ElevenLabs has agreed to acquire Kotoba Technologies, the Japanese voice AI lab behind the Koto speech model, in a deal reported by Nikkei Asia at more than $100 million. The agreement, announced on November 16, 2026, is ElevenLabs’ first acquisition in Japan.
Kotoba was founded in 2023 by researchers Noriyuki Kojima and Jungo Kasai. Its Koto model covers Japanese text-to-speech, speech-to-text and speech-to-speech conversion for simultaneous interpretation, and its 15-person team will join ElevenLabs. The deal is expected to close later in the fourth quarter of 2026, subject to regulatory approvals.
The acquisition gives ElevenLabs a specialist team for one of the hardest problems in voice AI: making agents sound natural in Japanese. The language often omits subjects, changes vocabulary with the relationship between speakers, and treats pauses as meaningful — all of which a voice agent must handle to hold a convincing conversation.
Koto already has an enterprise foothold. It powers Japanese-language conversations in Salesforce’s Agentforce Voice, and Kotoba’s technology has supported artists including YOASOBI. ElevenLabs says it will combine Kotoba’s East Asian language expertise with its own research and computing resources, starting with Japanese and then extending to Korean and other languages.
The deal underlines how competitive voice AI has become. ElevenLabs competes with speech specialists such as Deepgram and AssemblyAI, while Descript serves creators who need voice tools for editing and production. For ElevenLabs, owning Japanese speech research outright is a faster route into Asia’s voice-agent market than building it from scratch.
Source: Unite.AI