LuaZeroBack to supporting notes
Supporting note · not verified

Sources and claim boundaries

Checked 7 September 2026. Primary sources are preferred. The supplied draft is reference material; its assertions are not evidence of work performed.

ClaimAssessment and treatmentSource
Deadline 7 September, 23:59 GMT+7Confirmed in the live 2026 application introduction. Jakarta uses the same offset.Application
Minimum viable product and Vietnam pilot intentOfficial site states this eligibility requirement. A software proof of concept may not meet the organiser's threshold; the team must not imply field validation.NZC
One-million-hectare programme / up to USD 40m TCAF commitmentSupported as programme context. Does not establish a payment due to LuaZero, an active buyer agreement or a causal claim that payments are blocked only by software.World Bank
USD 33.3m approvedReported in September 2024 with conditions and a stated commitment window. Do not present the old article as proof of today's undisbursed balance.Government report, 24 Sep 2024
354,839 hectaresReported by VOV citing the ministry. A programme-area figure, not LuaZero adoption. Optional historical context; not used as the main submission anchor.VOV
Only 20,000 ha have MRV / under 6%Not independently substantiated sufficiently in this review. Dates, definitions and populations may differ. Removed from public copy and application.Original draft only
MRV gaps matterSupported by IRRI's regional analysis, which includes institutional capacity and data systems; not solely software.IRRI, 16 Apr 2026
Gold Standard country/scaling valuesSupported in the methodology. Prototype scope is narrower than certification requirements.Methodology
SOC risk in 2026Newer rule must be considered; conditional 5% adjustment when practices are absent/insufficient.SOC update
Regrow cannot handle riceFalse. Regrow documents rice methane and AWD work. Removed.Regrow, 11 Mar 2026
Radar independently proves every drainage eventUnsupported. Crop canopy, phenology and water-state effects need local separation and calibration. Search-indexed abstract supports caution; full page was rate-limited during this review.Research paper, Remote Sensing 18(7),1055
National guidance existsIRRI provides rice-MRV resources and Vietnam-specific guidance. Actual project mapping still needs a local expert.IRRI guidance
Vietnam carbon frameworkDecree 119/2025 amends Decree 06/2022. A live legal review is required before contracting or transfers.Official decree
Personal data lawInclude Law 91/2025/QH15 and current implementing rules in the deployment review. Local hosting alone does not establish compliance.Official law

Team-confirmed facts

Muhammad Ilham Saripul Milah and Dzaky Zahy Rabbani are the named founders. The user confirmed no institutional funding, no revenue, no official pilot partner, and team-controlled rights to use the LuaZero code without an assignment obligation to a university/company. No patent or exclusive right to the underlying methods is claimed.

Retrieved satellite evidence

On 7 September 2026 we queried the public Sentinel-1 archive over An Giang (Microsoft Planetary Computer STAC, collection `sentinel-1-rtc`, anonymous access) and read real backscatter windows for every scene on relative orbit 18 descending between 1 November 2025 and 31 May 2026. The record holds 30 distinct acquisition dates, a median gap of 6 days, and 7 gaps of 12 days. Sentinel-1A, 1C and 1D all appear.

This is a measurement of data availability, and the only figure on the site derived from retrieved data. It is not a drainage detection. No ground truth, field boundary or land-cover check was performed, so no sampled pixel is confirmed to be a rice paddy; one of the five sample points is almost certainly built-up ground. The median gap of 6 days confirms the two-satellite repeat, and the seven 12-day gaps show that 6 days is a best case rather than a planning assumption. Method and reproduction steps: `satellite-observed-revisit.md`.

Evidence inventory

Implemented: calculator, input validation, editable synthetic water-log sample, exception review, Word/CSV/JSON export, a retrieved Sentinel-1 acquisition record, 14 automated tests and a static website. Planned: satellite classification of water state, field validation, Zalo integration and external assurance. The example report records synthetic provenance explicitly. No accuracy percentage is an achieved field result.

Responsible commercial interpretation

USD 3–6/ha/season and 50,000 ha × two seasons imply USD 300,000–600,000 annual service revenue. This is arithmetic under an adoption assumption. A USD 5 price means USD 500,000. A cost target just below USD 5 would not create a strong margin; the team must demonstrate materially lower delivery cost or higher willingness to pay. Optional success fees are not committed revenue and must respect benefit sharing.

Neither TCAF nor Gold Standard has endorsed this prototype. A national MRV workflow, a results-based government programme, voluntary credit issuance and an Article 6 transfer have distinct eligibility, ownership and accounting requirements. No funding, issuance or transfer is guaranteed by a software output.