Valued at approximately 19.4 billion euros during a liquidity operation of 264 million euros, ElevenLabs is accelerating its transformation. The company born from speech synthesis now wants to become an agent platform capable of supporting end-to-end customer interactions. Its trajectory is putting pressure on French telephony and contact center specialists, but also American platforms and Chinese suppliers of voice models.
ElevenLabs has not just raised 264 million euros, but organized a share buyback offer of 300 million dollars, or approximately 264 million euros, from employees and existing shareholders. Investors acquire shares that have already been issued. The operation gives liquidity to the teams and above all establishes a new reference valuation, namely 22 billion dollars, or approximately 19.4 billion euros.
Wellington Management and T. Rowe Price led the transaction, joined by EQT, Goldman Sachs, GIC, Ontario Teachers’ Pension Plan, Sapphire Ventures and BDT & MSD. The signal is that of an institutionalization of capital. After raising $500 million in February 2026, on a valuation of $11 billion, ElevenLabs doubles its displayed value in eight months.
This mechanism also allows the company to defer the pressure of an IPO, while offering a first window of liquidity to employees. It comes as ElevenLabs transitions from a voice technology provider to an enterprise AI platform. Founded in 2022 by Mati Staniszewski and Piotr Dąbkowski, the company now has more than 800 people, spread across around twenty countries.
From synthetic voice to processing a customer request
ElevenLabs’ starting point was to generate synthetic voices expressive enough to be used in dubbing, localization, editing or content production.
As voice models multiply, the quality of synthesis becomes more accessible and the risk is that voice becomes a technical commodity, consumed via API and chosen primarily on price, latency or availability.
ElevenLabs seeks to escape from this logic with ElevenAgents, whose ambition is no longer just to make an artificial voice respond, but to connect an agent to a documentary database, a CRM, ticketing software or a planning system. The agent must understand a request, find the customer context, apply a business rule, perform an action and transmit the file to an advisor when it reaches a limit.
The company says B2B customers represent 55% of its revenue. Its agents now handle more than 15 million conversations per week, three times more than in February, and ElevenAgents’ annual recurring revenue has more than tripled over the same period. The cases cited (reimbursements, insurance renewals, changes to plans, making medical appointments or access to public services) illustrate the real target: industrializing the frequent reasons for contact which clog up customer relations centers.
We invite you to download our study on this subject, AI Agents, CRM, commercial performance, data and sovereignty…
A French team built for corporate deployment
France is becoming a priority market in this trajectory. ElevenLabs has set up a sales and deployment organization there led by Julien Lesaicherre, appointed General Manager France. The company is recruiting profiles in major account sales, partnerships, solutions engineering and forward-deployed engineeringthat is to say teams responsible for adapting and integrating products with customers.
The objective is to quickly recruit around twenty positions in France in 2026, with priority given to sectors where sales cycles are long, security requirements high and the volume of interactions high: banking, insurance, telecommunications, health and community services.
This local presence reveals an evolution of the model; until now, ElevenLabs could sell an API to a creator, a studio or a developer. In customer relations, you have to convince business management, IT teams, lawyers, compliance managers and support organizations. Above all, it is necessary to connect the agent to existing tools and demonstrate that it can produce an economic gain without degrading the experience.
Ringover, Diabolocom and Odigo: the telephone channel is already busy
This is where French competition takes on a particular dimension. ElevenLabs brings the model, speech synthesis and its agent platform. But other companies already control the phone number, routing, queues, registrations and connections to tools used by sales or support teams.
Ringover is one of the actors best placed to illustrate this shift. With AIRO, the company offers a voice agent integrated into the interactive voice server: it welcomes callers, qualifies requests, handles simple cases and routes the exchange to human teams. The agent can, for example, manage appointment scheduling, thanks to integrations with tools like Calendly or Cal.com.
Ringover’s position is interesting because it is both competitive and potentially complementary. Its telephony layer can also be linked to external agents, including ElevenLabs. The template provider provides conversational intelligence, Ringover maintains the customer’s channel, number, IVR and operational configuration.
Diabolocom and Odigo are located on neighboring land, more oriented towards the contact center and large organizations. The first combines voicebot, routing, interaction analysis and assistance to advisors. The second, which strengthened its offering with the acquisition of Akio, is pushing an omnichannel AI agent approach intended for complex customer relationship environments.
American competitors have the models, the cloud and the large accounts
American competition takes place on several levels. OpenAI now offers bricks dedicated to real-time voice agents, capable of dialoguing directly by voice, maintaining a conversation state and calling external tools. Its offers are explicitly aimed at support uses, order taking or telephone reservations.
Google has, with its Contact Center AI platform, a complete contact center offering combining omnichannel routing, virtual agent, assistance to advisors and interaction analysis. Amazon, for its part, is pushing agentic voice functions into Amazon Connect, where agents can understand, reason and act on voice or digital channels.
Compared to ElevenLabs, these groups have a distribution advantage; they are already present in the cloud environments, data systems and architectures of many major accounts. At the other end of the market, more specialized companies like Retell or Vapi focus on quickly assembling telephone agents and are already offered as integration options by Ringover.
The challenge for ElevenLabs is therefore to remain more than a premium voice provider. Its valuation assumes that it imposes its own orchestration layer, its integrations, its control tools and a direct relationship with companies.
China puts pressure on the technological layer
Chinese competition is less visible in incoming calls from European companies, but it is already weighing on the voice technology economy.
Alibaba offers Qwen real-time voice chat, voice synthesis and voice cloning models. Some integrate voice, function calling and research capabilities. Alibaba Cloud also supports voice solutions from MiniMax in its agent creation tools.
This competition does not automatically translate into a commercial breakthrough in Europe. Data access, localization requirements, regulatory compliance, and key account preferences limit direct deployments. But it exerts fundamental pressure on prices and the pace of innovation. The more powerful voice models become available at low cost, the more differentiation shifts toward integration, deployment, and trust.
This is probably where the battle will be decided, a company can change its voice model more easily than it can change its customer relations system, routing rules, document base or integration partner.
Conclusion: voice becomes a strategic layer of the company
The valuation of ElevenLabs not only rewards an advance in voice generation, but anticipates the creation of a new software layer between the company and its customers: agents capable of understanding, speaking, accessing information and acting in business systems. We invite you to read our study on this subject
The company has a brand and product lead in voice quality. It is now building local teams to transform this advance into corporate contracts. But it is entering a market where positions are already occupied: Ringover, Diabolocom and Odigo own a share of the infrastructure and uses in France; OpenAI, Google and Amazon have the models, cloud and global distribution, Chinese suppliers are gradually reducing the cost of the technology layer.
For ElevenLabs, the challenge is no longer to prove that an AI can speak like a human, but to demonstrate that it can become, in a controlled setting, a reliable and useful interlocutor for millions of customers.
