India's AI Startups Are Raising Real Money Now — Here's Who Got Funded This Week

The narrative about India's startup ecosystem has always arrived a few years late. The country produced the world's largest fintech by users (PhonePe), one of the most sophisticated government digital infrastructure stacks on earth (AADHAAR + UPI), and a generation of engineers who built large portions of the internet — and Western venture capital treated all of it as a footnote. The same pattern is repeating with AI, and the same mistake is being made.

This week's Indian startup funding data tells a story worth paying attention to.

The headline number from March 26, 2026: Deccan AI raised $25 million in a Series A led by A91 Partners, with participation from Susquehanna International Group (SIG) and Prosus Ventures. Deccan AI — based in Hyderabad — specializes in AI post-training: data generation, evaluation pipelines, and RLHF (reinforcement learning from human feedback). Their clients include Google DeepMind and Snowflake. The company was founded in 2024 and already has 10+ enterprise customers with active projects.

This is not a speculative bet. RLHF and post-training data infrastructure are among the most valuable services in the AI supply chain right now. Every frontier model being built — by OpenAI, Anthropic, Google DeepMind, Meta — requires massive volumes of high-quality human-generated training data and evaluation. Deccan AI sits directly in that supply chain, and they are building it out of Hyderabad.

The same day, Bengaluru-based Plum raised ₹193 crore (approximately $23 million) in a Series B led by Peak XV Partners. Plum is an insurtech platform focused on employee health benefits — not a pure AI play, but a company that has explicitly flagged AI-driven claims operations as a core investment area for the fresh capital. Peak XV (formerly Sequoia India) does not write cheques of this size into companies they do not believe will scale significantly.

Zoom out from the single week and the picture gets more interesting. Earlier in March, Kyvex — an AI infrastructure company based in Telangana — closed an angel round. Mayson, a generative AI startup from Uttar Pradesh, closed a seed round. These are not Bangalore companies. The geographic distribution of Indian AI investment is expanding beyond the traditional tech corridor.

The macro context matters here. Global startup funding into India dropped 22.93% in early 2026 compared to the same period in 2025, according to Tracxn data. Most sectors contracted. AI did not. The AI-specific funding buckets — post-training, inference infrastructure, vertical AI applications — held up while everything else pulled back. That is not a coincidence. It reflects where the global capital is flowing, and India is capturing a meaningful slice of it.

What is being built in India that does not get built in San Francisco is different in kind, not just in cost. The Deccan AI story is instructive: RLHF data generation at scale, in multiple Indian languages, with cultural context that a San Francisco team cannot replicate at any price. The question of who trains the next generation of LLMs — and in whose languages and value systems — is one of the most important questions in technology right now. India is actively positioning itself to be part of that answer.

Sam Altman has visited India twice in the last year. Dario Amodei has spoken publicly about Anthropic's interest in non-English language capabilities. The interest is real. The question is whether Indian founders build the infrastructure layer themselves before American companies commoditize it. This week's funding suggests they are moving.

Deep Dive

See our previous coverage on Bangalore's rise: Why Bangalore Is Becoming the AI Capital the West Refuses to Take Seriously

See also: Ilya Sutskever Raised $2B With No Product — Here's Why That Makes Sense