← All rulings

ANS-100/P — Telecom, Energy & Infrastructure Sector Calibration Note

Drafted 30 August 2026 · Ratification not separately logged — see the note on the rulings index · written before the first telecom, energy and infrastructure company was scored · Rater 1 · Companion to ANS-100/P v2.1 (frozen anchor set) · Sector: Reliance Industries, Bharti Airtel, Adani Enterprises

Ratification log: F11 ratified — Q32 classes N for retail; no score changes; retail files final as drafted pending right-of-reply.

1. Why these three need their own note — and why entity screening is the central issue

These are the frozen universe's final three companies and its three most structurally complex. All are conglomerates, and two carry the Index's hardest crossover problems:

The D9/D9b and F9 lessons apply at their sharpest here: verify each entity's structure at pack assembly, before scoring, and state the scored perimeter in the file header.

2. Proposed rulings (ratify individually: G1 ☐ … G10 ☐)

G1 — Q25 economic pair, by business type. Telecom: revenue per employee read with ARPU, subscriber churn, and network opex per site or per GB carried. Energy and infrastructure: revenue per employee read with plant or asset availability, capacity utilisation, and specific energy or throughput metrics. Conglomerates: use the segment where the AI evidence sits, and state which segment the economic pair is drawn from. A1 unchanged. Named confounds: commodity cycles, spectrum and licence costs, tariff regulation, capex phasing, and demerger effects.

G2 — Q27, the sector's strongest opportunity after steel. Telecom admits: network fault prediction and self-healing rates, AI-attributed churn-prediction accuracy, automated service-assurance or zero-touch provisioning shares, spam and fraud detection rates. Energy and infrastructure admit: predictive-maintenance-driven downtime reduction, AI-attributed yield or availability improvement, autonomous inspection rates. Excluded (the standing analogue of B2, C2, D3, E2 and F2): subscriber counts, data traffic volumes, network coverage percentages, digital-channel adoption, and app download figures — these measure scale, not automation.

G3 — "Network automation is not necessarily AI." SDN, NFV, OSS/BSS orchestration, SCADA and conventional network management predate machine learning and score nothing without evidence of learning, prediction or perception. Self-optimising networks with disclosed learning behaviour do score. This is D3 and F3 applied to network infrastructure.

G4 — Q6/Q7 autonomy anchor. Production autonomy means AI systems taking or triggering network or plant actions: self-healing networks, autonomous traffic steering, automated capacity allocation, closed-loop process control, autonomous inspection with reject authority. Recommendation engines and dashboards are level 2. B2-4 applies — predictive systems at disclosed scale in delivery-critical paths may reach 3 without actuation authority.

G5 — Q33 product-embedded AI. Level 3 requires customer-facing AI shipped with disclosed volumes: AI-powered fraud or spam blocking with subscriber reach, AI-driven digital services with usage figures, AI infrastructure sold as a service with contracted capacity. For infrastructure players, AI data-centre capacity contracted or commissioned counts, consistent with the L&T treatment in manufacturing.

G6 — Q31 exposure re-anchored. Telecom's AI exposure is pipe-ification: AI-native services capturing value above the network while operators carry the traffic, plus agentic customer interfaces bypassing operator channels. Energy and infrastructure exposure is demand-side disruption and competitive cost displacement, alongside the opportunity of AI-driven power demand. Mitigation = disclosed AI-native revenue lines, platform ownership, or AI infrastructure build-out with contracted capacity.

G7 — Q18. Telecom operators handle subscriber data under DoT and TRAI obligations; energy and infrastructure carry critical-infrastructure security duties. Regulatory baselines score nothing. Level 2 requires AI-specific governance with specifics — model governance, AI ethics policy, automated-decision disclosure, or an AI-specific security posture. Level 3 requires audited or certified AI management systems, or AI surfaces in a public bug-bounty scope.

G8 — Q24 and capital intensity. These are the most capital-intensive companies in the Index. The A4 ≥5%-of-TTM-revenue line applies only to identifiable AI investment. Network capex, refinery capex and airport capex are never AI investment. AI data-centre and compute capex counts where disclosed separately — the treatment established for TCS HyperVault, HCLTech's data-centre entry and L&T's Vyoma.AI.

G9 — Entity rulings (verify at pack assembly, per the D9 lesson). Reliance Industries: consolidated, including Jio Platforms and Reliance Retail; state explicitly that Jio's AI evidence is in scope as a consolidated subsidiary. Bharti Airtel: consolidated India and Africa operations, with Nxtra data centres and Airtel Payments Bank noted; Indus Towers is an associate and its evidence must not cross over. Adani Enterprises: the listed incubator entity only — evidence from separately listed Adani Group companies (Green Energy, Ports, Power, Energy Solutions, Ambuja) must be excluded, exactly as ITC Infotech was from ITC and Tech Mahindra from M&M.

G10 — Signal library additions. Integrated annual reports, quarterly investor presentations (telecom operators disclose operating metrics quarterly at unusual granularity), TRAI filings, network-technology partner announcements, data-centre capacity disclosures, and telecom-industry analyst assessments.

3. What does not change

All v2.1 anchors, P-caps, evidence floor, 18-month window, A1–A10, Batch-2 as applied, Appendix A ceiling arithmetic, the pack-based confidence rule, F-U full-text annual-report retrieval, and the evidence-language rule. Application, not amendment.

4. On ratification

Score in universe order: Reliance Industries → Bharti Airtel → Adani Enterprises. Expect Reliance to be the most AI-evidenced of the three and the hardest to scope; expect Airtel to be the cleanest test of G2's telecom categories; expect Adani Enterprises to be the sharpest test of G9's crossover discipline. A Q32 ruling (G11) is anticipated at the sector synthesis, after which the frozen universe is complete and the full cross-sector synthesis follows.

Published 9 September 2026 under CC BY-SA 4.0 as part of the Edition One replication material. Status changed from "proposed" to "ratified" on publication: the note was drafted on 30 August 2026 and written before the sector was scored; the ratification record for this set is described on the rulings index. No ruling text is changed by this publication.