Berlin-based voice AI company Deepslate has raised €7.7 million in a seed round led by Munich-based technology investor 42CAP, with participation from Alstin Capital, existing investor SIVentures and several business angels.
Deepslate develops its own speech-to-speech AI models, which process spoken language directly rather than first converting speech into text. The company focuses on European languages and hosts its technology entirely within the EU.
Conventional voice AI typically converts speech into text, processes it through a language model and then converts the response back into speech. Deepslate instead uses an end-to-end model in which audio is processed directly and returned as audio. The company says this approach reduces latency while retaining information such as tone, emphasis, dialect and other aspects of spoken language.
Its system consists of three components trained in-house: a speech encoder, a reasoning core based on an open-weights language model that Deepslate post-trains for specific languages, and a speech decoder that generates the spoken response. The architecture allows the underlying language model to be replaced without retraining the entire system.
Deepslate says its model recorded a response time of 440 milliseconds in an independent Artificial Analysis benchmark, making it the fastest speech-to-speech model measured by the benchmark as of September 2026. The company also says its technology achieved the best error rate in the comparison field for European languages in the CoVoST2 benchmark. The company is ISO 27001 certified.
The technology is already being used in production by insurers, contact centres and platforms. Deepslate offers the model through a self-service platform and API, while volume and self-hosting options are available for platform providers and enterprise customers.
For our customers, data sovereignty is not a nice-to-have, it is a prerequisite. And either it can be verified or it is worthless. That is why we disclose where the computing happens, who our subprocessors are and what is still open,
said Paskal Paesler, co-founder of Deepslate.
With the new funding, Deepslate plans to expand its model training and European data programme, with a particular focus on German street names, personal names and dialects, while working to reduce latency and improve voice quality.
The company will also grow its sales and marketing teams and
scale its production infrastructure in European data centres to support
increasing call volumes.