Analytics

India proposes zero tax on export cloud revenue run from Indian data centers until 2047

India wants hyperscalers to run overseas cloud workloads from Indian data centers by offering a tax holiday through 2047, while warning signs on power and water could cap real capacity....

India proposes zero tax on export cloud revenue run from Indian data centers until 2047
Feb 1, 2026
2 min read
By Emma Wilson

Key Takeaways

  • India proposed a tax holiday through 2047 for cloud revenue earned outside India if workloads run from Indian data centers.
  • Domestic sales must go through locally incorporated resellers, keeping Indian-customer revenue within India’s tax net.
  • A 15 percent cost-plus safe harbour would simplify transfer pricing for related-party data-center services.
  • Hyperscaler commitments cited include 15 billion dollars (Google), 17.5 billion dollars (Microsoft), and 75 billion dollars total planned (Amazon).
  • Power availability and water scarcity remain key constraints for AI-heavy data center expansion, per Reuters.

India is trying to pull more global compute into the country by making it dramatically cheaper to serve international customers from Indian data centers—an aggressive move as AI workloads push cloud infrastructure demand higher.

Tax holiday targets export cloud services and related-party operations

In India’s latest federal budget, finance minister Nirmala Sitharaman proposed a tax holiday that would effectively bring taxes to zero on revenue from cloud services sold outside India, as long as those workloads run from data centers located in India, according to the official budget speech (indiabudget.gov.in).

The proposal includes a structural constraint for go-to-market: sales to Indian customers would need to be routed through locally incorporated resellers and taxed domestically. For operators building and running the infrastructure, the budget also outlines a 15 percent cost-plus safe harbour for Indian data-center companies providing services to related foreign entities—important for multinational transfer pricing and how profit is booked.

For B2B marketers and SaaS operators, the signal is straightforward: India is attempting to become a default region for cross-border compute, potentially reducing unit economics for inference-heavy products, data pipelines, and latency-tolerant workloads.

Hyperscaler build-out meets power and water constraints

Big cloud providers are already scaling India. Google has disclosed a 15 billion dollar plan to expand data centers and build an AI hub (TechCrunch). Microsoft has announced 17.5 billion dollars in investment by 2029 (TechCrunch). Amazon said it will add 35 billion dollars by 2030, taking total planned spending to about 75 billion dollars (TechCrunch).

But incentives do not remove physical bottlenecks. Reporting has highlighted patchy power availability, high electricity costs, and water scarcity as constraints for energy- and cooling-intensive AI compute (Reuters). Expect a split outcome: attractive tax treatment on paper, and a real-world race for reliable power contracts, land, and state-level approvals.

The practical takeaway: founders should watch India-region pricing, but also plan for capacity and reliability risks if they are considering shifting non-India workloads there to lower COGS.

Stay Informed

Weekly AI marketing insights

Join 5,000+ marketers. Unsubscribe anytime.

Related Topics

Indiacloud computingdata centersGoogleMicrosoftAmazontax policy