India’s Indic AI startups raise $687 million, funding remains concentrated: Tracxn

India’s 34 companies building artificial intelligence (AI) for Indian languages have raised $687 million in cumulative equity funding, but capital remains concentrated among a handful of players, with the ecosystem yet to record a late-stage funding round, according to a report by market intelligence platform Tracxn.

The Indic AI: Building AI for Bharat report, based on data as of October 15, 2026, found that the companies raised $657 million across 56 rounds since 2021. The funding surge has been uneven, with a few large transactions accounting for a key share of the capital raised.

Sarvam drives 2026 funding surge

While 2024 recorded the highest number of funding rounds at 15, 2026 saw the largest capital inflow, driven primarily by two funding rounds raised by Sarvam. Excluding these transactions, the remaining companies raised $54 million across seven rounds during the year.

The report highlights how the ecosystem expanded in terms of the number of companies before funding sizes increased. However, all rounds recorded during the period were at the seed or early stages, indicating that the sector is yet to transition to late-stage financing.

The report also flags the risks of dependence on foreign AI providers. The suspension of access to Anthropic’s Fable 5 and Mythos 5 between June 12 and July 1 was cited as a recent example of the challenges associated with relying on external AI models and infrastructure.

Two distinct business models emerge

Tracxn divides the 34 companies into two categories based on whether they would retain a product offering without their Indic-language capabilities.

The first category comprises 18 companies for which language technology is the core product. These businesses develop speech recognition, translation and language models, or datasets used to train them. Sarvam, Gnani, Soket AI and BharatGen are among the companies in this segment.

The remaining 16 companies use Indic-language capabilities as a differentiator within broader enterprise offerings. These businesses would continue to have products even without the language layer. Skit.ai, Nuplay AI and GreyLabs are among them.

GreyLabs, which provides voice AI solutions to financial institutions, serves more than 50 institutions, including RBL Bank, AU Small Finance Bank, and IDFC FIRST Bank. Its customer base indicates growing enterprise demand for Indic-language voice AI.

Sarvam and Krutrim pursue different strategies

Sarvam and Krutrim, both founded in 2023, have achieved unicorn status and together account for $464 million in funding.

Sarvam has raised $389 million to build a full-stack sovereign generative AI platform. It is also among the four companies selected by the government to develop sovereign AI models. Its early language-model development adapted Meta’s Llama and Mistral’s open models rather than training models entirely from scratch.

Krutrim has raised $75 million and partly funded its expansion through debt collateralised against founder Bhavish Aggarwal’s Ola Electric shares. After pausing chip-design efforts in late 2025, it shifted from foundation-model development towards AI cloud infrastructure and has reported its first profit. The company no longer develops Indic AI models but remains in the report’s dataset because of its funding history.

India’s venture-led model faces global competition

The report identifies two acquisitions in the ecosystem, both involving Reliance group companies in 2019. Reliance Jio acquired approximately 87% of Mumbai-based conversational AI company Haptik in a transaction valued at $101 million, including growth and expansion investment. Reliance Industrial Investments & Holdings acquired a majority stake in Bengaluru-based Reverie Language Technologies for about $27 million.

Tracxn contrasts India’s investor-led ecosystem with state-backed initiatives elsewhere. Singapore’s SEA-LION programme has received S$70 million in funding while Africa’s LINGUA initiative offers grants of up to $250,000 per project across 26 projects. Saudi Arabia’s HUMAIN and the UAE’s G42 and Technology Innovation Institute represent other state-backed approaches.

India’s $687 million in equity funding compares with an estimated $77 billion cost for Saudi Arabia’s HUMAIN project, although the figures measure different types of investment.

The next phase will depend on how Indian startups compete with global providers. Google’s Gemini supports nine Indian languages while OpenAI has released an Indian-language benchmark developed with 261 domain experts. The first late-stage funding round in the ecosystem will be a key indicator of whether companies can move from building language models to scaling commercially viable businesses.