ElevenLabs' funding success shows growth of AI voice market
The startup has quickly raised a total of $101 million, reflecting the fast growth of new technologies like deep neural networks. Along with growth comes challenges.
AI voice startup ElevenLabs continues to grow, reflecting the strength of the AI voice technology market.
ElevenLabs on Monday revealed that it raised $80 million in a series B funding round led by Andreessen Horowitz.
The new round brings the New York-based startup to a total of $101 million in funding and comes a year after the vendor launched its VoiceLab platform.
In addition to the funding, ElevenLabs introduced new product developments. Its new Dubbing Studio workflow allows users to dub entire movies and generate and edit transcripts, translations and timecodes.
A Voice Library marketplace provides a platform on which users can earn from their AI voice clone. Users can create professional voice clones and share them for compensation when others use them.
Finally, an early preview of the Mobile App reader allows users to convert texts and URLs into audio.
A growing market
ElevenLabs' rapid growth in just under two years comes as new technology has paved the way for the AI voice market to grow and evolve.
New techniques have come in to have made a lot of conversational and voice capabilities more capable.
Bern ElliotAnalyst, Gartner
"New techniques have come in to have made a lot of conversational and voice capabilities more capable," said Gartner analyst Bern Elliot.
These include, for example, advancements in deep neural networks, and, more recently, large language models with the popularization of generative AI technology.
"These created a lot of interest in applications that are language-based, and this would include conversational ones," Elliot said.
As for ElevenLabs, it had been growing quietly for a while before its public launch, he noted.
Therefore, its success comes from introducing the right capabilities into the market at the right time.
For instance, ElevenLabs' new Dubbing Studio comes at a time when there is a lot of video streaming in different markets in which people wish to listen in their local language.
"Dubbing is very hard to do," Elliot said. He added that part of the difficulty is because a phrase in one language could include several more words in another language. "If [ElevenLabs is] able to improve the automation of dubbing, that will be well received."
ElevenLabs' popularity also comes from its platform's ease of use, said Forrester analyst Zeid Khater.
"They don't require a lot of source audio," he said. As long as users have about 15 seconds of clean source audio, the platform can generate audio quickly.
Moreover, ElevenLabs has also improved its features fairly rapidly, growing from text-to-speech and speech-to-speech.
"Most voice clone apps typically kind of sound a little robotic," Khater said. "That's the one thing ElevenLabs has abstracted away. It's very good at making it sound natural."
The challenge with AI voice
However, despite the continued growth and interest in AI voice technology, there remains a threat of misuse.
Also of concern is the threat of intellectual voice property theft.
"There's a lot of AI technologies right now that companies and individuals and even legislators are grappling with that represents new legal territory," Elliot said.
Even ElevenLabs has been accused of taking samples of voice actors without their knowledge.
While one of ElevenLabs' competitors, Replica Studios, has attempted to combat this by partnering with the SAG-AFTRA actors union, ElevenLabs' new voice library could be an approach for creators to make sure there's evidence of where a certain voice clone or generated voice comes from.
At the same time, the voice library could also lead to the theft of more voices, Elliot said.
"It's a dynamic that needs to be worked through as these technologies mature and evolve," he said.
It appears that ElevenLabs has been working on making sure its technology is secure based on subgroub posts from Reddit users who reported their accounts being locked based on violating some terms, Khater said.
Some of those terms may be including unauthorized audio that was acquired secretly.
"We see tighter restrictions kind of just from the user community, but nothing that I've seen at least directly from them," he said. "Understandably, they do need to as a company vet to make sure that no one is using the technology in a way that is potentially harmful."
Ultimately, problems with AI voice technology correlate with the larger intellectual property and ownership issues in the generative AI market, Khater continued.
"The world of voice and video are far more sensitive," he said. "A lot of the underlying problems are going to be in license laws or if you can use my likeness for a commercial or creative endeavor."
ElevenLabs says it plans to use the new funding to advance research, expand infrastructure, develop new products and enhance safety measures for responsible and ethical AI development.
Also participating in the new funding round were Nat Friedman and Daniel Gross. They were joined by Sequoia Capital, SV Angel, Smash Capital, BroadLight Capital and Credo Ventures.
Esther Ajao is a TechTarget Editorial news writer covering artificial intelligence software and systems.
