Summary: ElevenLabs closed a $500 million Series D on February 4, 2026 at an $11 billion valuation, led by Sequoia Capital, after ending 2025 with reported annualized revenue above $330 million. The company is positioning itself as an enterprise voice-agent platform across agents, creative audio, and developer APIs. New funding is earmarked for conversational agents, research, and international expansion.
ElevenLabs, the voice AI company, closed a $500 million Series D on February 4, 2026, at an $11 billion valuation with Sequoia Capital leading the round. Twelve months earlier, in January 2025, it had pulled in $180 million at a $3.3 billion valuation — meaning the company’s price tag grew more than threefold within a year. Founded in 2022, the startup has now collected $781 million across five rounds of funding.
The more consequential change sits behind the dollar figure: ElevenLabs is recasting its identity, moving away from “voice tool for content creation” and toward “enterprise voice agent platform.” Per TechCrunch and CNBC, the company ended 2025 with annualized revenue above $330 million, and co-founder Mati Staniszewski says it is building “toward IPO and beyond.”
How One Round Left the Valuation Three Times Larger
The Sequoia lead comes with partner Andrew Reed joining the board. What speaks loudest, though, is the conviction of the existing cap table: Andreessen Horowitz grew its stake fourfold, while ICONIQ tripled its exposure. Newcomers include Lightspeed Venture Partners, Evantic Capital, and BOND, and returning names include BroadLight, NFDG, Valor Capital, AMP Coalition, and Smash Capital. CNBC’s report adds NVIDIA to the list of supporters.
The pricing logic is easy to trace: within a single year the valuation traveled from $3.3 billion to $11 billion, and what carries it is $330 million-plus of ARR along with an enterprise adoption curve — not model demos alone.
A Three-Layer Stack: Agents, Creative, API
Three layers now carry the commercial structure. ElevenAgents is built for enterprise voice and conversational agents, with Deutsche Telekom, Square, Revolut, and the Ukrainian government as customers, across customer support, commerce, training, and sales. ElevenCreative covers audio generation, editing, and localization in more than 70 languages, and Duolingo, NVIDIA, and TIME number among its users. ElevenAPI is the low-latency infrastructure layer for developers — Meta, Epic Games, Salesforce, MasterClass, and Harvey all run on it, and by the company’s count, platforms built on it combined reach over one billion users.
The core model family ranges across text to speech, speech to text, sound effects, dubbing, music, and conversational AI — with the input side and the output side of the audio pipeline both shipped as products.
Where the $500 Million Goes: Conversational Models and International Buildout
Three destinations absorb most of the $500 million. The first is growing ElevenAgents, which includes a substantial voice agent upgrade: quicker responses, stronger expressiveness, redesigned turn-taking behavior, plus the Eleven v3 Conversational model. The second is expanding the research team, aimed at emotional conversational models, dubbing, and what the company calls audio general intelligence. The third is international go-to-market, as the team spreads across 14 cities that include London, New York, Tokyo, São Paulo, and Singapore.
The Enterprise Pivot of Voice AI
Read against the wider industry, the round marks a shift: voice is moving out of “the last mile of content production” and into “the primary interface for agents.” The moment an agent has to place a phone call, pick up a support line, or carry on a real-time conversation, voice stops being a feature and is the product itself. Billing follows the same drift — per-minute API pricing is ceding ground to per-agent, per-seat enterprise contracts, and unwinding one of those contracts costs far more than comparing API rates ever would.
That $11 billion price is a wager on exactly this conversion from tool to platform. Sequoia and the insiders who multiplied their stakes four- and threefold are betting that enterprise voice agents turn into a standard budget line — as contact-center software once did — over the next few years.
The Takeaway for Developers and Product Teams
Three points matter in practice. One: when voice sits at the heart of your product, rethink procurement — a platform contract (agent orchestration, turn-taking, call interruption handling) binds you at a different depth than a bare TTS API, so work out what leaving would cost before you sign. Two: low-latency conversation is the next competitive axis — put Eleven v3 Conversational’s turn-taking and expressiveness through tests inside your own scenarios instead of trusting vendor demos. Three: open-source speech models are maturing in parallel; wherever privacy or cost dominates, mixing self-hosted and API speech infrastructure is becoming the default. Build your vendor abstraction with care — it holds up better than staking everything on a single platform.