Sarvam dominates Indic AI funding in 2026 with 87% share: Tracxn report

India’s frontier model company Sarvam bagged almost 90 per cent of the total funding received by the country’s startups building artificial intelligence (AI) for India’s languages in 2026, according to a report by data-intelligence platform Tracxn.

Sarvam, which received two rounds of funding of $255 million and $90 million earlier this year, thus accounted for $345 million, or 87 per cent, of the $399 million raised by Indic AI companies. Excluding Sarvam, the other companies raised $54 million across seven rounds, up from $23 million across nine rounds last year.

Between October 2021 and 2026, companies of the Indic AI ecosystem raised a cumulative $657 million across 56 rounds. The funding style shows investors gravitating towards a select few companies that they believe have the potential to score big. Besides Sarvam, Gnani AI—whose speech recognition is trained on a decade of real Indian call centre audio—closed a $4.2 million funding round in July.

The report says funding for sovereign AI models received a boost as the US government on June 12 this year issued an export-control order against Anthropic’s Fable 5 and Mythos 5 models days after their public launch and suspended access to both. That raised questions of AI sovereignty and whether US frontier model companies were becoming too big a tool in the hands of the Donald Trump administration to wield undue influence.

“The shutdown turned an abstract argument about foreign AI dependency into a concrete one: frontier access could be revoked by a foreign government’s order, without warning, although access was restored. Sarvam, one of the government’s own four sovereign model picks, raised twice in the eight weeks that followed (the US government action),” the report added.

Besides Sarvam and Gnani, Soket will develop India's first open-source 120 billion parameter foundation model optimised for the country's linguistic diversity, targeting sectors such as defence, healthcare, and education. And Gan AI will create a 70 billion parameter multilingual foundation model targeting text-to-speech capabilities.

Sarvam, backed by HCLTech and Nvidia, is expected to launch a trillion parameter model in the next six months, a significant leap from its existing billion parameter models. The company also plans to open an office in San Francisco to tap AI talent.

The report covers 34 companies, which are divided into two categories: language as a product and language as a differentiator. Sarvam, Gnani, and BharatGen fall into category one, while Skit.AI, Nuplay, and GreyLabs are in the second category.

Two developments, the report explains, will shape the next phase of growth. Global model providers are moving directly into the language gap these companies were formed to address: Gemini natively supports nine Indian languages and is training on more than 100, and OpenAI has released a dedicated Indian language benchmark built with 261 domain experts across 12 languages. The product tier competes most directly with that push, while the differentiator tier—whose value rests on enterprise workflow and customer relationships—may be less exposed to that competition.

“The first late-stage round in the set will signal that at least one company has cleared the threshold from building to scaling,” the report said.