- Voice AI startups raised over $398 million in VC funding in 2024, per PitchBook data.
- The technology is expanding into enterprise uses like customer service and assistants.
- BI spoke to investors about the untapped opportunities in the nascent voice AI space.
Voice is fast becoming the new AI battleground.
From buzzy virtual assistants to speech synthesis tools, the technology has taken off in the past year.
While AI voice tech isn’t new, the tools have rapidly become more sophisticated, driving adoption from the call center to recruitment agencies.
Its use cases are vast, from real-time audio transcriptions to generating synthetic voices from text prompts.
Investors looking for the next opportunity in the highly competitive AI market have thrown their checkbooks behind startups. According to PitchBook data, startups developing voice AI technology raised over $398 million in VC funding in 2024.
London-based PolyAI, which has developed voice assistants for call centers, secured $50 million in a funding round from Hedosophia. London and New York-headquartered ElevenLabs, which has developed a voice cloning technology, raised $100 million in January 2024 — and is said to be raising a further $200 million, Business Insider first reported.
“Recent breakthroughs in real-time speech-to-speech processing have unlocked new use cases, including virtual assistants, customer support, and voice-based productivity,” said Sivesh Sukumar, an investor at VC firm Balderton. “Companies like ElevenLabs and OpenAI are at the forefront of this space, with ElevenLabs releasing a real-time API expected to drive further adoption.”
Voice AI is a comparatively nascent space, so there isn’t an established incumbent yet — but it’s triggering investor excitement for the untapped opportunities in the sector, Sukumar added.
An expanding ecosystem
Startups are quickly identifying how to tailor voice technology to a host of enterprise and consumer needs. And with agentic AI a hot topic for CEOs, its overlap with voice technology could hold high potential.
PlayAI, a startup that is developing an AI platform for text-to-speech models and AI voice agents, raised $21 million in seed funding in November.
“We’ve seen a massive increase in interest in building voice agents, which a human can speak to just like it’s another human,” said PlayAI cofounder Hammad Syed. “Voice AI is going mainstream and will be a key interface in how people interact with technology. Investors definitely realize this opportunity,” he added.
VCs scouring the ecosystem to make their next big bet are now looking at voice AI as a technology stack, said Steve Jang, founder and managing partner at Kindred Ventures, which also backed PlayAI. The firm’s investment thesis is to back startups “with multiple layers with many use cases in consumer, enterprise, and creativity.”
“First, there are specialized and foundational models. Second, there are infrastructure services and tools, which offer access and integration with these models. And then, perhaps most importantly, there is the vast vertical application space,” he told BI.
The sector is also attractive to investors because voice is an easy category to cash in on. “You can price it by the outcome, so it’s pretty easy to monetize,” said Jonathan Userovici, general partner at VC firm Headline. “That’s why you have so much revenue traction — it’s pretty easy to get a return on investment, especially if you’re replacing a human doing that labor.”
Consumer appetite for voice AI has also skyrocketed. With more users preferring to take in information through audio formats like podcasts, Sukumar highlighted the growing consumer demand for voice control and audio interfaces. He built PersuAIsion, a voice AI platform that allows users to practice real-world conversations — from job interviews to first dates — because he saw the scope for voice to cater to such consumer needs.
“If OpenAI can capture the consumer voice agent, they’ll be what Siri was supposed to be,” he said. “I think there’s going to be a lot more interfacing with personal devices, and there’s just going to be better e-commerce consumer experience on that front.”
Frontier labs are catching up
Despite its growing popularity, voice AI doesn’t seem to have one established juggernaut just yet. Part of the reason could be that frontier labs have largely stayed away from the space, possibly due to a concern that a misuse of voice generation capabilities could result in a potential backlash, according to Air Street Capital’s 2024 State of AI report.
“Despite scraping huge quantities of audio and video data, frontier labs have been slow to release text-to-speech products,” said Nathan Benaich, founder and general partner of Air Street Capital. He pointed to OpenAI’s Advanced Voice Mode, which was repeatedly postponed, and Google’s NotebookLM, which “is relatively locked down.”
AI experts had sounded the alarm about the possible rise of deepfakes in a year marked by global elections — but that didn’t end up being the case.
“In all likelihood, labs were keen to avoid being dragged into the panics about deepfakes that often accompany major elections. I think it’s inevitable they will play more in this space, just because the potential commercial opportunity is so large,” Benaich said.
Big Tech may be slowly pivoting toward the trend. Amazon’s plans to ramp up its voice assistant offerings through Alexa were delayed until 2025, and Apple recently bulked up its Siri feature by adding ChatGPT capabilities.
Still, Benaich noted that it won’t be an easy task for any one company to take the crown. “Displacing companies like ElevenLabs, who already enjoy widespread adoption and have been optimizing their tools for enterprise users for years now, could prove challenging,” he said.
Source link
lol
