ElevenLabs invests hundreds of millions in India as the voice-AI company builds local teams, develops models for Indian users and prepares support for all 22 official languages. Co-founder and chief executive Mati Staniszewski disclosed the commitment on October 6 but did not give a precise total or spending timetable.

 

The expansion has four concrete parts:

  • A nine-figure investment in Indian operations
  • A target of 100 local employees
  • Voice services across all 22 official languages
  • A Karnataka partnership on services, safety and voice restoration

 

Related Research

 

ElevenLabs India Investment Expands Local Voice AI

Reuters reported that the money will fund additional local teams, model development and work on Indian languages. ElevenLabs is also open to acquisitions in the country, although Staniszewski did not identify targets or say how much of the commitment could be used for deals.

 

The company already works with more than 250 businesses and enterprises in India. Named customers include Razorpay, TVS Motor, Urban Company, PhysicsWallah and Havells, spanning financial services, transport, consumer services, education and electrical equipment.

 

Indian businesses recorded about 100 million conversations on the ElevenAgents platform over the past year. More than 70% were in Hindi, Kannada, Tamil and Telugu, according to figures presented at the company’s first India summit in Bengaluru.

 

ElevenLabs currently supports speech-to-text and text-to-speech services in 14 Indian languages. The company plans to cover all 22 languages recognized in the Eighth Schedule of India’s Constitution within the next year, a substantial localization task involving accents, dialects, code-switching and limited training data for some languages.

 

Karnataka Partnership Adds Public-Service Pilots

The Karnataka Innovation and Technology Society signed a memorandum of understanding with ElevenLabs on October 6. It is the company’s first collaboration with an Indian state government and is designed to identify pilots rather than award a procurement contract.

 

The parties will explore voice interfaces for citizen helplines, grievance redressal, access to government information and entitlements. Kannada is central to the initial work, although the agreement leaves room for other languages used across the state.

 

A second workstream focuses on people who have lost the ability to speak because of oral or throat cancer, ALS and other neurological or degenerative conditions. Eligible participants could receive synthetic versions of their own voices, including Kannada support, through ElevenLabs’ One Million Voices program.

 

KITS is engaging the All India Institute of Speech and Hearing in Mysuru to examine implementation and identify beneficiaries. The memorandum is voluntary and non-binding, and specific projects remain subject to separate agreements, approvals and applicable law.

 

Karnataka Voice AI Safety Framework Takes Shape

The partnership also covers a platform-agnostic safety framework for synthetic speech. Planned areas include deepfake detection, audio provenance, watermarking, consent for voice cloning, restrictions involving public figures, and protection against voice fraud and impersonation.

 

That policy work is important because better speech generation has two effects at once. It can make digital services accessible to people who struggle with text-based interfaces, while also making fraudulent audio more convincing and harder for citizens or officials to authenticate.

 

The parties plan knowledge-sharing and training for government officials and agencies. They will also consider safeguards for public-facing representatives and institutions whose voices could be cloned without authorization.

 

Karnataka’s approach gives ElevenLabs a public-sector test bed, but the safeguards are not yet final rules. The partners have not published technical standards, audit requirements, pilot dates or success measures, leaving those decisions for the next stage.

 

India Tests the Economics of Multilingual Voice Agents

The investment follows a rapid rise in ElevenLabs’ commercial scale. The company reached a $22 billion valuation in a September employee tender, double its February valuation, and has reported a global annualized revenue run rate above $600 million.

 

India’s appeal is not limited to its population. A large base of feature-phone users, uneven literacy, many active languages and extensive call-center operations create conditions where voice can be more practical than another app or text chatbot.

 

The difficult part is execution. Serving 22 official languages requires more than translating a prompt: models must handle regional pronunciation, mixed-language speech, names, noisy phone audio and culturally specific phrasing without weakening consent and anti-fraud controls.

 

The announced spending establishes ambition, not a guaranteed outcome. Investors and customers will need clearer milestones for capital deployment, hiring, language coverage and public-service pilots. The most useful evidence will be measured accuracy, latency, safety performance and adoption across languages beyond India’s largest commercial markets.