Twenty minutes with the CEO of ElevenLabs, now reportedly valued at $22 billion

The AI voice generation market is maturing rapidly, with significant capital flowing into companies that can deliver high-quality, human-sounding speech for enterprise applications.

ElevenLabs, a company specializing in AI voice technology that makes text sound human, is reportedly valued at $22 billion. The company, founded four years ago, claims an annual recurring revenue (ARR) of $600 million, with over 55% coming from enterprise clients. ElevenLabs' technology is used by major companies like Klarna, Deutsche Telekom, and Cisco for customer service, as well as by creators for audiobooks and dubbing. CEO Mati Staniszewski acknowledged increased competition, including from former customers like Decagon, but stated the company prioritizes market share growth even if it means lower gross margins. ElevenLabs also disclosed that governments are customers, requiring specific model deployments based on local regulations and data residency needs. The company is preparing for a potential IPO in the coming years, though the exact timing depends on market conditions.

ElevenLabs' reported $22 billion valuation and $600 million ARR indicate a massive market opportunity for sophisticated AI voice technology. The company's strategy of prioritizing market share over immediate gross margins suggests an aggressive growth phase, aiming to capture a dominant position in conversational AI. This approach, coupled with their work with governments and large enterprises, signals a move towards deeply integrated AI solutions that are becoming indispensable for customer service and communication.

The blurred lines between model providers, platforms, and application companies, as described by ElevenLabs' CEO, highlight a fundamental shift in the AI industry structure. Customers are increasingly choosing flexible "reasoning layers," opting for frontier models where high accuracy is critical and open-weight models for less sensitive tasks. This dynamic requires AI companies to offer a diverse range of model options and robust data handling, especially when serving government clients with strict data residency and security requirements.