Editorials, Opinions & Explained2 Items

Core TopicImportantConcise

OpinionsSigned Op-Eds

01India's DPI Playbook for AI — Making Intelligence Free02SHANTI Framework — Bay of Bengal as India's Maritime Laboratory

OpinionsSigned Op-Eds · The Hindu / Indian Express

01

India's Fourth Utility: Applying the DPI Playbook to Make Artificial Intelligence Free

Core TopicOpinionGS-III · Economy — S&T, Digital Economy, Governance of TechnologyPrelims + MainsThe Hindu · Opinions · 31 Jul 2026

India turned identity, payments, and data into free public utilities through deliberate infrastructure design — and the argument being made now is that artificial intelligence should be the fourth, applying the same DPI playbook to make inference as cheap as a gigabyte of data.

◈ Background: India's DPI Stack — What It Is and How It Was Built

Digital Public Infrastructure (DPI) refers to shared, open, interoperable digital systems built by or for the state, made available to the entire economy as a utility rather than a proprietary service. India's DPI stack is distinguished globally by the simultaneous existence of three such layers.

• Aadhaar (identity layer): Launched in 2009 under the Unique Identification Authority of India (UIDAI), established by the Aadhaar (Targeted Delivery of Financial and Other Subsidies, Benefits and Services) Act, 2016. Enrolled approximately 1.4 billion individuals. Converts identity verification — previously a paper-intensive, time-consuming process — into a low-cost API call accessible to any enrolled entity. The Supreme Court in K.S. Puttaswamy v. Union of India (2018) upheld Aadhaar but restricted its mandatory linkage to private entities.
• UPI (payments layer): Developed by the National Payments Corporation of India (NPCI), a not-for-profit entity set up under the Payment and Settlement Systems Act, 2007, with ownership shared among banks. UPI went live in 2016. It made digital payments effectively free by abstracting inter-bank transfers into a common protocol. As of 2025–26, UPI processes approximately 20 billion transactions monthly — making India the world's largest real-time payments market by volume, ahead of China, the US, and the EU combined.
• DEPA / Account Aggregator (data layer): The Data Empowerment and Protection Architecture (DEPA), conceptualised by the Ministry of Electronics and Information Technology (MeitY) around 2019–20, enables consent-based data portability. The Account Aggregator (AA) framework — operationalised through the Reserve Bank of India's AA Master Direction, 2016 — allows financial data to flow between regulated entities with explicit user consent. This is India's functional equivalent of Europe's Open Banking / PSD2 directive.
• How data became free: The dramatic fall in India's data cost was not a government subsidy programme. Between 2016 and 2019, the cost of 1 GB fell from approximately $4 to under $0.30 — among the cheapest globally. This resulted from the entry of one large private player that absorbed the fixed costs of a nationwide 4G network and priced at near-marginal cost, forcing incumbents to match. The state's role was to release spectrum and set competition rules; the market crashed the price. Around 500 million Indians came online in roughly five years as a consequence.

India's Position in the Global AI Economy: The Extractive Trade

• India currently occupies an asymmetric position in the global AI value chain — supplying inputs at the bottom and purchasing outputs at the top. Indian engineers staff and fine-tune frontier models built by large technology companies headquartered abroad. Indian-generated data (digitised public records, transactions, linguistic corpora) feeds training datasets for these models.
• Indian workers perform the labour-intensive annotation and reinforcement learning from human feedback (RLHF) processes that make models safer and more accurate. Yet Indian startups must access the resulting intelligence by purchasing API tokens priced in US dollars, hosted on foreign infrastructure, and subject to export control regimes (notably the US Export Administration Regulations).
• The article's framing — "ship out the cotton, buy back the cloth" — invokes the colonial-era drain of wealth critique. While the analogy is polemical rather than precisely calibrated, it accurately captures an asymmetry: India generates significant value in the AI supply chain but captures relatively little of the rent from the finished product.
• Scale of the imbalance: Global AI API revenues are overwhelmingly concentrated among a small number of US-headquartered firms. India, despite ranking among the top-5 countries in AI talent concentration (per Stanford AI Index 2024), has no frontier foundation model of comparable scale or adoption.

The Three-Pillar Proposal: IndiaAI Token Economy

• Pillar 1 — Compute: The IndiaAI Mission, approved by the Union Cabinet in March 2024 with an outlay of ₹10,372 crore over five years, anchors India's public compute strategy. It operates on a public-private partnership model — empanelling private cloud providers and aggregating government demand — rather than building state-owned data centres. The target is 1,00,000 GPUs accessible to eligible startups and researchers at approximately ₹65 per GPU hour, a fraction of international market rates (typically $2–4/GPU hour for comparable hardware). The article adds an energy dimension: cheap electrons are the new cheap spectrum. India's National Electricity Plan does not yet treat AI inference load as a distinct planning category — an oversight the author argues should be corrected, with dedicated renewable or nuclear generation earmarked for compute clusters.
• Pillar 2 — Open Models: Currently, the most capable foundation models are proprietary. The proposal is that any AI model developed using state-subsidised compute or public datasets must be released under an open-weights licence. This mirrors the UPI philosophy: the government builds the rails and makes the protocol free; private applications compete on user experience rather than on model ownership. The state's asset here is its linguistic data — legal rulings, agricultural records, educational curricula in all 22 Scheduled languages — which should be anonymised, aggregated, and made available exclusively for open-source model training.
• Pillar 3 — Unified Intelligence Interface (UII): A national API gateway modelled on UPI's interoperability principle. Just as UPI abstracts inter-bank complexity behind a common protocol, the UII would let any application — a government agency, a school, a startup — call any model (sovereign or private, open or proprietary) through standardised interfaces covering identity, consent, billing, and safety. A freemium tier — verified through Aadhaar — would give Indian students and startups a monthly token allotment at no cost, subsidised by the state.

▤ Key Data Points for UPSC

• UPI transactions: ~20 billion/month (2025–26) — world's largest real-time payments system by volume
• Aadhaar enrolment: ~1.4 billion individuals
• Data cost fall: $4/GB (2016) → under $0.30/GB (2019) — ~93% decline in three years
• Indians coming online: ~500 million in roughly five years post-2016
• IndiaAI Mission outlay: ₹10,372 crore over five years (approved March 2024)
• GPU target: 1,00,000 GPUs; current onboarding: 38,000+
• Compute pricing: ~₹65/GPU hour (IndiaAI) vs $2–4/GPU hour (international market)
• NPCI: Not-for-profit, owned by consortium of banks; operates under Payment and Settlement Systems Act, 2007
• UIDAI: Statutory authority under Aadhaar Act, 2016; under MeitY

Critical Evaluation

• Strengths of the argument: The DPI analogy is well-grounded — India's success with Aadhaar and UPI rested on specific structural features (interoperability, open protocols, state-mandated access) that are replicable in AI distribution. The compute subsidy focus avoids the trap of trying to compete with frontier model training, where capital requirements run into tens of billions of dollars.
• Data sovereignty challenge: Aggregating and anonymising public data across 22 languages, multiple ministries, and state governments involves significant governance complexity. India does not yet have a comprehensive data protection framework for non-personal public data — the Digital Personal Data Protection Act, 2023 covers personal data but leaves the public data aggregation question largely unaddressed.
• Open-weights risk: Open-weight model releases carry dual-use risks — the same model that enables a rural doctor to diagnose better can be fine-tuned for disinformation or cyberattack. Governance frameworks for open-weight AI are still nascent globally (the EU AI Act treats open-source models with general-purpose AI provisions; the US currently relies on voluntary commitments).
• Infrastructure gap: India's AI ambitions are constrained by power and cooling infrastructure. Data centres require uninterrupted power at scale; India's grid reliability outside major metros remains variable. The National Electricity Plan (NEP) 2023 targets 900 GW of installed capacity by 2032 but does not disaggregate AI/data centre demand — a planning gap the article rightly identifies.
• Comparison with other models: The EU's approach (regulation-first, with the AI Act) prioritises rights protection but risks innovation lag. The US approach (private sector-led, state backstops through NIST frameworks) has produced frontier models but concentrates market power. India's proposed model is a third path — public infrastructure + open models + private applications — with precedent in its own DPI experience.

Figure 1 — India's DPI Stack: Three Layers and the Proposed Fourth

LAYER 4 (PROPOSED)Intelligence / AI Inference — Unified Intelligence Interface (UII)NEWLAYER 3Data — DEPA / Account Aggregator (consent-based portability)2019–LAYER 2Payments — UPI (NPCI; ~20 bn txns/month; near-zero cost)2016–LAYER 1Identity — Aadhaar / UIDAI (1.4 bn enrolled; API-based KYC)2009–No parallelEU PSD2Brazil PixEstonia e-IDGlobalparallel

India is uniquely positioned having all three existing DPI layers integrated and interoperable — no other country has this combination. The proposal is to add a fourth: AI inference as public utility.

✎ Mains Practice Question

India's success with Digital Public Infrastructure — Aadhaar, UPI, and Account Aggregator — rests on specific design principles. Critically examine whether these principles can be extended to make artificial intelligence inference a public utility, and what institutional and regulatory prerequisites such an extension would require. 15 marks · 250 words

02

SHANTI and the Bay of Bengal: India's Evolving Maritime Security Doctrine

Core TopicOpinionGS-II · IR — India's Maritime Policy, Regional Organisations, Indo-PacificPrelims + MainsThe Hindu · Opinions · 31 Jul 2026

India has articulated a new maritime security framework — SHANTI — and the argument is that the Bay of Bengal, with its shared geography, shared vulnerabilities, and existing institutional architecture under BIMSTEC, is the natural laboratory in which to translate that framework from principle to practice before extending it across the wider Indo-Pacific.

◈ Background: India's Maritime Doctrine — From SAGAR to MAHASAGAR to SHANTI

India's maritime security articulation has evolved through three successive conceptual frameworks over a decade, each building on — and expanding the scope of — its predecessor.

• SAGAR (Security and Growth for All in the Region), 2015: Articulated in March 2015 during a visit to Mauritius. Established India's aspiration to be a net security provider in the Indian Ocean and to pursue 'equity in development' for Indian Ocean states. SAGAR was primarily a vision statement — it identified what India wanted to achieve but did not specify mechanisms or a governance architecture.
• MAHASAGAR (Mutual and Holistic Advancement for Security and Growth Across Regions), 2025: Announced in 2025, MAHASAGAR expanded the geographic and conceptual scope of SAGAR — from the Indian Ocean sub-region to the wider Indo-Pacific, and from bilateral development partnerships to recognising the interconnectedness of security threats horizontally (across nations) and vertically (across threat types). It acknowledged non-traditional security threats as co-equal with traditional maritime rivalry.
• SHANTI (Securing Holistic Advancement through Norms, Trust and Integrity), 2026: Introduced by the External Affairs Minister on 13 July 2026, while announcing India's candidature for the UN Security Council for the 2028–29 term. SHANTI is described in the article as the "grammar" — the method — of maritime security, distinct from the "vision" (SAGAR) and the "scope" (MAHASAGAR). It shifts the emphasis from deterrence and competition to norm-building, trust-based cooperation, and governance of shared maritime commons.

The Bay of Bengal: Geography, Geopolitics, and Shared Vulnerabilities

• Historical centrality: The Bay of Bengal connected South and Southeast Asia through trade and cultural exchange for centuries before the Indo-Pacific entered diplomatic vocabulary. The ancient maritime Silk Route passed through the Bay; the Chola Empire (9th–13th centuries CE) projected naval power across the Bay to the Malay Peninsula and Sumatra.
• Contemporary strategic importance: The Bay links India's Act East Policy with ASEAN, provides access to the Malacca Strait (through which an estimated 40% of global trade passes, including 80% of China's oil imports), and connects the eastern Indian Ocean to the South China Sea.
• The Malacca Dilemma: China's heavy dependence on the Malacca Strait for energy and trade has driven its strategic calculus in the Bay — including port development (Chittagong in Bangladesh, Kyaukphyu in Myanmar), infrastructure projects under the Belt and Road Initiative, and submarine deployments. This intensifies geopolitical competition in a region that India considers its natural sphere of influence.
• Shared non-traditional threats: The Bay of Bengal is one of the world's most cyclone-prone regions — 6 of the 10 deadliest tropical cyclones in recorded history struck the Bay's coastline. Its littoral states (India, Bangladesh, Myanmar, Thailand, Sri Lanka) face near-identical challenges: disaster risk reduction, coastal erosion, fisheries depletion, undersea cable vulnerability, and the effects of sea-level rise. These threats do not respect national boundaries.
• Institutional gap: Unlike the South China Sea, the Bay has no functional multilateral maritime governance mechanism. Bilateral arrangements dominate, creating coordination gaps when disasters or security incidents cross borders.

BIMSTEC as the Institutional Vehicle

• BIMSTEC (Bay of Bengal Initiative for Multi-Sectoral Technical and Economic Cooperation) was established in 1997, initially as BIST-EC (Bangladesh, India, Sri Lanka, Thailand Economic Cooperation). It expanded to include Myanmar (1997) and Nepal and Bhutan (2004), reaching its current seven-member composition. Secretariat is based in Dhaka.
• BIMSTEC is the only regional organisation that exclusively connects South Asia with Southeast Asia through the Bay of Bengal — making it the natural institutional vehicle for SHANTI's operationalisation in the Bay.
• Recent momentum: The July 2026 BIMSTEC National Security Advisers' meeting in New Delhi produced two significant outcomes: member-states adopted common principles for maritime law enforcement and humanitarian assistance and disaster relief (HADR), and agreed to hold BIMSTEC's first-ever joint maritime security exercise in November 2026. A white shipping information-sharing agreement remains under discussion.
• White shipping: "White shipping" refers to the sharing of Automatic Identification System (AIS) data on commercial vessels among partner navies — a transparency measure that improves maritime domain awareness without involving sensitive military intelligence. India has bilateral white shipping agreements with the US, France, Japan, Australia, and others; a BIMSTEC-wide arrangement would be a significant norm-building step.

SHANTI's Conceptual Contribution: From Deterrence to Resilience

• The article's central argument is that SHANTI represents a qualitative shift in India's maritime security posture — from deterrence (projecting power to deter adversaries) to resilience (building shared capacity to manage shared vulnerabilities). This maps onto a global trend in security studies: the recognition that non-traditional security threats require cooperative governance frameworks rather than competitive balance-of-power responses.
• The article positions SHANTI as shifting emphasis from influence to "non-prescriptive institution-building" — meaning India proposes norms and mechanisms but does not dictate outcomes, differentiating Indian regional leadership from Chinese infrastructure-led influence or US security-alliance models.
• India's "preferred security partner" and "first responder" positioning in the Indo-Pacific is reinforced by a track record of HADR operations: Operation Maitri (Nepal earthquake, 2015), Operation Vanilla (Cyclone Kenneth, Mozambique, 2019), Operation Samudra Setu (COVID-19 evacuations, 2020), and Cyclone Mocha relief operations (2023).

Critical Evaluation

• Institutional fragmentation remains: The Bay of Bengal has BIMSTEC, the Indian Ocean Rim Association (IORA), ASEAN, and various bilateral mechanisms — but no single maritime governance body. SHANTI provides a normative framework but not a new institution; its success depends on existing institutions (primarily BIMSTEC) developing operational capacity.
• China factor: China is not a member of BIMSTEC but has deep infrastructure and economic ties with Bangladesh and Myanmar. Any Bay of Bengal maritime governance architecture that excludes China will face a credibility gap in states that depend on Chinese investment — yet including China would fundamentally alter SHANTI's norm-building character.
• Myanmar's internal crisis: Since the February 2021 military coup, Myanmar's civilian government has been displaced. Its participation in BIMSTEC security cooperation is diplomatically complex for democratic member-states. The July 2026 NSA meeting outcome did not publicly address this dimension.
• Nomenclature proliferation: SAGAR → MAHASAGAR → SHANTI represents India's tendency to articulate frameworks through acronyms without always following through with institutional architecture. The July 2026 BIMSTEC outcomes are concrete steps, but sustained follow-through — particularly on the white shipping agreement and the November exercise — will determine whether SHANTI has operational content or remains rhetorical.

Figure 2 — India's Maritime Security Doctrine: Evolution from SAGAR to SHANTI

SAGAR2015VisionSecurity & GrowthIndian Ocean focusNet security providerMAHASAGAR2025ScopeIndo-Pacific reachNon-traditional threatsGlobal South linkageSHANTI2026MethodNorms, Trust, IntegrityBay of Bengal labBIMSTEC vehicleExpanded scopeAdded method

SAGAR (2015) set the vision; MAHASAGAR (2025) expanded the geographic and threat scope; SHANTI (2026) provides the method — norms, trust, and integrity — to translate both into cooperative practice.

✎ Mains Practice Question

India has articulated a succession of maritime security frameworks — SAGAR, MAHASAGAR, and SHANTI — over the past decade. Critically analyse how SHANTI represents a doctrinal evolution in India's approach to maritime security, and examine whether the Bay of Bengal, through BIMSTEC, provides an adequate institutional basis for operationalising this framework. 15 marks · 250 words