People · AI & Deep Tech

Sarvam appoints Devendra Singh Chaplot as advisor

By Startup Enthusiast · 30 July 2026

What they do

Sarvam builds AI models, infrastructure, and enterprise products for Indian languages

It appointed Devendra Singh Chaplot, a Mistral AI founding team member, as advisor

The company plans to build a trillion-parameter frontier AI model in India within six months

Sarvam appointed Devendra Singh Chaplot as an advisor at its Epoch 2026 developer conference.
Chaplot was a founding team member at Mistral AI and worked at Thinking Machines Lab and xAI.
He contributed to models like Mistral 7B, Mixtral 8x7B, and Mistral Large.
Sarvam plans to build a trillion-plus parameter frontier AI model from scratch in India.
The company is building an end-to-end AI stack including models, infrastructure, and enterprise software.
Sarvam opened a San Francisco office to support its global ambitions.
Chaplot said a model with ~3 trillion total parameters and ~100 billion active parameters could be pre-trained in ~2 months using 10,000 Blackwell GPUs.
He argued India needs its own models for linguistic, cultural, educational, legal, and public policy needs.

Sarvam became India's 130th unicorn about a month before this announcement, after raising $300 Mn in a Series B round at a $1.5 Bn valuation.
It raised $234 Mn as part of a $300 Mn Series B round at a $1.5 Bn post-money valuation.
The round was led by HCLTech ($150 Mn commitment) with participation from Bessemer, Khosla, and Peak XV.
Funds will be used for next-gen models (agentic AI, coding, cybersecurity), compute infrastructure, and enterprise deployments.
Sarvam's FY26 revenue was ₹45.1 Cr, up from ₹1.5 Cr in FY25.

Sarvam develops AI models, infrastructure, and enterprise products across language, speech, vision, and document intelligence.
Its model lineage includes OpenHathi-v1, Sarvam-2B, Sarvam-1, Sarvam-30B, and Sarvam-105B (enterprise reasoning, multi-turn conversations, long-document understanding in 22+ Indian languages).
AI products are used in banking, insurance, government, and defence.
Its conversational AI platform handles 2M+ interactions daily; inference platform processes 10M+ API calls daily.
Products include Voice Agents, Text to Speech (11 Indic languages), Speech to Text (12 Indic languages), Translation (23 languages), and Document Digitisation.
APIs offered via REST API, Python SDK (`pip install sarvamai`), and Playground.
It has served 10B+ tokens with median latency <100ms and 99.9% uptime SLA.
Deployment options: Sarvam Cloud, Private Cloud (VPC), and On-Premises (air-gapped for regulated industries).
Compliant with SOC 2 Type II, ISO 27001, and DPDP; provides role-based access, audit trail, and data residency controls.

Use Series B funds for next-gen models (agentic AI, coding, cybersecurity), compute infrastructure, and enterprise deployments.

Product: Foundation models and agentic AI systems with voice capabilities and enterprise applications
Customers: Enterprise customers across finance, aviation, telecom and government sectors
How it makes money: SaaS and licensing model for AI models and enterprise AI applications
What sets it apart: India-built sovereign AI stack with focus on cost-effective deployment and localization