Smallest.ai raises $13M Series A for ultra-low-latency voice AI

Smallest.ai raised a $13 million Series A led by Seligman Ventures, with participation from Sierra Ventures and 3one4 Capital, bringing total funding above $21 million. Founded in late 2024, the company builds specialised voice models for real-time conversational agents rather than relying on large language models.

The company says its model processes speech as it arrives instead of waiting for a complete audio input, producing what it describes as virtually zero response lag. Named customers include RingCentral and Truecaller. It competes with ElevenLabs and Cartesia.

Why it matters for digital humans: Conversational latency is the single most common reason a real-time digital human stops feeling live. Investment flowing specifically to the latency problem, rather than to voice quality, is a signal about where the perceived bottleneck now sits.