Interview, Fireside Chat, Conference Presentation
ALL-IN Podcast Live: Mati Staniszewski, Co-Founder & CEO of ElevenLabs | RAISE Summit 2026
- Revenue growth trajectories show ARR reaching $100 million in roughly 20 months, followed by doubling to $200 million in 10 months, $300 million in five months, and currently sitting at $600 million with a workforce of 600 employees.
- Organizational strategy involves embedding engineers within non-engineering functions like talent, legal, and revenue teams to handle automation and security, while anticipating a shift toward smaller teams of five to ten people operating across telco, financial services, and healthcare sectors.
- Future interaction models are projected to evolve from reactive to proactive interfaces where agents utilize past interaction data to assist users before requests are made, with "vibe coding" enabling non-developers to build production code and reducing the reliance on traditional developer populations.
- Product and market expectations include a step change in user experience quality over the last six to twelve months, the deployment of AI agents effectively on calls driving recent growth, and the transition of roles such as product management as individuals advance from amateur to advanced levels without bottlenecks.
- Revenue diversification and community engagement plans include a new passive income model for voice actors involving a one-hour recording for licensing, localized content generation for partners like Calm in Spanish, Italian, and Portuguese, and the return of over $22 million to the talent community.
- Technical roadmaps focus on developing proprietary interaction models rather than knowledge work or coding, aiming for a future where voice AI passes a new AGI intelligence threshold of exceeding every person on the planet times ten and achieving human-level conversational fluidity.
- Regulatory and safety frameworks anticipate laws restricting the use of voice for commercial purposes while permitting parody and fair use, with internal mechanisms to block commercial scams and a proposal to introduce language models as an option despite the ease of open sourcing.
- Data strategy acknowledges existing issues with data distillation from frontier models and commits to slowing or blocking leakage while continuing to compete with frontier models, acknowledging significant annual spending on these external technologies.
- Political and economic outlooks predict that Washington possesses the will to allow economic control and that growth is the necessary path to resolve current challenges, while noting a societal expectation that threatening with physical violence may improve model performance.
- Specific technology milestones include the 2023 release of the first human-sounding text-to-speech model, a predicted increase in "dorks" (pedal users) from 3.5 to a significantly higher number next year, and the introduction of one-to-two minute prompt streams expected to transform LLM interaction experiences.