ENERGY AI OBSERVATORY / SUNYA — READING PACK ASSEMBLED 25 OCTOBER 2026

The web register receives selected updates. A newer publication date does not mean every profile was re-audited. PDFs and decks are fixed selections, with their own edition dates.

Initial source screen: 26 September 2024–26 September 2026. Broad classification review: 8 October 2026. Targeted Presidio, Aramco and EOG updates, plus three field additions: 10 October 2026. Flywheel platform and recruitment review: 12 October 2026. CNX safety, Aker BP implementation and anonymous failure review: 14 October 2026.
306 profiles; 220 company-use profiles. Profile counts can overlap and do not measure industry adoption.

Start with download-guide.html for the file list. For the current interactive report, charts and sources: https://energyai.sunya.ai/

The twelve-project comparison also includes 24 source-linked HTML evaluation briefs, one operator and one AI-provider perspective per case. These are Sunya proposals, not company plans.

CURRENT EXPORTS
operating-results.html | 2026-10-26 | Seventeen selected comparisons and assessments, including stopped use cases, historical limited-value trials and technical tests. Each retains population, period, source and limitations.
operating-results.csv | 2026-10-26 | Same seventeen bounded comparisons and assessments; different measures are not a ranking or complete ROI.
slb-chemical-dosing.svg | 2026-10-25 | Nearly 1,400 Permian wells; separate 20-month performance and first-nine-month supplier-reported net savings.
company-brain.html | 2026-10-25 | Twelve selected comparisons; broad review 9 October, with KOC SANAD added 18 October and PTTEP X.brain and Devon ChatDVN checked 21 October.
operating-technology.json | 2026-10-26 | Automation, internal software and technology foundations, with source and AI-attribution limits.
operating-technology.csv | 2026-10-26 | Same technology observations in tabular form.
operating-technology.md | 2026-10-26 | Readable source-linked technology observations.
research-watch.html | 2026-10-26 | Ecopetrol IT operations and AI rollout limits, YPF supplier-reported refinery support, and Dixstone software selection; earlier reviews retained.
visual-review.json | 2026-10-26 | Dated exhibits plus reported TotalEnergies use-case non-retention, with corroboration limits.
operator-failures.md | 2026-10-26 | Historical searches and dated follow-ups, including source limits for reported non-retention.
oct14c-research-review.html | 2026-10-25 | Aramco core labs, clarified Exxon/Ecopetrol/PETRONAS/Chord profiles, TotalEnergies digital planning and two technical studies.
oct14b-research-review.html | 2026-10-25 | Xinjiang diagnostics, CNX methane, HG Oil, Santos and two KPMG surveys; source dates and benefits kept separate.
industry-reports.json | 2026-10-26 | Current report entries, populations, methods and limitations.
industry-reports.csv | 2026-10-26 | Same current industry publications in tabular form.
oct14-research-review.html | 2026-10-25 | CNX current-use disclosure, Aker BP implementation announcement, anonymous failure note and existing Expand evidence.
current-initiative-register.csv | 2026-10-25 | All 306 profiles, with targeted source checks through October 25; supplementary technology context is reported separately.
research-register.xlsx | 2026-10-25 | Same rows and values as the CSV, with filters.
research-data.json | 2026-10-26 | Records, source entries, disclosures, provenance and dossiers.
initiative-histories.csv | 2026-10-25 | Current histories, including the field additions and targeted Presidio, Aramco and EOG updates. Account briefs retain their own dates.
flywheel-energy.html | 2026-10-25 | Existing AI app platform, hiring intentions, conventional software context and source-access limits; reviewed 12 October.
industry-brief.html | 2026-10-25 | Three industry findings, six workflows, four measurement examples and source-linked follow-up questions; editorial edition of 25 October.
industry-brief.pdf | 2026-10-25 | Portable six-page edition of the October 25 synthesis; source links and measurement limits retained.
industry-brief.json | 2026-10-25 | Canonical evidence counts joined to selected workflow explanations and proposed questions.
workflow-comparison.html | 2026-10-25 | Action authority, asset scope, measurement limits and proposed evaluations; includes original source locations.
workflow-comparison.csv | 2026-10-25 | The same case fields and proposed evaluation questions as the interactive comparison.
workflow-comparison.json | 2026-10-25 | Structured editorial comparison and source metadata; no new adoption classifications.
slb-data-ai.html | 2026-10-25 | Original-source assessment, reported Digital economics, data rights and questions for the investment case.
slb-data-ai.json | 2026-10-25 | Source metadata and reported segment figures; AI-only economics remain unknown.
slb-data-ai.md | 2026-10-25 | Portable text edition of the targeted October 24 assessment.
oct10-field-cases.html | 2026-10-25 | BPX, Burk and Expand: operating details, measurement limits and original sources; EOG context kept separate.
oct10-research-review.html | 2026-10-25 | Findings, research leads and limits; not additional deployment counts.

DATED SNAPSHOTS — THESE DO NOT AUTOMATICALLY INCLUDE LATER UPDATES
research-sample.pdf | 2026-10-08 | Eight selected upstream companies; not the complete register.
energy-ai-exhibits.pdf | 2026-10-08 | Selected exhibits. Does not include the new chart gallery or October 10 source updates.
energy-ai-exhibits.pptx | 2026-10-08 | Editable version of the same dated exhibit selection.
industry-takeaways.pdf | 2026-10-08 | Industry evidence counts and selected exhibits as of this edition.
research-atlas.pdf | 2026-10-07 | Four exhibits from the October 7 assessment.
latest-research-additions.pdf | 2026-10-09 | Selected supplementary notes; not all later findings or an updated full report.

Older reports retain their historical text and date labels. Use the initial-screen dates above for precise cohort framing.
Sources establish what was disclosed. They do not independently establish performance, net returns or a complete worldwide project census.
The selected comparison covers eight companies. Other files include historical supporting notes with their own review dates. Original publishers may restrict source access.
