In the race to wire real-time voice intelligence into enterprise software, a relatively young San Francisco outfit called Smallest.ai just landed the kind of validation that should turn a few heads—and perhaps more investor phone calls its way.
The startup secured $13 million in Series A funding led by Seligman Ventures, a round that pushes its total capital raised past $21 million. Not bad for a company that only opened its doors in November 2023. The financing, disclosed at the end of July, arrived just nine months after Smallest.ai closed an $8 million seed—a timeline that suggests the market's appetite for conversational AI infrastructure hasn't exactly cooled.
Seligman Ventures took the lead this time around. Sierra Ventures, which had steered the seed round back in October 2024, returned for the Series A, joined by 3one4 Capital. A handful of other backers—Better Capital, Upsparks Capital, Schema Ventures, Tiny VC, DeVC, and Mission Street Capital among them—also participated, alongside a roster of angel investors. The company announced the close in a blog post the day after the round went public.
Building for milliseconds that matter
Smallest.ai's pitch centers on what it calls "Voice 4.0," a sweeping term for real-time voice AI that's meant to feel less robotic and more responsive than earlier generations of speech technology. The startup's flagship offerings include Lightning, a text-to-speech engine; Pulse, which handles speech-to-text conversion; and Hydra, a full-duplex system designed to process spoken audio in both directions with minimal lag. Low latency, in this context, isn't just a technical nicety—it's the difference between a conversation that flows naturally and one that feels like talking through molasses.
The company also built Electron, an in-house large language model positioned as a drop-in alternative to OpenAI's Chat Completions API, and Atoms, a platform for assembling and deploying voice agents across phone lines and web interfaces. For industries navigating thickets of regulation—healthcare, finance, and the like—Smallest.ai offers on-premises deployment and touts compliance with SOC 2 Type II, ISO 27001, and GDPR standards, though its terms and conditions note that the services aren't specifically tailored for certain regulations like HIPAA. In other words, it's designed to play nice with enterprise security teams, within limits.
According to documentation the company updated between May and July 2026, Pulse notched a 5.42% word error rate on the Open ASR Leaderboard, tying for second place in English-language accuracy. Benchmarks like these matter in an increasingly crowded field where every basis point of precision can sway a procurement committee.
Early traction, familiar names
Smallest.ai lists RingCentral, Truecaller, Readymode, Piramal, Kogta, and Pocket among its customers, though the company hasn't disclosed contract sizes or how deeply its technology is embedded in these organizations. Still, the presence of recognizable brands—RingCentral especially—lends some credibility to the startup's enterprise aspirations.
The company was co-founded by CEO Sudarshan Kamath and Akshat Mandloi. By the time the Series A closed, the team had grown to nearly 60 employees, a headcount that reflects both ambition and the challenges of scaling technical infrastructure in a market where talent is expensive and competition for engineers fierce.
What the capital buys
The funding announcement made clear that Smallest.ai plans to double down on Hydra and the broader Voice 4.0 vision. Translation: expect continued investment in refining its real-time speech-to-speech capabilities and expanding the toolkit available to enterprises building voice-first applications. Whether that's enough to carve out sustainable differentiation in a space where giants like Google, Amazon, and OpenAI also compete remains an open question.
For now, at least, Smallest.ai has bought itself more runway—and perhaps a better shot at answering it.
