Deepslate, a Berlin-based voice AI startup, has raised €7.7 million in a seed funding round led by 42CAP. Alstin Capital, existing investor SIVentures and several business angels also participated in the round.
Founded in 2024 by Paskal Paesler and Jan Brachthäuser, Deepslate develops native speech-to-speech AI models for real-time conversational applications. Paesler serves as CEO and Brachthäuser as CTO.
The company trains its own models to process spoken audio directly rather than relying on the conventional sequence of speech-to-text, a text-based language model and text-to-speech systems. Deepslate focuses on European languages and offers EU-hosted deployments as well as self-hosting on customer infrastructure. Its technology is already being used in production by insurers, contact centres and platforms.
Deepslate’s Opal model recorded a time to first audio of 0.44 seconds in Artificial Analysis’ speech-to-speech benchmark. Artificial Analysis currently lists it as the lowest time to first audio among the models tracked in its comparison.
The company plans to use the seed funding to advance its speech-to-speech model research, expand European training data, grow its go-to-market team and scale infrastructure across European data centres.
