India proposes lighter rules for AI startups, tighter oversight for global giants, and a shift toward “accountable data portability” instead of forced localisation.
India is rethinking its data governance playbook for the AI era—with an eye on innovation, accountability, and economic value retention.
In the Economic Survey 2025–26, tabled by Finance Minister Nirmala Sitharaman in Parliament today, the government pitched a startup-friendly data framework designed to strike a balance between global data flows and domestic value creation.
Prepared by Chief Economic Advisor V. Anantha Nageswaran’s office, the Survey proposes a future-forward model where data remains portable across borders, but firms processing Indian data at scale must be accountable—especially those building AI foundation models.
Accountable Portability, Not Data Lock-In
India is not mandating strict localisation yet. Instead, it’s proposing “accountable portability”:
- Firms can move Indian data abroad, but must ensure it’s auditable, traceable, and retrievable.
- Those training general-purpose AI models or generating significant revenue from Indian data would be held accountable by Indian regulators.
Why now? Because India’s 1 billion+ internet users generate some of the most diverse, real-world data globally—a strategic asset in AI development.
“India can’t afford to be just the data mine for the world’s AI,” a senior policymaker said. “The value must cycle back.”
Startup Focus: Light Compliance, Big Opportunity
Crucially, the Survey emphasises light-touch compliance for startups and sovereign AI efforts:
- Smaller firms, research labs, and India-focused AI builders would be shielded from the heavy regulatory burdens meant for Big Tech.
- Compliance scales only with sensitivity, volume, or revenue tied to Indian data.
Why this matters: With the GPU crunch already limiting early-stage AI development, regulatory overreach could further choke innovation.
“Let startups breathe,” the Survey implies—until their scale demands scrutiny.
Smarter Categorisation, Smarter Governance
Not all data is created equal. The new framework calls for risk-based data categorisation:
- High-risk data (like large behavioural or inferred datasets) would face greater oversight if used for AI training or monetisation.
- Low-risk data could move freely, reducing friction for research and smaller applications.
The goal: avoid blanket restrictions while securing economically or strategically sensitive data.
Creating Domestic Value Without Lockdown
Rather than walling off Indian data, the Survey proposes economic value capture through contribution, not control:
- Global AI firms could be nudged to train models locally, fund AI research, or support Indian skilling and compute infrastructure.
- Public data trusts, AI sandboxing, and compute sharing could help build a sovereign data ecosystem without harming ease of doing business.
And there’s a carrot, not just a stick. Companies operating within certified Indian data environments may get:
- Faster regulatory clearances
- Fewer audits
- Early access to national AI programmes
Human Data is the New Oil—And It’s Drying Up
The Survey also highlights an emerging bottleneck in AI training: the shortage of fresh, high-quality human-generated data.
As synthetic data proliferates, its limits are becoming clear. Human interactions, behaviors, and decisions are still crucial to train trustworthy AI models.
Given India’s population and digital intensity, its data is becoming more valuable—and must be governed accordingly, the Survey argues.
Broader Digital Ecosystem Signals
- The Survey flags the GPU shortage as a bigger hurdle than funding for AI startups—potentially spurring new domestic GPU manufacturing or compute incentives.
- It also calls for age-based access controls on social media and digital platforms, citing concerns around digital addiction and youth exposure.
TL;DR
India’s Economic Survey 2025–26 proposes a progressive, startup-friendly data governance framework. It favors cross-border “accountable portability” over localisation, smarter risk-based regulation, and ensures AI firms building on Indian data contribute back—without stifling innovation or growth.
AI summary
- New data framework supports AI innovation, cross-border flow
- Startups, labs face light compliance; Big Tech held accountable
- High-risk data sees tighter rules; low-risk data flows freely
- Firms monetising Indian data may fund local AI ecosystem
- Human data scarcity, GPU crunch flagged as urgent issues
