DPI Deep Dive — Friday | September 04, 2026

Layer: L5 — Sectoral Infrastructure (ABHA/ABDM, AgriStack, eCourts) Window covered: August 28 – September 4, 2026

This week, all three sectoral stacks hit the same wall from different directions: the registries are built, but usage is the hard part. Health — Haryana’s numbers show how thin private-sector adoption of ABDM actually is. Agriculture — the Centre paid Telangana for finishing AgriStack milestones while the state photographs 1.93 crore farm fields and quietly links season-end procurement to registration. Justice — the Supreme Court bumped a “guilty-by-default” design flaw in the V-Courts portal up to its e-Committee. And a vendor breach abroad showed what is at stake when court data lives on commercial servers. Four stories, one theme: the sectoral layer has moved from construction to utilisation, and utilisation is where consumer interests actually get decided.

1. ABDM’s adoption bill comes due — Haryana shows the private sector isn’t showing up

Haryana is one of India’s richer, better-administered states, which is exactly why its ABDM scoreboard published this week is worth reading closely. As of September 1, the state had created 1.9 crore ABHA IDs — 64.76% of its reported 2.9 crore population — and onboarded 97.9% of targeted government facilities (4,016 of 4,099). But barely 12% of private and PPP healthcare facilities have registered, and just 1% of private nurses (The Tribune). In Gurgaon — a city whose healthcare market is overwhelmingly private — the state’s additional chief secretary has warned that low private-sector adoption could slow the entire mission (Times of India). State mission director Dr Virender Yadav’s framing is honest: the system is designed to let a patient share records, prescriptions and diagnostics across providers with consent — but only if participating providers actually generate and link those records.

The national baseline makes the gap starker. At the third Mission Steering Group review (chaired by Health Minister J.P. Nadda on July 10), ABDM stood at 93.95 crore ABHA numbers, over 105 crore linked health records, 5.33 lakh registered facilities and 9.85 lakh professionals, with 2.72 lakh facilities running ABDM-enabled software and nearly 24 crore Scan & Share OPD tokens generated (PIB). Nadda’s own directive — that the next phase must focus on “maximum adoption and utilisation” — is precisely what Haryana’s private-sector numbers show isn’t happening yet. The spread across states is wide: Nagaland has ~8.96 lakh ABHAs, about 38% of its population (Eastern Mirror), while Rajasthan reported 7.1 crore ABHAs with 89.7% saturation by May, plus integrations of state platforms like PCTS into the national mesh (Digital Health News).

Analysis: Government facilities are captive audiences — a circular gets a district hospital registered. Private hospitals answer to patients and insurers, not the civil surgeon, and ABDM’s incentive scheme (payments per digitally enabled transaction) has clearly not offset the workflow cost of integrating registration, records and consent flows into private OPD software. The consumer stakes are concrete: an ABHA card without linked private records means your diabetes history at a corporate hospital lives nowhere near your government OPD record — the “longitudinal health record” pitch fails exactly where most tertiary care happens. Watch the right metric: not ABHA counts (saturation is largely a push-registration artefact) but the share of records generated by private providers. Until that moves, ABDM is a public-sector archive with a consent architecture waiting for participants.

2. AgriStack pays Telangana — and starts photographing its fields

Two Telangana stories this week capture AgriStack’s twin engines: fiscal reward and data extraction. On August 31, state agriculture minister Tummala Nageswara Rao announced the Centre had sanctioned ₹175 crore for 2026-27 under the Special Assistance to States for Capital Investment (SASCI) scheme, rewarding Telangana for completing its first AgriStack milestone — an AgriStack-based fertilizer distribution system. That is on top of ₹424 crore already sanctioned for completing the farmer registry (The Hindu). The minister’s pitch: linking farmers, crop extent, crops grown and fertilizer supplied digitally will make welfare delivery transparent.

The same week, the state’s digital crop survey went into overdrive. For the Kharif season, Telangana will photograph 1.93 crore farm fields across all 10,622 revenue villages — a first for the state — using around 10,000 volunteers working alongside agriculture extension officers, on a phone app designed to work offline and sync when signal returns (Deccan Chronicle). Buried in the same report is the enforcement hook: the department has indicated that end-of-season procurement will be linked to farmers registering their crop data during the exercise. The rollout is going national-by-default: Goa launched its own AgriStack Digital Crop Survey this week (Times of India), and Ladakh ran district-level staff training on September 2 (Voice of Ladakh).

Analysis: Two design choices deserve scrutiny. First, milestone-based fiscal incentives make AgriStack the first major DPI where states are literally paid per completed component — a faster federal adoption engine than ABDM’s per-transaction incentives, but it rewards completion, not data quality. A photographed field with the wrong crop, or a field whose cultivator isn’t its recorded owner, gets locked into the benefits pipeline anyway. Second, conditionality: linking procurement to registration turns a survey into a gate. For clear-title landholders, fine. But tenant cultivation is extensive in Telangana, and cultivators often aren’t the recorded landholders — for them, the gate can exclude exactly the farmers welfare schemes claim to target. Layer on the data question — Aadhaar-keyed farmer IDs tied to geo-tagged photographs of land — and you have sensitive personal data flowing through registries whose consent architecture predates operational DPDP rules. Thursday’s commerce-layer story (ONDC, Beckn-based agri networks) is worth re-reading in this light: AgriStack registries are quietly becoming the substrate for agri-commerce rails, which raises the cost of getting consent and data quality wrong.

3. “Guilty by default”: Supreme Court sends V-Courts’ payment-plea confusion to its e-Committee

On August 31, the Supreme Court directed its e-Committee to consider a plea seeking changes to the V-Courts portal so that citizens can compound traffic offences without pleading guilty, calling it an issue of “paramount importance” (LiveLaw; Bar & Bench). The grievance behind the plea: the portal treats payment of a traffic fine as a plea of guilt for compoundable offences — meaning citizens who simply wanted to settle a challan were unwittingly recording convictions against themselves (Lawbeat; ETV Bharat).

The scale is what makes this a DPI story rather than a curiosity. Virtual courts exist precisely to mass-adjudicate petty traffic challans — over 5.26 crore cases have already been processed through them (PIB). Meanwhile the National Judicial Data Grid puts system-wide pendency at roughly 5.65 crore cases — 95,316 before the Supreme Court, 64.7 lakh before High Courts, over 5.06 crore before district courts — and a sharp August 29 analysis argues NJDG should additionally track how many cases sit frozen under interim orders, a category it currently doesn’t measure (LiveLaw).

Analysis: In a physical courtroom, the difference between compounding an offence and pleading guilty is a lawyer’s alphabet. At portal scale, a default user-interface choice becomes mass adjudication: a button that records guilt manufactures conviction histories for people who just wanted to pay a fine and move on — with real consequences for jobs, visa applications and future cases that ask “have you ever been convicted?” This is the deepest consumer issue in DPI: defaults are policy. The fix here is cheap — an explicit compounding option with plain-language disclosure of consequences — which is why routing it to the e-Committee matters: its templates govern every district court portal at once, so one design decision can repair the experience for crores of users. The second, quieter lesson is about measurement. NJDG counts pendency beautifully, but nobody publishes how many virtual-court closures were compounds versus guilty pleas. An invisible statistic about a visible harm is exactly the kind of gap DPI governance keeps leaving open.

4. The C-Track breach is a preview, not a foreign story

On September 2, Thomson Reuters disclosed that an unauthorised party accessed files belonging to C-Track, its cloud-hosted court case management platform. The intrusion happened in March; it was detected on June 30; and the fallout now spans appellate courts in 11 US states plus Ontario — with notices warning that social security numbers, medical information and some sealed or confidential court documents may have been exposed (Infosecurity Magazine; TNW). Ontario’s chief justices have publicly confirmed the incident and flagged possible exposure of personal information of people involved in court proceedings.

Analysis: India’s eCourts Phase III leans heavily on NIC and a growing roster of private vendors for case management, video conferencing and e-filing. NJDG already publishes crores of case records; virtual courts hold identity and challan data keyed to driving licences; and Indian courts do run sealed workflows — in-camera proceedings, POCSO matters, commercially sensitive disputes. The C-Track episode is the cautionary architecture tale for all of it: vendor custody of court records concentrates risk outside judicial control, and three months passed between intrusion and detection. As India digitises justice, the trust layer (Sunday’s beat) has to mean more than consent notices — it means procurement clauses with breach-notification duties for justice data, jurisdiction over where court records may be stored, and keeping sealed records off commercial cloud stacks entirely.

Cross-layer connections this week

• L1 (Identity): ABHAs and farmer IDs are Aadhaar-keyed; identity saturation is high while service saturation lags. The hard part was never issuing the number — it’s the linkage.
• L6 (Governance): The e-Committee referral is grievance redress by architecture — one design fix, applied to every court portal simultaneously.
• L7 (Trust & Privacy): DPDP rules are still not fully operational for consent-heavy stacks (ABDM records, AgriStack field data), and C-Track shows what vendor custody risk looks like when it matures.
• L4 (Commerce): AgriStack registries are becoming substrate for agri commerce — fertilizer distribution, procurement linkage, and eventually open agri networks. Sectoral data is upstream of the commerce layer.

The week in numbers

Sources: PIB (3rd Mission Steering Group, July 10, 2026; Virtual Courts); The Tribune and Times of India (Haryana ABDM adoption, Sept 1–3); Eastern Mirror (Nagaland ABHA); Digital Health News (state digital health mapping); The Hindu (Telangana SASCI incentive, Aug 31); Deccan Chronicle (Telangana digital crop survey); Times of India (Goa crop survey); Voice of Ladakh (AgriStack training, Sep 2); LiveLaw, Bar & Bench, Lawbeat, ETV Bharat, Free Press Journal (SC e-Committee referral on V-Courts, Aug 31 – Sep 2); LiveLaw (NJDG pendency analysis, Aug 29); Infosecurity Magazine, The Next Web (Thomson Reuters C-Track breach, Sep 2).