ElevenLabs

• 1. Introduction

  ›

  ElevenLabs, in March 2026, pledged a billion dollars in free voice restoration technology to one million people living with permanent voice loss. The company had by then grown far beyond its 2022 roots. Its software could replicate the human voice with enough precision to fool voters in a presidential primary. Behind that capability is a founding story that begins not in Silicon Valley but in Poland.

  19505115

• 2. The Dubbing Frustration

  ›

  Piotr Dąbkowski left Google's machine learning team to co-found ElevenLabs in 2022. His co-founder, Mati Staniszewski, had previously worked at Palantir as a deployment strategist. Both men were raised in Poland. Their inspiration for starting the company reportedly came from watching American films that had been inadequately dubbed. That frustration pointed toward a genuine and unsolved problem in voice technology.

  Before raising money, the pair considered working with a startup accelerator and eventually chose to seek independent funding. In January 2023, they announced a $2 million pre-seed round led by Credo Ventures, with Concept Ventures also participating. ElevenLabs' focus on AI voice intelligence helped attract backers because the field remained largely underdeveloped in Europe. The same month the funding was announced, the company opened its beta platform to the public.

  1792024

• 3. The Emotional Voice Engine

  ›

  Speech Synthesis, ElevenLabs' core product, was built around the idea that context shapes how words should sound. The company stated that its models analyzed text to adjust intonation and pacing, and that the underlying system aimed to detect emotions such as anger, sadness, happiness, and alarm. ElevenLabs was pursuing patents on this approach.

  VoiceLab, the company's voice cloning tool, let users create a new synthetic voice from just a few short audio snippets. A Voice Library stocked with more than 1,000 community-created profiles gave other users a ready pool of voices to choose from. Paying subscribers could also upload their own recordings to train a custom voice style.

  In July 2023, ElevenLabs announced Projects, a tool for generating long-form spoken content such as audiobooks using synthetic or custom voices. It launched that September. August 2023 brought an expansion to 28 languages, with automatic detection covering Korean, Dutch, and Vietnamese among others. The company also declared that its technology had officially exited its beta phase.

  By October 2023, AI Dubbing had arrived, translating speech into more than 20 languages while preserving the original speaker's voice, emotions, and intonation. Proprietary methods handled noise removal, speaker differentiation, transcription, and synchronization.

  The ElevenLabs Reader App launched on iOS and Android in June 2024, letting users listen to articles, PDFs, and ePubs through synthetic voices. Voice Isolator followed a month later, stripping background noise from audio. In November 2024, Conversational AI launched as a developer platform for building interactive voice agents. In February 2025, the company introduced a platform enabling authors to create and publish audiobooks directly through the Reader app.

  Scribe, also released in February 2025, addressed transcription: it processed spoken audio with character-level timestamps and identified individual speakers. ElevenLabs said Scribe's word error rate led the industry, citing internal benchmarks and tests by Artificial Analysis. Eleven v3, launched in June 2025, added support for more than 70 languages, natural multi-speaker dialogue, and audio tags including excited, whispers, and sighs. In May 2024, ElevenLabs had shared early samples from a text-to-music model it said would become available for public use in July 2025.

  2829313236

• 4. Investors and Growth

  ›

  In June 2023, ElevenLabs raised $19 million in Series A funding at a valuation of around $100 million. Andreessen Horowitz co-led the round alongside former GitHub chief executive Nat Friedman and entrepreneur Daniel Gross. Additional investors included Mike Krieger, a co-founder of Instagram, and Brendan Iribe, a co-founder of Oculus. Mustafa Suleyman, who co-founded DeepMind, also participated, along with Tim O'Reilly of O'Reilly Media and SV Angel. Andreessen Horowitz also took a seat on ElevenLabs' board.

  Between the January 2023 launch and June of that year, ElevenLabs passed one million registered users. The platform drew praise for the quality of its voice output, its generation speed, and a free tier described as generous. Reviewers also noted that the system handled unusual name pronunciations better than competing tools, which tended to favor Western names.

  On the 22nd of January 2024, ElevenLabs raised $80 million in Series B funding, bringing its valuation to $1.1 billion. Andreessen Horowitz, Friedman, Gross, and Sequoia Capital co-led the round. The announcement came alongside new products: a Voice Marketplace, an AI Dubbing Studio, and a mobile app.

  On the 30th of January 2025, ElevenLabs closed a $180 million Series C round at a valuation of $3.3 billion. a16z and ICONIQ Growth co-led the investment, joined by NEA, World Innovation Lab, Valor, Endeavor Catalyst Fund, and Lunate. Strategic investors included Deutsche Telekom, LG Technology Ventures, HubSpot Ventures, NTT DOCOMO Ventures, and RingCentral Ventures. In April 2025, the company was named to the Forbes AI 50 list. On the 4th of February 2026, ElevenLabs announced a $500 million raise at an $11 billion valuation and indicated it was considering a public offering.

  138141618

• 5. The Robocall Problem

  ›

  In the lead-up to the January 2024 New Hampshire Democratic primary, thousands of residents received automated voice calls that appeared to feature the voice of Joe Biden. The calls encouraged voters to skip election day. The New Hampshire attorney general's office investigated and traced the calls to a company based in Texas. Audio experts who examined the recording concluded it had been produced using ElevenLabs' technology. CEO Mati Staniszewski said the company was dedicated to preventing the misuse of audio AI tools but provided no comment on that specific incident.

  After its January 2023 launch, users on 4chan had generated hateful messages in the vocal styles of celebrities and public officials, drawing wide attention to how the technology could be misused.

  On the 20th of June 2023, ElevenLabs released the AI Speech Classifier, a tool it described as the first of its kind. Accessible through an API, the classifier was designed to determine whether an audio sample had been generated by ElevenLabs' own technology. ElevenLabs stated its goal of collaborating with other AI developers to build a universal detection system that could be adopted industry-wide.

  154781043

• 6. Music, Film, and Agents

  ›

  Eleven Music, launched in August 2025, was built in collaboration with record labels, publishers, and artists. It was cleared for commercial use across film, television, podcasts, social media videos, advertisements, and gaming. Users could generate studio-grade tracks from natural language prompts, with control over genre, style, structure, and whether the result included vocals. The platform supported multiple languages including English, Spanish, German, and Japanese. A proprietary model developed by ElevenLabs' own research team powered the generation, capable of blending genres and instruments in real time.

  ElevenLabs had already tested the creative space with The Next, an album featuring vocals produced through its voice synthesis tools. Songs on that project blended varying levels of human and synthetic contribution, and streaming earnings went back to all participating artists. ElevenLabs also produced the documentary series 11 Voices, which premiered at SXSW in 2026. The series featured interviews with people working across artificial intelligence, audio production, and related fields, covering voice synthesis, ethics, and creative practice.

  ElevenAgents' Expressive Mode, released in February 2026, combined two updates to the company's conversational voice technology. The first was Eleven v3 Conversational, a text-to-speech model built on Eleven v3 and optimized for real-time dialogue. The second was a redesigned turn-taking system aimed at producing fewer interruptions and better-timed responses. The Eleven v3 Conversational model was described as emotionally intelligent and context-aware, designed to handle the rhythms of real spoken exchange rather than scripted prompts.

  3738485039