India's 31-year-old tax treaty with the US is facing significant challenges as the country's AI ambitions intensify amid a broader regional AI investment cycle. According to reports from Business Standard, the Convention for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion, signed in 1989, was designed for a pre-cloud, pre-AI era and now constrains rather than supports digital innovation. The treaty defines royalties as payments for physical equipment and industrial processes, while modern AI ecosystems operate on subscription-based cloud services and automated intelligence models. Recent analysis shows that India remains underrepresented in the AI hardware supply chain, with the country's technology ecosystem heavily concentrated in IT services, software development, and business-process outsourcing facing growing disruption risks as generative AI automates previously outsourced tasks.
India's Budget 2026-27 has announced significant tax incentives for AI infrastructure development. As reported by Business Standard, the budget introduces a 15% cost-based safe harbour for non-resident foreign companies providing cloud services globally using Indian data centres, valid until 2047. Additionally, all IT services including software development and knowledge process outsourcing are grouped under a single category with a 15.5% common safe harbour margin. These measures are designed to attract major tech companies like Amazon, Microsoft, Google, and Meta to establish AI and cloud infrastructure in India. The AI-led capex cycle has lifted demand for semiconductors, memory, servers, networking equipment and related electronics, helping offset tariff uncertainty and muted consumer demand globally.
The software industry is experiencing unprecedented disruption as AI coding threatens traditional software-as-a-service models. According to Mergermarket analysis, around $300 billion was wiped from software company valuations earlier this year, with investors becoming increasingly cautious about basing investment decisions on traditional assumptions about growth and market position. UBS analysts warned that "LLM providers, AI-native entrants, and DIY initiatives are dismantling some moats" and that SaaS providers face challenges to "remain relevant." However, HSBC analysts suggested that productivity boost concerns from AI tools had been overestimated, citing strong double-digit growth figures from companies like Adobe and Salesforce. The disruption particularly affects "purely workflow-oriented SaaS tools" where AI agents can readily replicate user interfaces and process coordination layers, while applications with embedded governance, regulatory complexity, and high customization levels remain more resistant to AI disruption.
The current treaty structure creates potential fiscal risks as India increases spending on US AI applications. According to Business Standard analysis, cross-border payments to US entities for AI services risk being characterized as royalties or business profits under the treaty, potentially resulting in lost withholding tax revenue. India risks losing revenue while US firms may argue that payments are not attributable to permanent establishments in India, creating unilateral interpretation risks and disputes. The benefits of the AI boom are far from evenly distributed, with economies like Taiwan delivering robust double-digit export growth supported by its dominant role in advanced semiconductor fabrication, while India and other economies with limited AI exposure have captured fewer gains from the AI investment cycle.
Experts recommend comprehensive treaty amendments to address AI-related taxation challenges. As reported by Business Standard, proposed reforms include amending Article 5 to define 'significant digital presence' with OECD-adapted frameworks, updating Article 12 to explicitly exclude AI model payments from royalties, and aligning Budget 2026-27 safe harbour provisions with treaty obligations. The recommendations also suggest establishing a bilateral AI tax working group to develop frameworks for cross-border AI revenue attribution and prevent disputes through transparent documentation and coordination. The AI investment cycle is highly capital-intensive, with benefits concentrated among companies, sectors, and regions closest to the technological frontier, potentially widening economic disparities across different segments of society.
