Add Product 
Bengaluru- and San Francisco-based Maya Research has raised $1.9 million in a seed round to build voice AI that speaks, thinks, and responds like a native speaker — a bet on the billions of people who interact with technology through conversation rather than text. The round was led by South Park Commons.
Inside the $1.9 Million Round Led by South Park Commons
The $1.9 million seed funding led by South Park Commons will help Maya train larger conversational models, expand its deployment infrastructure, and study how users in voice-first markets actually talk to machines. Founded in 2025 by New York University graduates Dheemanth Reddy and Bharath Kumar Kakumani, the startup frames voice not as a feature bolted onto a chatbot, but as the primary interface for the next wave of internet users.
“The internet was built around English and text, which quietly left most of the world outside the interface,” said Prateek Mehta, general partner at South Park Commons. “As voice becomes the way billions of people interact with technology, the company with the richest multilingual speech data and the strongest conversational models will have a defining advantage.”
Aditya Agarwal, also a general partner at the firm, put it more bluntly: “Maya is India’s ChatGPT moment.” The deal lands amid a busy stretch for domestic AI financing, from Sarvam’s unicorn round to Equal’s Series B.
Why Maya Research Is Building Voice AI for the Next Billion Users
For most companies, voice AI is a layer: a user speaks, a language model writes a reply, and a text-to-speech engine reads it aloud. Maya argues that pipeline misses the point. Its conversational models are designed to handle the hesitation, affirmation, and emotion that real speech carries — deciding not just how to talk, but what to say and when.
“The next four to five billion people will not use AI the way today’s power users do,” cofounder and CEO Dheemanth Reddy told ET. “For them, voice is not a feature. It is how they live. The interface has to think and speak like them.”
That conviction was shaped partly by the founders’ time at NYU under AI pioneer Yann LeCun. “People think AI has solved everything because models can generate text,” Reddy said, “but that’s only a small part of intelligence.”
Maya 1 Speech Model Ranks Among the World’s Best
The company’s emotive Maya 1 speech model, released open-source under the Apache 2.0 licence, currently ranks sixth globally among open-weight models on Speech Arena, a benchmark for conversational speech systems. With a Quality Elo score of 1,051, Maya says the model sits roughly on par with OpenAI’s GPT-Realtime-2 — and that it is the only Indian company on the leaderboard.
The traction extends beyond benchmarks. Maya’s models have crossed 440,000 downloads on Hugging Face, while its consumer app has surpassed 3 million downloads across India, Southeast Asia, and the Middle East and North Africa.
Voice-First Markets: Telugu, Tier-II Towns, and Local Trust
Maya’s strongest signal comes from its voice-first markets, where adoption looks nothing like Silicon Valley’s. Telugu-speaking users form its largest base today, followed by Uttar Pradesh and West Bengal, with heavy usage among women discussing everything from shopping and parenting to devotion and family life.
To capture how people genuinely speak, the company sends staff into villages and small towns to record real conversations, collecting regional dialects instead of relying on standardised datasets. “We first wanted the model to speak like a local,” Reddy said. “That creates trust.” The biggest obstacle, he added, isn’t competition but connectivity: “Voice experiences are extremely sensitive to latency. If the internet connection is poor, the conversation breaks.”
What’s Next for Maya Research’s Conversational AI
Rather than chase subscriptions, Maya Research’s conversational AI aims to become a discovery and navigation layer — eventually helping users find financial products, government schemes, healthcare services, and commerce offerings they might never have located on their own.
“There are so many things people still need to discover,” Reddy said. “Technology has to help them navigate.” Cofounder and CTO Bharath Kumar Kakumani frames the mission in plainer terms: “When technology feels native, real adoption happens.”
It enters a crowded field that includes ElevenLabs, Cartesia, and Indian players like Gnani.ai, Skit.ai, and Yellow.ai — and a wider national push toward homegrown AI, echoed by raises like Innefu Labs’ sovereign-AI round. But if Maya’s wager on culturally fluent, voice-first AI holds, the company building for the next billion may end up defining how they come online at all.
For Indian startup funding and the voice AI race News visit KnowStartup.
