'Adopt, adapt, advance.' Evaluate how far the IndiaAI Mission is aligned with this framework, and what foundational gaps remain.

The World Bank's World Development Report 2026 asks developing countries to adopt existing AI tools, adapt them to local conditions (which gives the biggest benefits) and advance to frontier models selectively, since that path costs the most [1]. India's IndiaAI Mission, approved in 2024 with ₹10,371.92 crore over five years [2], fits this framework closely on paper. Delivery and foundations, however, are lagging.

Alignment with the framework

• Adopt: The Compute Capacity pillar builds a pool of 10,000+ GPUs through public-private partnership. This lets startups and ministries use existing models without owning hardware [2].
• Adapt (the strongest fit): The Application Development and Datasets Platform (AIKosh) pillars target local problems [2]. Results so far:
• Baramati sugarcane yields rose 40% with 30% less water and fertiliser.
• TB detection rose 35%.

• An AI monsoon forecast built partly on open-access models is now being followed by Ethiopia [3].

• Advance: The Innovation Centre funds indigenous foundation models [2]. This is a measured bet on sovereign capability, not a race to the frontier.

• Governance: The Safe & Trusted AI pillar covers bias mitigation and deepfake detection [3]. This fits the WDR's call for countries to act as pragmatic regulators [1].

Foundational gaps

• Execution: PRS has flagged underutilisation of IndiaAI Mission funds [4]. Approved money is not turning into compute, datasets and startups quickly enough.
• Infrastructure: The WDR treats reliable electricity, internet and education as preconditions [1]. Without them, adaptation stays stuck at the pilot stage.
• Skills: The ILO finds generative AI will mainly augment jobs rather than replace them [5]. Yet FutureSkills mostly expands degree-level AI courses [2] and does little for clerical, health and farm-extension workers.
• Dependence: A few firms in a few countries control the advanced models, chips and data centres. The WDR says this creates dependency risks [1].
• Equity: AI could widen inequality within countries [6] if benefits stay with users who are urban and already online.

Verdict: The mission is closely aligned in design but only partly in delivery.

Overall, IndiaAI is strong on adopt and adapt and cautious on advance, but uneven foundations and slow spending limit what it achieves. Four steps would help turn pilots into scale: faster release of funds, preferring open models in public buying, building rural power and broadband alongside AI, and skilling that centres on workers. These would make India a credible model for the Global South and advance SDG 9's goal of inclusive innovation.