Neuphonic raises £3M fo text-to-speech Voice AI

Posted by Admin October 17, 2024 Tech

Neuphonic, a UK startup redefining human-AI communication with the world’s fastest text-to-speech technology, has raised £3 million in pre-seed funding. The round was led by Moonfire VC, one of the top 10 data-driven VCs in the world, based on the percentage of engineers in the team, with participation from Tiny VC, Salica Oryx Fund, and Cur8 Capital.

Until now, Conversational AI’s potential has been held back by major tech constraints – text-to-speech models are too large, slow, expensive, and unnatural-sounding. Neuphonic is changing this: its patent-pending algorithm enables real-time, incremental speech generation with ultra-low latency of just 25 milliseconds— making it the world’s fastest text-to-speech solution. This incremental methodology also allows Neuphonic to work with any Large Language Model in a way that’s more human-like and language agnostic. Neuphonic’s API is available to customers who want to create human-like speech in their products through an exclusive closed beta program.

“High latency in Voice AI prevents natural interaction and slows growth in key fields like gaming, conversational AI, digital avatars, and real-time translation” said Sohaib Ahmad, Co-founder and CEO of Neuphonic. “People are struggling to really interact with Voice AI as a result. We want to reach a point where AI feels like a natural extension of ourselves – intuitive and effortless. Ideally people then spend less time staring at screens and more time actually communicating.”

Neuphonic was founded by former Papercup co-founder Jiameng Gao and former hedge fund quant trader Sohaib Ahmad, who met at Cambridge University whilst studying Machine Learning. As multilingual first-generation immigrants with roots in China, Ireland, and Pakistan, Sohaib and Jiameng have a unique insight into language barriers and cultural nuances, which is what led them, alongside their passion for voice technology, to create Neuphonic and solve the challenges faced by existing text-to-speech solutions.

“By generating speech word-by-word as text arrives, we unlock a wide range of use cases for Text-To-Speech that wasn’t possible before – we’re in talks with businesses in customer service, digital reception, humanoid robotics, ed-tech, storytelling, and content creation. This goes beyond speed improvements and allows us to create AI interactions that feel as natural and responsive as human conversation” added Jiameng Gao, Co-founder and CTO of Neuphonic. “Just as how people speak immediately, our models bypass the need for full sentences and in doing so significantly cut down latency.”

“Voice AI has been a sleeping giant, held back by technical limitations that Neuphonic is now solving. Their technology has the potential to unlock significant value across multiple industries” commented Akshat Goenka, Partner at Moonfire. “In customer service, it could enable more natural, efficient interactions. For content creators, it opens up new possibilities in localisation and accessibility. In emerging fields like digital avatars and AI gaming, it could be the key to creating truly immersive experiences. We see Neuphonic’s solution as a catalyst for innovation in these sectors and beyond, potentially unlocking billions in economic value. They could ultimately enable entirely new business models and user experiences that weren’t possible before.”

“Neuphonic’s breakthrough in real-time speech synthesis will create a paradigm shift in human-machine interaction,” said Professor Steve Young CBE, Emeritus Professor of Information Engineering and former Senior Pro-Vice Chancellor of Cambridge University. “By reducing latency to near-human levels, they are paving the way for seamless voice interaction that could replace screens in many aspects of our daily lives.” Professor Young, an advisor and investor in Neuphonic’s current fundraise, highlighted the company’s potential to redefine the future of voice technology.

Headquartered in King’s Cross, London, Neuphonic plans to use the funds to expand its language capabilities and voice offerings, enhance model performance by expanding research, and develop on-device solutions. With a growing team and a waiting list of hundreds of potential users and businesses, the company is positioned for rapid growth in a voice AI market projected to reach USD 41.39 billion by 2030.