In this report
ElevenLabs is acquiring the Japanese voice AI lab behind Koto
TOKYO, Japan — ElevenLabs has announced that it is acquiring Kotoba, the Japanese voice AI lab that builds the speech model Koto. In its announcement (opens in a new tab) of 16 November 2026, ElevenLabs describes Kotoba as the leading Japanese voice AI lab and says the two companies will together accelerate the adoption of AI agents and voice interfaces in the world’s fourth-largest economy. The acquisition is expected to close later in Q4 2026, subject to customary regulatory approvals and closing conditions. Financial terms were not disclosed.
Kotoba was founded in 2023 by two AI researchers, Noriyuki Kojima and Jungo Kasai, and builds voice AI from Tokyo. Its model, Koto, handles text-to-speech, speech-to-text and speech-to-speech translation for simultaneous interpretation, and ElevenLabs notes that Koto already powers Japanese-language conversations in Salesforce’s Agentforce Voice. In its own announcement (opens in a new tab), Kotoba says its team will continue to work across research, product development and deployment support, and will also contribute its research to the models and products ElevenLabs delivers worldwide.

Generated image, not a photograph of ElevenLabs’ or Kotoba’s people or premises. Both companies describe voice as the interface they are building for.
The announcement follows a busy autumn for both companies. ElevenLabs closed a $300 million employee tender offer (opens in a new tab) that values it at $22 billion, and used its New York Summit on 11 November (opens in a new tab) to introduce new models, agents and deployment options. Kotoba announced a partnership with Cresta (opens in a new tab) on the same day, bringing its speech-to-text and text-to-speech models to Cresta’s Voice AI agents. Its most recent funding round (opens in a new tab), an additional USD 10 million in seed funding announced in June 2026, took Kotoba’s total funding to date to USD 23 million.
Kotoba has done exactly that for Japan, building state-of-the-art voice models that have already been adopted by some of the world’s leading enterprises.
This report follows the announcement through the two companies’ own words. Every figure comes from an ElevenLabs or Kotoba publication, and the calculations in the data workbook are computed from those published figures, so each number can be traced back to its source.
The deal in five points
Key facts
- Acquirer Disclosed
- ElevenLabsVoice and conversational AI company founded in 2022.
- Target Disclosed
- Kotoba TechnologiesJapanese voice AI lab founded in 2023 by Noriyuki Kojima and Jungo Kasai.
- Announced Disclosed
- 16 November 2026Published by ElevenLabs and Kotoba on the same day.
- Expected close Disclosed
- Later in Q4 2026Subject to customary regulatory approvals and closing conditions.
- Core technology Disclosed
- KotoSpeech-to-text, text-to-speech and speech-to-speech translation model family.
- Named deployments Disclosed
- Agentforce Voice and CrestaJapanese-language voice for Salesforce and a Cresta partnership announced 11 November 2026.
- Kotoba funding to date Disclosed
- USD 23 millionAs stated by Kotoba on 24 June 2026.
- ElevenLabs valuation Disclosed
- $22 billionStated by ElevenLabs on 30 September 2026 with its employee tender offer.
- Disclosed
- Stated by a party to the transaction
The numbers behind the deal
Each figure has its own permanent link and a ready-made citation. Journalists, analysts and researchers are welcome to quote them with credit and a link to Acquiry.
- USD 23 millionDisclosed
Kotoba total funding to date
Stated by Kotoba on 24 June 2026.
- $22 billionDisclosed
ElevenLabs valuation
Stated by ElevenLabs on 30 September 2026.
- $500M+Disclosed
ElevenLabs ARR
Surpassed in the first four months of 2026, stated 5 May 2026.
- 15 million+Disclosed
Weekly agent conversations
ElevenAgents, stated by ElevenLabs on 30 September 2026.
- 180,000+Disclosed
Kotoba app users
Stated by Kotoba on 24 June 2026.
Part I of III
The agreement
What ElevenLabs and Kotoba announced, who founded Kotoba and where the close sits on the calendar.
ElevenLabs and Kotoba set out one plan for voice AI in Japan and East Asia
Both companies published their announcements on 16 November 2026, ElevenLabs focusing on the market and Kotoba on research and product.
ElevenLabs and Kotoba published their announcements on the same day, and the two texts describe the same plan from complementary angles. ElevenLabs frames the acquisition around the market. It says it will accelerate the adoption of AI agents and voice interfaces in the world’s fourth-largest economy, and that its goal is to solve how people interact with AI in every language and culture. Kotoba frames it around research and product (opens in a new tab). It says joining ElevenLabs brings together Kotoba’s expertise in East Asian language research and product development with ElevenLabs’ research capabilities, computing resources and global business infrastructure.
Japanese first, then Korean and other East Asian languages
Kotoba’s announcement is specific about sequence. With these resources, it says, the team will improve voice AI across East Asian languages, starting with Japanese and extending to Korean and others, while accelerating its adoption in products and services. For businesses, Kotoba says, this will create more opportunities to bring voice agents capable of handling complex Japanese conversations, and technology that bridges language barriers, into everyday operations.
ElevenLabs adds a statement about how it intends to organise the work in Japan. It describes the acquisition as the first step in a long-term plan to make AI intuitive and trusted for everyday interactions across Japan, and says that in the coming months it plans to build out local sales leaders and continue to invest in infrastructure in Japan. It also looks forward to partnering even more closely with Japanese enterprises, and with multinationals operating and expanding in Japan.

Generated image, not a photograph of ElevenLabs’ or Kotoba’s people or premises. The acquisition is expected to close later in Q4 2026.
Kotoba’s research stays in Japan and also feeds ElevenLabs worldwide
Kotoba describes what it brings to the combination in two directions. Within Japan, the team wants to help strengthen Japan’s role as a centre for cutting-edge voice AI research and development, collaborating with researchers around the world and drawing on large-scale computing resources to bring insights from real-world use in Japan into its models and products. Beyond Japan, as a frontier AI lab with global ambitions, it will contribute its research to the models and products ElevenLabs delivers worldwide. Piotr Dąbkowski, co-founder and CTO of ElevenLabs, describes the technical benefit directly.
Bringing their research together with ours will improve Japanese across the entire stack ElevenLabs builds.
Kotoba also reflects on what the transaction represents for founders in Japan. It says the next chapter shows a path for an AI startup based in Japan to develop distinctive technology, join a global leader, and continue its research and product development on a much larger scale, and it hopes this offers researchers and entrepreneurs in Japan another model for growth.
Two PhD researchers from Cornell and the University of Washington founded Kotoba in 2023
ElevenLabs profiles Noriyuki Kojima and Jungo Kasai, whose backgrounds span Hugging Face, Fugaku-LLM and academic research.
ElevenLabs introduces the founders in its announcement. It says Kotoba was founded by AI researchers Noriyuki Kojima and Jungo Kasai in 2023, with the aim of developing voice AI for a multilingual world, and quotes the founders’ own line: “Speech shouldn't stop where language does.” Kojima, who goes by Nori, holds a PhD in computer science from Cornell. He worked as a research scientist at Hugging Face and co-built Japan’s Fugaku-LLM, an open-source model trained from scratch on the RIKEN Fugaku supercomputer.
Kasai holds a PhD in computer science and engineering from the University of Washington and was an assistant professor at the Toyota Technological Institute at Chicago. ElevenLabs describes him as one of Japan’s most-cited researchers. Kotoba’s June 2025 funding announcement (opens in a new tab) describes the company as founded in 2023 by Cornell and University of Washington PhDs Noriyuki Kojima as CEO and Jungo Kasai as CTO, and notes that it had recently appointed Yusuke Shimomura as Chief Business Officer.

Generated image, not a photograph of Kotoba’s people or premises.
A mission to build a frontier AI lab from Japan
Both founders describe the ambition in the same terms. Kojima says that when they founded Kotoba the goal was to build a frontier AI lab from Japan, and that three years later the team is more determined than ever to build technology that benefits society. Kasai says that since he began his AI research career he has dreamed of building a frontier AI research lab in Tokyo and creating products that transform people’s lives, from voice AI agents to simultaneous translation, and that the journey is just beginning.
Since I began my AI research career, I’ve dreamed of building a frontier AI research lab in Tokyo and creating products that transform people’s lives, from voice AI agents to simultaneous translation.
Investors who backed the research before the acquisition
Kotoba’s funding history shows a research-first approach. In its Seed 2 announcement the company said it had closed a USD 11.83 million round led by Globis Capital Partners and Boost Capital, with continued participation from SIP Capital and a group of global angel investors. In June 2026 it announced an additional USD 10 million in seed funding led by Kindred Ventures, with participation from Salesforce Ventures and Sony Innovation Fund, bringing total funding to date to USD 23 million. Ken Asada, Partner and Japan Head at Salesforce Ventures, says the firm has invested in both ElevenLabs and Kotoba.
From a 2023 start to a close expected later in Q4 2026
Dated milestones for both companies, drawn from their own publications.
The timeline below places the acquisition in the context of each company’s recent announcements. On the Kotoba side, the sequence runs from the Seed 2 funding round in June 2025 to the alpha release of its API and SDK in May 2026, the additional USD 10 million in seed funding in June 2026, Koto’s selection for Agentforce Voice in August 2026 and the partnership with Cresta on 11 November 2026. On the ElevenLabs side, it runs from the Series D in February 2026 to the $500 million ARR milestone in May 2026, the employee tender in September 2026 and the New York Summit on 11 November 2026.
Figure 1
Kotoba and ElevenLabs milestones, from 2023 to the expected close
2023Kotoba
Kotoba is founded
AI researchers Noriyuki Kojima and Jungo Kasai found Kotoba to develop voice AI for a multilingual world.
30 Jun 2025Kotoba
Kotoba closes a USD 11.83 million Seed 2 round
Led by Globis Capital Partners and Boost Capital, with continued participation from SIP Capital.
4 Feb 2026ElevenLabs
ElevenLabs raises a $500 million Series D
At an $11 billion valuation, led by Sequoia Capital.
5 May 2026ElevenLabs
ElevenLabs passes $500 million ARR
After ending 2025 with $350 million in ARR. The announcement is the third close of the Series D.
28 May 2026Kotoba
Kotoba releases its API and SDK alpha
Speech-to-speech translation, streaming text-to-speech and streaming speech-to-text for evaluation.
24 Jun 2026Kotoba
Kotoba adds USD 10 million in seed funding
Led by Kindred Ventures, with Salesforce Ventures and Sony Innovation Fund, bringing total funding to USD 23 million.
7 Aug 2026Kotoba and Salesforce Japan
Agentforce Voice becomes available in Japan
Koto powers its Japanese-language conversations. Kotoba announced the selection on 10 August 2026.
30 Sep 2026ElevenLabs
ElevenLabs closes a $300 million employee tender offer
The tender values ElevenLabs at $22 billion, double its Series D valuation.
11 Nov 2026Kotoba and Cresta
Kotoba announces its Cresta partnership
Kotoba’s speech-to-text and text-to-speech models are offered to Cresta’s enterprise customers.
11 Nov 2026ElevenLabs
ElevenLabs holds its New York Summit
New models, agents and deployment options are introduced.
16 Nov 2026ElevenLabs and Kotoba
ElevenLabs announces the acquisition of Kotoba
Both companies publish their announcements on the same day.
Later in Q4 2026ElevenLabs and KotobaExpected
Expected close
Subject to customary regulatory approvals and closing conditions.
Show method and sourcesHide method
Dates and terms are from the ElevenLabs and Kotoba publications listed in the sources. Items marked expected remain open until the companies announce them.
The two tracks converge in the week of 11 November 2026, when each company made its own announcement five days before the acquisition was announced. The transaction track then has a single open item: the close, expected later in Q4 2026.

Generated image, not a photograph of any party’s premises. Closing is expected later in Q4 2026, subject to customary regulatory approvals and closing conditions.
The remaining step is the close
Both companies use the same wording for the remaining step: the acquisition is expected to close later in Q4 2026, subject to customary regulatory approvals and closing conditions. Kotoba’s documentation describes its APIs as being in private alpha, available to selected customers who request access, which gives readers a second way to follow Kotoba’s activity while the transaction completes.
Part II of III
The technology and its customers
Koto, the deployments already running on it and why Japanese calls for dedicated research.
Koto covers speech recognition, synthesis and interpretation in one model family
Kotoba’s model runs in the cloud, on-premises and on-device, with Japanese, Korean, Chinese and English as core languages.
Koto is the technology at the centre of the acquisition. Kotoba describes it as a frontier voice model (opens in a new tab) for AI agents and speech translation, available on-premise, on-cloud and on-device. On its website, Kotoba lists sub-50ms text-to-speech, ultra-fast speech-to-text, and speech-to-speech translation at the speed of a professional simultaneous interpreter, and says Koto delivers exceptional performance in East Asian languages across text-to-speech and speech-to-text.
The model family is organised around three capabilities, set out in Kotoba’s documentation (opens in a new tab). Speech-to-text turns audio into text, in a live streaming mode over WebSocket and in a file mode through a REST API. Live mode supports Japanese, while file mode supports Japanese, English, Korean and Chinese. Speech-to-speech translation ingests audio in one language and emits both an incremental transcript of the source and synthesized speech in the target language over a single connection, with English, Japanese, Korean and Chinese supported. Text-to-speech synthesizes audio from text and streams it back as soon as the model produces it, with Japanese as the supported language and two Japanese speakers, male-01 and female-01.

Generated image, not a photograph of a Kotoba product. Kotoba says Koto runs on-device, including on smartphones and wearables.
An API and SDK alpha for developers
On 28 May 2026 Kotoba announced the alpha release of its API and SDK (opens in a new tab), opening its speech models to developers and enterprises for evaluation. The alpha includes three model families: speech-to-speech translation, streaming text-to-speech and streaming speech-to-text. Kotoba says its core supported languages are Japanese, Chinese, Korean and English, with Spanish also supported, and that the alpha gives builders direct access to the same foundation models trusted by large enterprises across the U.S. and Japan. The Python SDK installs with pip install kotoba-sdk.
Kotoba also publishes benchmark results from its own evaluations. For Japanese agent text-to-speech, it reports a mean opinion score of 4.15 and a correctness score of 0.80 for its own system, compared with scores between 1.99 and 4.12 for four anonymised competing systems. For speech-to-speech translation, it says its April 2026 model leads a competitor’s translation API on Japanese to English and Korean to Japanese quality, and is faster on English to Japanese and Korean to Japanese. Acquiry presents these as Kotoba’s own results from its own tests and has not independently verified them.
Deployment options for enterprise customers
Kotoba lists three ways to deploy Koto. The hosted API suits teams that want to evaluate or build quickly. On-premise deployment serves customers with data residency, security or compliance requirements that rule out public cloud inference, which Kotoba says are common across finance, healthcare, government and large enterprise IT in Japan and the U.S. Select models are also available in on-device form factors for mobile, automotive, robotics and other edge use cases where latency, privacy or offline operation matter. In its June 2026 funding announcement, Kotoba says Koto already runs on-device with its enterprise customers across Asia and the US.
Koto already powers Japanese calls in Agentforce Voice and is rolling out through Cresta
Two named deployments show where Koto is used: Salesforce’s Agentforce Voice in Japan and Cresta’s contact centre platform.
Kotoba announced on 10 August 2026 (opens in a new tab) that Koto had been selected to power Japanese-language conversations in Agentforce Voice, the voice capability of Salesforce’s autonomous AI agent platform, Agentforce, provided in Japan by Salesforce Japan Co., Ltd. Agentforce Voice became available in Japan on August 7, 2026. Kotoba describes it as a native voice layer that transforms conventional IVR (interactive voice response) systems, enabling AI agents to handle customer interactions autonomously over the phone, and says that with Koto as the foundation it delivers fluid, high-quality dialogue that handles honorific speech and the idioms of everyday spoken Japanese.
In Japan, where the shortage of contact center staff continues to intensify, Agentforce Voice delivers a more seamless customer experience by allowing AI agents to handle phone interactions.
Mito adds that Salesforce adopted Koto to achieve natural, high-quality voice conversation in Japanese, and that it will continue working with Kotoba to help customers in Japan create new customer experiences built on collaboration between people and AI agents. Kotoba notes that it is also a portfolio company of Salesforce Ventures, and that the two companies continue to deepen their relationship on both the investment and business fronts.

Generated image, not a photograph of any customer’s premises. Agentforce Voice is described as handling customer interactions over the phone.
Cresta brings Kotoba’s speech models to contact-centre agents
On 11 November 2026 Kotoba announced a partnership with Cresta (opens in a new tab), which it describes as the AI operating system for customer experience. Kotoba will provide its speech-to-text and text-to-speech models to Cresta, which plans to integrate them into its AI platform. Speech-to-text transcribes callers in real time, powering AI agent responses, live assistance for human agents and conversation analytics. Text-to-speech voices AI agent responses instantly, with natural intonation and pacing. The models cover Japanese, Korean, Chinese, English and more, are rolling out in phases during 2026 and are offered to Cresta’s enterprise customers through the Cresta platform.
Ping Wu, CEO of Cresta, says delivering great customer experiences in Japan requires AI that understands Japanese, not just what is said but the nuances of how it is said, and that integrating Kotoba’s Japanese voice models will bring natural conversational experiences to Cresta’s customers in Japan. Kotoba says the partnership opens the door to automating calls such as inquiry intake, identity verification, bookings and changes, and FAQs while keeping the conversation natural, and that it plans to deepen the partnership by expanding supported languages and use cases and tailoring the experience to specific industries.
Honorifics, unstated subjects and pauses make Japanese a demanding language for voice AI
Both companies explain, in their own words, why Japanese needs dedicated research.
The rationale in both announcements starts with language itself. ElevenLabs writes that the way people speak and interact is deeply local, governed by a complex, often unstated mix of social and cultural norms, and that there are few places where local nuance is more important than Japan. It describes sentences that do not always contain a subject and an object, vocabulary that changes depending on who you are speaking to, apologies that are common because they carry little liability, and pauses that signal thought rather than awkwardness. It also points to the Japanese phrase kūki o yomu, or “reading the air,” which means people are expected to pick up what isn’t said.

Generated image, not a photograph of any party’s people. Levels of politeness shift depending on who is speaking to whom, as Kotoba describes.
Kotoba describes the same challenge from the engineering side. In its announcement it says conversations leave subjects unstated, levels of politeness shift depending on who is speaking to whom, and brief acknowledgments and pauses carry meaning. For voice AI to feel natural in everyday life and at work, it says, accurate speech recognition and generation are only the beginning: the system must also understand context, follow the flow of a conversation and know when to respond.
The telephone adds its own requirements
In its Agentforce Voice announcement, Kotoba says Japanese is not an easy language for voice AI because of honorific registers, idioms particular to spoken conversation and meaning that shifts with context, and that delivering conversations that hold up in a business setting requires ultra-low-latency, high-accuracy speech recognition and speech synthesis built on a speech foundation model that reflects how Japanese actually works. In its Cresta announcement it adds that Japanese brings its own challenges, from honorific speech to an abundance of homophones, on top of the demands of line quality and turn-taking that apply to any phone call.
Research rooted in local understanding, supported by global resources
Kotoba’s founders present this as the reason for joining a larger platform. They write that making natural interaction possible in each language takes teams that understand the nuances of local languages and cultures, respond to regional needs and move quickly, and that connecting research rooted in that local understanding with resources on a global scale is exactly the approach Kotoba has pursued. Mati Staniszewski makes the matching point from the ElevenLabs side: the language people speak and the social norms they follow need to be reflected in the technology’s architecture from the very start.
Part III of III
The buyer and the road ahead
ElevenLabs’ growth, its platform and the steps both companies have named.
ElevenLabs reached a $22 billion valuation and more than 800 people before signing Kotoba
A year of company milestones: a Series D, $500 million in ARR, an employee tender and a growing team.
ElevenLabs announced a $500 million Series D (opens in a new tab) on 4 February 2026, led by Sequoia Capital with Andrew Reed joining the board, at an $11 billion valuation that it says is more than three times its valuation one year earlier. Andreessen Horowitz and ICONIQ participated, and ElevenLabs said the funding brought total funding to $781 million across five rounds since its founding in 2022. It said it would use the funding to continue to expand research and product and its international expansion, with locally embedded go-to-market teams in cities that include Tokyo.
On 5 May 2026 ElevenLabs announced the third close of its Series D (opens in a new tab) and said its business was accelerating: it ended 2025 with $350 million in ARR and, in the first four months of 2026, had already surpassed $500 million ARR. It said this growth is driven by enterprises deploying voice agents across their businesses, from customer support and sales to hiring and marketing operations. New institutional investors named in that announcement include BlackRock, Wellington, D.E. Shaw and Schroders, and the company said it then had 530 teammates across over 50 countries.

Generated image, not a photograph of ElevenLabs’ people or offices. ElevenLabs says there are now over 800 people at the company.
On 30 September 2026 ElevenLabs said it had closed a $300 million employee tender offer (opens in a new tab) that values the company at $22 billion, double its valuation at the Series D in February. The round was led by Wellington and T. Rowe Price, and BDT & MSD, EQT, GIC, Goldman Sachs, OTPP and Sapphire Ventures invested in ElevenLabs for the first time. The company said there are now over 800 people at ElevenLabs, and that it is committed to giving its team regular opportunities for liquidity so they can participate in the value they are creating.
The workbook sets two of these changes side by side
The data workbook later in this report computes how these published figures relate. Taking the Series D valuation and the September tender valuation as inputs, it computes the step-up between them. Taking the ARR figure at the end of 2025 and the figure for the first four months of 2026, it computes the growth rate. Both calculations use only the companies’ own numbers and link back to their sources.
ElevenAgents handles more than 15 million conversations a week, up 3x since February
ElevenLabs describes a full-stack approach across research, product and deployment, and a platform used across industries.
In its September announcement, ElevenLabs describes a platform that tackles voice AI at three levels. At the research level it builds AI models that speak, listen and translate at a human level, in over 90 languages spoken natively by more than 5.5 billion people. At the product level, its fastest-growing product is ElevenAgents, its conversational agents platform. At the deployment level, its forward-deployed engineers and researchers work with organizations to orchestrate and refine models and agents for their specific conditions. By owning the full stack, it says, model improvements reach customers immediately, and feedback from real deployments allows it to continually refine both the models and the products.
The company shares usage figures to show the scale. It says its agents now handle more than 15 million conversations every week, up 3x since February, and that ElevenAgents’ ARR has increased more than 3x over the same period. It says enterprise now accounts for 55% of its revenue, and that its technology is used in daily operations at five of the world’s ten largest tech companies, five of the ten largest insurers and four of the ten largest telecoms. Organizations it names as having rolled out ElevenAgents since February include Admiral, Customers Bank and Cadence, along with the governments of Ukraine and Greece.

Generated image, not a photograph of the ElevenLabs Summit. The summit took place in New York on 11 November 2026.
The New York Summit set out the interaction layer
On 11 November 2026, at its New York Summit, ElevenLabs introduced new models, agents and deployment options. Staniszewski said that within a few years conversational agents will become critical infrastructure for every enterprise, as essential as cloud or payments. The company previewed Interaction Engine, which handles turn-taking and interruptions and can pick the right language model for each part of an exchange, and described Memory and Routines, two updates coming to ElevenAgents, along with ElevenAI, an app that works as an AI teammate for enterprise teams. It also said its partner network has grown to more than 550 partners in 44 countries, and launched the ElevenLabs Partner Program and the ElevenLabs Marketplace.
The summit announcements returned often to local nuance. Staniszewski said businesses need models that reflect what makes people and businesses distinctive, and ElevenLabs said a domain-optimized model for the financial services industry is in the works. It also said it is adding 15 MW of new Vera Rubin compute capacity in an expanded partnership with NVIDIA, which relates to the large-scale computing resources that Kotoba says it will draw on after joining.
Local sales leaders, infrastructure in Japan and Korean language support are the named next steps
ElevenLabs and Kotoba each name concrete steps after the close.
Both companies name the steps that follow the announcement. ElevenLabs says this is just the first step in its long-term plan to make AI intuitive and trusted for everyday interactions across Japan. In the coming months it plans to build out local sales leaders and continue to invest in infrastructure in Japan, and it looks forward to partnering even more closely with Japanese enterprises and with multinationals operating and expanding in Japan.

Generated image, not a photograph of a specific location. ElevenLabs says it plans to continue to invest in infrastructure in Japan.
Kotoba’s announcement ends with its own list of ambitions. The team says it will improve voice AI across East Asian languages, with Japanese first and Korean next, accelerate adoption in products and services, and contribute research to the models and products ElevenLabs delivers worldwide. Its founders thank their customers, users, investors, partners, friends and families, and say that together with ElevenLabs they will build the next generation of voice AI in Japan and bring it to the world.
What Kotoba’s app and API users can follow
Kotoba’s products continue to be described on its own site. The Kotoba app, which Kotoba says has surpassed 180,000 users, delivers simultaneous translation, note-taking and AI summaries across 21 languages with five primary target languages, and in June 2026 Kotoba shipped an update with 11 new features. Its API and SDK are in alpha for selected customers who request access. Readers who use these products can follow Kotoba’s news page and the ElevenLabs blog for updates around the close.
The numbers behind the deal
18 sourced figures, each labelled by source type, and 7 calculations built from them. Every calculation shows its formula and inputs.
- Months from last funding round to announcement
- 5
- Years from founding to announcement
- 3 years
- ElevenLabs valuation change from the Series D to the employee tender
- 100%
Open the full workbook
Acquiry calculations
| Measure | Result | Formula and why it matters |
|---|---|---|
| Months from last funding round to announcement | 5 | Days between the two dates ÷ 30.44How long the target stayed independent after its last round. |
| Years from founding to announcement | 3 years | Announcement year − founding yearTime to outcome for founders and early investors. |
| ElevenLabs valuation change from the Series D to the employee tender | 100% | (ElevenLabs valuation at the employee tender ÷ ElevenLabs valuation at the Series D − 1) × 100Shows how ElevenLabs’ own published valuation moved between its February and September 2026 announcements. |
| ElevenLabs ARR growth from the end of 2025 to the first four months of 2026 | 42.9% | (ElevenLabs ARR surpassed in the first four months of 2026 ÷ ElevenLabs ARR at the end of 2025 − 1) × 100The commercial momentum the buyer brings to the combination. |
| ElevenLabs team growth between the Series D third close and the September tender | 50.9% | (ElevenLabs people ÷ ElevenLabs teammates at the Series D third close − 1) × 100Headcount growth shows how quickly the buyer is scaling its teams. |
| Share of Kotoba total funding raised in its latest round | 43.5% | Kotoba latest funding round (additional seed funding) ÷ Kotoba total funding to date × 100Shows how much of Kotoba’s capital came in its most recent seed round. |
| ElevenLabs Series D relative to Kotoba total funding to date | 21.7x | ElevenLabs Series D round ÷ Kotoba total funding to dateCompares the scale of the buyer’s single largest round with the target’s whole funding history. |
Sourced inputs
| Figure | Value | Source |
|---|---|---|
| Announcement date | 16 November 2026 | 1. ElevenLabsCompany announcement |
| Kotoba founded | 2023 | 4. Kotoba Technologies (Business Wire release)Press release |
| Kotoba latest funding round (additional seed funding) | USD 10 million | 3. Kotoba TechnologiesCompany announcement |
| Date of Kotoba latest funding round announcement | 24 June 2026 | 3. Kotoba TechnologiesCompany announcement |
| Kotoba total funding to date | USD 23 million | 3. Kotoba TechnologiesCompany announcement |
| ElevenLabs people | over 800 people | 8. ElevenLabsCompany announcement |
| Kotoba Seed 2 round | USD 11.83 million | 4. Kotoba Technologies (Business Wire release)Press release |
| Kotoba app users | 180,000 users | 3. Kotoba TechnologiesCompany announcement |
| ElevenLabs Series D round | $500 million | 10. ElevenLabsCompany announcement |
| ElevenLabs valuation at the Series D | $11 billion | 10. ElevenLabsCompany announcement |
| ElevenLabs valuation at the employee tender | $22 billion | 8. ElevenLabsCompany announcement |
| ElevenLabs employee tender offer | $300 million | 8. ElevenLabsCompany announcement |
| ElevenLabs ARR at the end of 2025 | $350 million | 9. ElevenLabsCompany announcement |
| ElevenLabs ARR surpassed in the first four months of 2026 | $500 million | 9. ElevenLabsCompany announcement |
| ElevenLabs teammates at the Series D third close | 530 teammates | 9. ElevenLabsCompany announcement |
| ElevenLabs total funding after the Series D | $781 million | 10. ElevenLabsCompany announcement |
| ElevenAgents conversations handled every week | 15 million conversations | 8. ElevenLabsCompany announcement |
| ElevenLabs partners | 550 partners | 11. ElevenLabsCompany announcement |
Calculations use only the sourced inputs above, as stated at publication. Cite as Acquiry Deal Intelligence with the article URL.
Frequently asked questions
What is ElevenLabs acquiring?
ElevenLabs is acquiring Kotoba, the Japanese voice AI lab behind the Koto speech model, which covers speech-to-text, text-to-speech and speech-to-speech translation. Kotoba’s team will continue to work across research, product development and deployment support.
When is the acquisition expected to close?
Both companies say the acquisition is expected to close later in Q4 2026, subject to customary regulatory approvals and closing conditions.
What was the price?
Financial terms were not disclosed.
Who founded Kotoba?
Kotoba was founded in 2023 by AI researchers Noriyuki Kojima, who is CEO, and Jungo Kasai, who is CTO.
Where is Koto already used?
Koto powers Japanese-language conversations in Salesforce’s Agentforce Voice, which became available in Japan on August 7, 2026, and Kotoba announced on 11 November 2026 that it will provide its speech-to-text and text-to-speech models to Cresta.
What does ElevenLabs plan to do in Japan?
ElevenLabs says it plans to build out local sales leaders and continue to invest in infrastructure in Japan in the coming months, and that it looks forward to partnering even more closely with Japanese enterprises.
Methodology and limitations
This analysis is built from the ElevenLabs announcement of 16 November 2026, Kotoba’s own announcement of the same date, Kotoba’s earlier funding, partnership and product announcements, Kotoba’s documentation and website, and ElevenLabs’ announcements of 4 February, 5 May, 30 September and 11 November 2026. The text of each source is saved with the article’s audit trail and every figure was checked against it.
Every figure is labelled by basis. Company-disclosed figures are stated directly by ElevenLabs or Kotoba. Benchmark results are presented as the company’s own results from its own tests. Calculations in the data workbook are computed from published inputs and are labelled as calculated.
Acquiry’s own interpretation is marked as such throughout and is clearly separated from facts stated by the companies. Statements about the future are the companies’ stated intentions, not outcomes.
Independence. Acquiry was not engaged by any party. This is independent research from public sources and is not a solicitation, investment advice, or an offer to buy or sell any security. Acquiry holds no disclosed position in ElevenLabs or Kotoba Technologies, Inc.
Sources
Numbered to match the superscript citations. Every source is a primary company announcement, filing or named publication.
- 1Kotoba is joining ElevenLabs
ElevenLabs · · Company announcement
- 2Kotoba Is Joining ElevenLabs
Kotoba Technologies · · Company announcement
- 3We’ve Raised $10M in Seed Funding to Expand Real-Time Voice AI Across East Asia
Kotoba Technologies · · Company announcement
- 4Kotoba Technologies Raises $11.83 Million in Second Seed Round
Kotoba Technologies (Business Wire release) · · Press release
- 5Natural Japanese, Over the Phone: Koto Powers Agentforce Voice in Japanese
Kotoba Technologies · · Company announcement
- 6Kotoba Partners with Cresta to Bring Real-Time Voice AI to Contact Centers Worldwide
Kotoba Technologies · · Company announcement
- 7Announcing the Kotoba API & SDK Alpha: Speech-to-Speech, Streaming STT, and TTS for Builders
Kotoba Technologies · · Company announcement
- 8ElevenLabs reaches $22bn valuation after employee tender
ElevenLabs · · Company announcement
- 9ElevenLabs crosses $500M ARR and welcomes new investors
ElevenLabs · · Company announcement
- 10ElevenLabs raises $500M Series D at $11B valuation
ElevenLabs · · Company announcement
- 11ElevenLabs Summit NYC: The Interaction Layer for AI
ElevenLabs · · Company announcement
- 12Kotoba Technologies APIs documentation
Kotoba Technologies · · Company website
- 13Kotoba: the foundational voice AI model for a borderless world
Kotoba Technologies · · Company website
Cite this report
Boyton, J. (2026, November 17). ElevenLabs to Acquire Kotoba to Advance Japanese Voice AI. Acquiry Deal Intelligence. https://www.acquiry.com/deal-intelligence/elevenlabs-to-acquire-kotoba/
@online{boyton2026elevenlabskotoba, author = {Boyton, Joash}, title = {ElevenLabs to Acquire Kotoba to Advance Japanese Voice AI}, organization = {Acquiry Deal Intelligence}, date = {2026-11-17}, url = {https://www.acquiry.com/deal-intelligence/elevenlabs-to-acquire-kotoba/} }
Joash Boyton is a technology sector analyst, publisher, and the founder of Acquiry, where he executes buy-side and sell-side M&A mandates across digital assets, software, and gaming technologies. He is the author of peer-reviewed corporate finance literature indexed across institutional repositories including Google Scholar and the ORCID Registry. Joash publishes Acquiry Deal Intelligence to deliver independent, forensic strategic reviews and valuation benchmarks of global technology acquisitions, compiling primary data directly from corporate disclosures, SEC filings, and regulatory ledgers.
Research support: Acquiry Deal Intelligence.
