Berlin's Deepslate raises EUR 7.7M seed for European voice AI
Tags AI / ML · Startups
Berlin-based voice AI company Deepslate raised EUR 7.7 million in a seed round led by Munich's 42CAP, with Alstin Capital, existing investor SIVentures and business angels participating. Deepslate develops speech-to-speech models that process spoken language directly rather than converting speech to text first, focused on European languages and hosted entirely within the EU. The company says its model recorded a 440-millisecond response time in an independent Artificial Analysis benchmark, the fastest speech-to-speech model measured as of September 2026, and achieved the best error rate for European languages in the CoVoST2 benchmark. The funding will expand European training data, sales and infrastructure in European data centers.
Technical significance
Speech-to-speech models avoid the latency and information loss of a text intermediate step, and a 440-millisecond response time is close to conversational turn-taking. EU-only hosting and European-language focus position the company for buyers with data-residency requirements.