer surprises when a product moves from anechoic-chamber testing to a noisy kitchen, a moving vehicle, or a conference room with glass walls. If your roadmap includes audio AI features, simulation-based testing is becoming a practical alternative to months of field recording.
Your membership also unlocks:
