Activate, a $75M AI-focused fund launched by former Haptik CEO Aakrit Vaish and Together Fund’s Pratyush Choudhury, is backing early-stage Indian startups building core and applied AI — with NVIDIA as a strategic partner announced in February 2026. Arkam Ventures published “India AI: The Asymmetric Opportunity” in March 2026, identifying four structural pillars including population-scale consumer AI and B2B applications. The signal is clear: institutional capital is shifting toward India’s AI application layer — and most Indian VCs are still funding infrastructure.

The Pattern

The global AI narrative from 2024 to 2026 split into two camps. Camp 1 (US-Europe): build better foundation models and training infrastructure. Camp 2 (Asia-Pacific): build applications on commodity models. Camp 2 is winning. Foundation models are now a commodity good — Anthropic, OpenAI, and Google compete on price. Margin and moat moved to applications, not to chips or model weights.

India has 958 million active internet users according to IAMAI’s 2025 report. GDP stands at $4.15 trillion with IMF-projected growth of 6.2% for 2026. Hurun’s 2025 report counts 73 unicorn startups worth over $300 billion. But examine the India AI landscape and you find a missing segment: there are few significant India-based AI application companies at scale. The funded cohort skews toward infrastructure — Sarvam AI (sovereign language models, raising $300-350M at ~$1.5B valuation), Dubpro.ai (formerly Deepsync; AI video dubbing), Veri5Digital (AI-powered KYC verification). Vertical AI applications in agriculture, healthcare, fintech, and operations remain largely unfunded at meaningful scale.

Three structural reasons explain the gap.

Venture capital misalignment. Indian VCs learned from the 2014–2020 cycle when marketplace and fintech apps scaled to unicorn status. They’re applying the same playbook to AI: build a platform that every industry can use. The problem: most industries don’t want a generic platform; they want a vertical-specific solution — AI-powered farm advisory, AI-powered clinic triage, AI-powered supply chain planning. The platform bet assumes a category that doesn’t exist yet in India.

Talent distribution. The best AI product engineers in India are split roughly between MAANG (captured by US salaries), India fintech, and AI startups. There isn’t enough product-focused AI talent to build verticals at scale — yet. The pipeline is filling, but the funding is ahead of the talent.

Go-to-market mismatch. Enterprise decision-making in India is slower than in the US. A US SaaS founder can land a $10K/month enterprise contract in 60 days. An India-based founder is looking at 120–180 days to the same result. Venture economics only work if the team is exceptionally capital-efficient.

What’s Changing

Three signals indicate the window is opening.

Activate’s $75M debut fund plans to back 25-30 AI startups at $500K to $3M per investment. Its LP roster — including Aravind Srinivas (Perplexity), Vinod Khosla, and Ronnie Screwvala — signals conviction from operators, not just allocators. When a fund of this size makes a thesis-level commitment, deal flow follows the money. Founders who were building elsewhere will now build India AI applications.

The pace of AI startup formation is accelerating. Events like the BuildAI Pitch Event under the IndiaAI Mission and the India AI Impact Summit 2026 are drawing growing cohorts of AI-first founding teams. The rate of change matters more than any single snapshot — when founders see a path to capital, they build toward it.

Arkam’s report is the third signal. Arkam is one of the more analytically rigorous India VCs. Their four-pillar thesis — population-scale consumer AI, B2B AI for India, India-to-world AI services, and indigenous infrastructure — names the application layer explicitly as the primary opportunity. When a firm of their calibre publishes a thesis-level report, other VCs follow or get outmanoeuvred.

The Charaka View

The 18–24 month window before fully-funded international teams ship India-specific products is real but not guaranteed. The playbook for founders: identify a vertical (agriculture, micro-credit, telemedicine, supply chain), identify the most inefficient manual process within it, build a vertical-specific AI agent that solves that one process, and price at 20–30% below incumbent cost while delivering measurably superior quality.

For India VCs: every rupee allocated to infrastructure competes with OpenAI, Anthropic, and Google at scale where they have structural advantages. Every rupee allocated to vertical applications competes with founders who haven’t been born yet and international teams who don’t understand India’s market structure. The asymmetry is obvious. The question is whether the capital is fast enough to capture it.

This analysis draws on Analytics India Magazine, TechCrunch, Business Standard / IAMAI, Outlook Business / Hurun India, StartupNews / Arkam Ventures, Inc42, and IMF / Worldometer. Human editorial oversight applied.

This analysis is informational and does not constitute investment advice, a research report, or a recommendation to buy, sell, or hold any security.

Charaka Notes by Manthan Intelligence. Subscribe