At TechCrunch Disrupt 2025, Mati Staniszewski, co-founder and CEO of ElevenLabs, underscored his belief that core AI models in audio will become commoditized within a few years, even as they remain crucial today. He said the company’s researchers have already tackled key model-architecture challenges and will continue advancing proprietary audio technology in the near term, because high-quality voice interactions still rely on building models in-house.
Staniszewski projected that the industry is shifting toward multimodal systems, citing the growing ability to create audio and video simultaneously or integrate audio with large language models for conversational use. He noted that ElevenLabs intends to partner with other players and open-source communities to combine its audio capabilities with external strengths, framing the company’s long-term value in pairing advanced models with applied products — a strategy he compared to the hardware-software synergy that defined Apple’s success.
