ElevenLabs CEO says AI audio models will be ‘commoditized’ over time

AI audio company ElevenLabs co-founder and chief executive Mati Staniszewski believes that AI models will be commoditized over time. This is a revealing comment for a company currently focused on building them. Speaking on stage at the TechCrunch Disrupt 2025 conference on Tuesday, the founder discussed his short-term and long-term views of the AI audio space.

Staniszewski said that his company’s researchers have been able to solve some model architecture challenges, and this focus will continue in the audio space for the next year or two. He stated that over the long term, the technology will become a commodity. He believes that even if differences remain for some voices or languages, those differences will become smaller.

When asked why ElevenLabs would focus on building models if he believed they would be commoditized, Staniszewski explained that in the short term, they are still the biggest advantage and the most significant step change available. For instance, if AI voices or interactions do not sound good, that is a problem that still needs to be solved. He said the only way to solve it is by building the models yourself, though he acknowledged that other players will also solve that problem over the long term.

He also noted that those looking for reliable and scalable use cases would likely use different models for different situations. However, in the next year or two, Staniszewski said an increasing number of models will move into multi-modal or fused approaches. This means you will create audio and video at the same time, or audio and large language models together in a conversational setting. He pointed to Google’s Veo 3 as an example of what can be achieved by combining models.

The founder said ElevenLabs plans to launch partnerships with other companies and work with open source technologies. The goal is to see if the company can combine its audio expertise with the expertise of other models. For ElevenLabs, the objective is to focus on both model building and applications to create long-term value. He added that in the same way software and hardware were the magic combination for Apple, he believes the product and AI will be the magic for creating the best use cases for this generation.