ElevenLabs CEO Discusses $22 Billion Valuation and AI Voice Strategy

ElevenLabs, a four-year-old voice AI company, is reportedly valued at $22 billion following a $600 million annual recurring revenue run rate. The Toronto-based firm, co-founded by Mati Staniszewski, has established itself as a foundational provider of text-to-speech infrastructure, powering first-line customer support for enterprises like Klarna, Deutsche Telekom, Cisco, and Adobe, while also serving a growing portfolio of government agencies. During a recent appearance at the Nrth conference, Staniszewski outlined the company strategic priorities, technical architecture, and long-term growth trajectory amid an increasingly competitive AI landscape. Enterprise clients now constitute over fifty percent of ElevenLabs revenue, with the remainder driven by small businesses, developers, and creators utilizing the platform for audiobooks, media dubbing, and audio synthesis. The company operates with a flexible model deployment strategy, allowing clients to select between frontier proprietary models and open-weight alternatives based on use-case risk. While open-source solutions adequately handle informational customer service queries, highly regulated sectors like finance and healthcare require the precision and authentication of frontier models. Despite facing margin pressure from inference costs, Staniszewski confirmed the company prioritizes market expansion over short-term profitability, passing infrastructure savings directly to enterprise customers. Technological development remains heavily focused on data annotation and emotional intelligence rather than raw synthetic data volume. Staniszewski noted that thousands of internal contractors and professional voice coaches are utilized to map speech patterns, accent detection, and conversational dynamics. The competitive boundary between model providers and application developers continues to blur, as evidenced by firms like Decagon, which trained its voice product on ElevenLabs infrastructure before launching a competing service. The company maintains strict cybersecurity and compliance protocols, including mandatory KYC checks and the exclusion of self-replicating agent architectures, positioning itself differently from broader open-weight model hubs. On regulatory and ethical fronts, Staniszewski advocates for mandatory AI disclosure in customer interactions, though he predicts a cultural shift within five years as consumers grow accustomed to AI-mediated service. The firm supports industry-wide pacing on frontier model deployment but clarifies that its focus remains strictly on audio synthesis and conversational execution rather than foundational intelligence training. Looking ahead, ElevenLabs is preparing for a potential public offering, with leadership targeting 2028 as a realistic timeframe, contingent on market conditions and strategic readiness. The company trajectory underscores a broader industry transition toward specialized, enterprise-grade AI voice infrastructure built on scalable, compliant, and emotionally nuanced generation capabilities.