ElevenLabs has released Dubbing v2, an upgraded AI dubbing model that preserves a speaker's emotion, tone, and identity across more than 90 languages, the latest in a rapid sequence of product launches positioning the company as an audio platform rather than a point solution.
For most of media history, dubbing a film or series for a foreign market meant hiring translators, casting local voice actors, scheduling studio time, and waiting weeks or months for a finished cut. The result was often uncanny: the emotional register of the original performance would get lost somewhere between the script adaptation and the recording booth. ElevenLabs thinks it has solved that problem. Dubbing v2, released this month, conditions on the source audio performance rather than a transcript, which means the model carries across not just the words, but the hesitation, urgency, warmth, or tension of the original speaker's delivery. The voice on the other end sounds like the same person, just speaking a different language.
That is a technically meaningful distinction. Earlier dubbing tools, including ElevenLabs' own first-generation product, worked primarily from text, which made them good at accurate translation but limited in conveying the performance underneath the words. Dubbing v2 skips the transcript as a dependency and treats the audio signal itself as the source of truth. Combined with automatic voice cloning that requires no manual setup, the full pipeline, translation, cloning, dubbing, and synchronization, runs end-to-end without human intervention.
ElevenLabs is not alone in targeting this market. HeyGen has built a strong position in video dubbing with lip-sync capabilities across 175 languages and has become a popular choice for content creators who need the visual layer matched to the audio. Papercup, whose IP was acquired by localization giant RWS, serves the broadcast and documentary end of the market with a human-review layer built into every delivery. Deepdub focuses on emotional preservation for entertainment and streaming clients. Each of these players is angling for a piece of a global content localization industry that runs into the tens of billions of dollars annually and still runs largely on slow, labor-intensive human workflows.
ElevenLabs is competing on the quality of its voice cloning first, with lip-sync as an optional add-on. That is a deliberate positioning choice: the company's core strength has always been voice synthesis, and Dubbing v2 extends that advantage into the localization vertical. Whether enterprise streaming clients, the kind of buyers who make multi-year platform deals, are committing to ElevenLabs at scale is not yet publicly confirmed, but the product is clearly being built for that audience. API access is rolling out initially to select enterprise clients, with broader self-serve availability still ahead, a sequencing that suggests ElevenLabs is prioritizing high-value relationships over volume at launch.
The platform thesis is becoming clearer
What makes the Dubbing v2 launch worth watching beyond its feature set is the company's cadence. As TechCrunch reported, just two days before the dubbing announcement, ElevenLabs shipped Music v2, a generative music model capable of switching genres mid-track, sustaining fast rap delivery, and embedding sound effects within compositions without breaking musical coherence. The company simultaneously cut Music API pricing by up to 50 percent and launched ElevenMusic, a platform for listening, remixing, and creating original tracks. That is a lot of product activity in a very short window.
The pattern is intentional. ElevenLabs is building what it describes as the audio layer of the internet, organized around three pillars: voice synthesis, music generation, and now professional-grade dubbing. Each product feeds the same API, the same subscription tiers, and the same platform infrastructure. For enterprise clients, that bundling matters: a media company that uses ElevenLabs for voice-over work can now extend the same relationship to localization and original music without adding a new vendor. For investors watching the company's monetization trajectory, the question shifts from whether ElevenLabs can grow its user base to whether it can convert that base into high-value enterprise contracts.
The localization market is a meaningful test case. Human dubbing workflows for major studio productions routinely cost hundreds of thousands of dollars per title per language. The economics for volume buyers are dramatically different with AI dubbing, even accounting for quality review and post-processing. The more relevant question is not whether AI dubbing is cheaper, it clearly is, but whether Dubbing v2's quality is consistently good enough for professional deployment at scale. Performance-conditioned voice cloning is a real step forward on that front, but broadcast-grade clients in particular will want to see rigorous testing across diverse speaker types, accents, and content categories before committing.
ElevenLabs has roughly 18 months of momentum to convert into durable enterprise revenue before the next wave of competitors close the capability gap. Music v2 and Dubbing v2 launching days apart suggests the company is not waiting around. The practical signal to watch: whether major streaming platforms or broadcast networks publicly announce localization partnerships with ElevenLabs in the second half of 2026. That would be the inflection point that confirms the platform strategy is working beyond the developer and creator tier.
Also read: Liquid AI is betting that smaller edge models can beat bigger rivals • SpaceX trims its IPO valuation target to at least $1.8 trillion as Starship stays grounded and the June countdown begins • Liquid AI just made edge models honest enough for government work
