Rime Raises $24M to Accelerate Speech AI

The big picture: San Francisco-based Rime has raised $24 million in a Series A led by M13, with participation from Twilio Ventures, Corazon Capital, Cadenza Ventures, and Unusual Ventures. The company is developing speech-to-speech foundation models that make AI conversations sound natural and responsive, with the new funding accelerating research and product development.

Why it matters:

• Natural interaction: Current voice AI often struggles with naturalness, leading to interruptions, unnatural pauses, and misreading emotions, which degrades user experience.
• Trust in regulated sectors: In critical industries like healthcare, banking, and customer support, these shortcomings quickly erode trust, making reliable voice AI essential.
• Linguistic precision: Rime’s approach prioritizes linguistic precision over raw model size, a strategy validated by independent studies showing lower hang-up rates and faster completion times compared to competitors.

How it works:

• Speech-to-speech models: Rime builds real-time speech-to-speech foundation models designed to understand, generate, and respond in natural-sounding conversations.
• Enterprise scale: Their voice models currently power nearly 100 million phone calls monthly for enterprise clients, including Mayo Clinic, Dialpad, Upstart, and Asurion.
• Linguistics-first design: Founded by PhD linguists, Rime integrates linguistic judgment and taste into its research, ensuring speech appropriateness and relevance in complex conversational contexts.

The catch: The voice AI market is highly competitive, with well-funded players like ElevenLabs, Hume AI, Cartesia, and Deepgram also vying for market share. Rime’s focus on the modeling layer and traction in regulated industries like healthcare and financial services (where HIPAA compliance and pronunciation accuracy are critical) differentiates it, but sustained innovation is required to maintain its edge against broad-based and specialized competitors.