Carbon markets and carbon-pricing systems have spent more than a decade building the back end: registries, reporting platforms, tracking systems and accounting frameworks. These tools are important, and many have improved substantially. But this is not where MRV credibility is ultimately decided.
The credibility question sits at the other end of the pipeline: the first mile, where the physical reality of an emission or a reduction becomes a data point. Logistics has a last-mile problem: getting the package to the customer. MRV has a first-mile problem: producing a reliable emissions number in the first place.
That first mile is where technology has moved fastest, where the cost economics have changed most, and where verification and accreditation have moved slowest. The result is the central asymmetry of MRV today: measurement and reporting are becoming digital, continuous and data-rich, while verification is still often organized around periodic site visits, manual document checks and conservative evidence rules.
This article examines that asymmetry across project-based carbon markets, the Article 6.4 mechanism and compliance systems such as CBAM. The settings are different, and they should not be treated as interchangeable. But they face a common practical question: when better data exists, what does the system recognize as credible evidence?
The first mile in plain terms
The downstream infrastructure of carbon markets is increasingly robust. Credits can be issued and transferred. Allowances can be tracked. Corresponding-adjustment and authorization processes are being built. But none of this solves the upstream question: how reliable is the original emissions or reductions number?
A conventional MRV process still often relies on reports, checklists, spreadsheets, manual evidence and periodic site visits. That approach can work in many settings. It also has limits. It is expensive, slow and difficult to scale when projects are highly distributed, when installations are outside established compliance systems, or when monitoring parameters are technically complex.
The first-mile problem is therefore not only a technology problem. It is a recognition problem. Digital tools can generate stronger evidence, but that evidence only matters if methodologies, verification bodies, accreditation bodies and regulators are prepared to accept it as primary evidence rather than as a supplement to the traditional audit file.
What the technology already delivers
Many core technologies are no longer research-stage. Continuous emissions monitoring systems, automated fuel-flow metering, industrial telemetry, IoT sensors, satellite imagery, drones and digital reporting platforms are commercially available. Their suitability varies by sector, project type and monitoring parameter, but the direction of travel is clear: the technical capacity to observe, record and transmit emissions-related data has improved faster than the rules for verifying it.
The technology is strongest where the monitored activity is instrumented, regular and data-rich. Industrial facilities, power generation, renewable-energy generation, pay-as-you-go energy systems and many distributed devices can produce frequent digital records. Satellite and remote-sensing tools can also materially improve land-use monitoring, especially for activity data such as land cover and land-use change.
The technology is weaker where the relevant parameter is harder to infer remotely. Forest biomass is the best example. Satellite and airborne remote sensing can strengthen forest-carbon MRV, but biomass estimates still require calibration and validation against high-quality field or reference data, especially in high-biomass and high-uncertainty areas. Digital MRV can improve the verification chain; it does not automatically replace ground truth.
The same caution applies to claims about automated baselines, end-to-end blockchain integrity and fully remote forest-carbon verification. These tools are promising, and some are already useful. But they are not yet a universal substitute for method design, calibration, quality assurance and professional judgment.
Projects and installations: different settings, related bottlenecks
Project-level carbon finance and installation-level compliance MRV are different systems. In a crediting project, the central question is how much a discrete intervention reduced or removed against a baseline. In an ETS, carbon tax or CBAM setting, the question is how much a regulated facility or imported good actually emitted. The evidence architecture, legal consequences and institutional responsibilities differ.
The first-mile challenge appears in both settings, but it appears differently. In compliance contexts, digital monitoring is more mature. The EU ETS has regulated the use of continuous emissions monitoring systems since the start of its third trading period in 2013, including quality-assurance requirements based on EN 14181 under the Monitoring and Reporting Regulation. The United States Acid Rain Program required CEMS under EPA rules in the 1990s. Regulated installations have therefore produced instrumented emissions data for many years.
Project-based markets are moving more unevenly. Some project types are naturally suited to digital monitoring; others are not. Highly distributed activities, such as cookstoves or smallholder agriculture, create difficult economics because sensors, platforms and verification systems must be deployed across many small units. At current carbon prices, advanced digital MRV can still be expensive relative to the credit value generated by each unit.
This is why the debate should not be framed as technology optimism versus traditional verification. The real question is proportionality: what level of evidence is required, what level of uncertainty is acceptable, who pays for the additional integrity, and how can the system avoid excluding smaller or lower-income-country projects from the market?
What market practice now shows
Two groups of carbon-crediting programs are visible. The first group consists of established programs that are adding digital MRV through pilots. In February 2026, Verra approved the first credits under its digital MRV pilot for high-frequency issuance, generated by the Foumbouni–Mitsamiouli solar farm in the Comoros, Verra Project 3788. Gold Standard approved three digital MRV pilots in 2025 covering electric cooking, biomass cooking and sustainable rice cultivation.
These pilots matter because they show that large incumbent programs can issue credits where monitoring systems are built around digital primary evidence. But they remain pilots. They do not yet change the operating model across the existing project base.
The second group consists of newer programs designed with digital evidence more deeply embedded in their architecture. Isometric and Puro.earth provide examples in durable removals; other programs and methodologies are moving in the same direction for land use, engineered removals and smaller distributed technologies. Their approaches differ, but the common feature is that digital data is not treated as an afterthought. It is built into the monitoring architecture, platform workflow and risk assessment.
Scale remains the decisive qualifier. The Berkeley Carbon Trading Project’s Voluntary Registry Offsets Database (current release v2026-04) shows the continued dominance of the large established registries by issuance volume. The programs that are architecturally most digital remain much smaller than the programs that carry most of the market volume. The center of gravity has not yet moved.
Why the V has not modernized
The technology has been ready in many areas for years. Costs have fallen. Yet digital MRV has not displaced conventional verification across high-volume programs. The reason is structural. Verification is governed by standards, accreditation practice, program rules and regulation. These systems are designed to be cautious. That caution protects credibility, but it also slows adoption.
The main international standard for verifying GHG statements is ISO 14064-3:2019. Verification bodies commonly operate under ISO/IEC 17029:2019 and ISO 14065:2020. These standards do not prohibit digital evidence. They are principle-based and allow risk-based evidence gathering. The problem is not that the standards ban remote or digital verification. The problem is how the sufficiency of evidence is interpreted in practice.
For many verification bodies, the safe professional choice remains the physical site visit, the manual check and the conventional audit file. A verifier that moves faster than its accreditation body, or faster than the program rules allow, risks challenge during witness assessment or surveillance. A verifier that stays conservative protects its accreditation but does not change the system. This is how a permissive standard can still produce conservative practice.
A related point is worth clarifying, because it is often misread: the DMRV-native programs do not operate outside the established accreditation stacks. They sit inside them. Puro.earth’s Validation and Verification Requirements require VVBs to be accredited under ISO 14065, ISO/IEC 17029, ISO 17065, ISO 17020 or ISO 14034, or to be UNFCCC-approved DOEs. Cercarbono requires validation and verification in accordance with ISO/IEC 17029:2019 and ISO 14065:2020. Rainbow (formerly Riverse) requires audits under ISO 14064-3 and ISO 14065 and VVBs accredited under the IAF MLA framework. In practice, the same ISO-accredited VVBs work across Verra, Gold Standard, Puro, Isometric and the rest. What the newer programs do differently sits one layer up — in methodology design, platform workflow and risk frameworks engineered so that remote-by-exception becomes the operating norm. The accreditation regime underneath is the same.
Article 6.4 shows both the progress and the remaining calibration problem. The Article 6.4 validation and verification standard for projects (A6.4-STAN-AC-003 v03.0) gives remote inspection an explicit place in the rules. Appendix 1 treats remote inspection as an alternative means to on-site inspection, using information and communication technologies under a risk-assessment framework. That is a meaningful advance.
At the same time, the same standard requires an on-site inspection at verification in specific cases, including the first verification by the DOE, more than three years since the last on-site inspection, or more than 300,000 tCO2e achieved since the last verification with an on-site inspection. The first two triggers are understandable. The volume trigger is more debatable. A utility-scale project with strong telemetry may be easier to verify remotely at large scale than a small distributed land-use project with weak ground data. The better variable is the verifiability of the first mile: measurement architecture, data quality and risk profile, not credit volume alone.
ISO has now moved further with ISO 14064-5:2026, published in February 2026, which provides formal ISO guidance on remote activities and techniques in GHG verification and validation. This is an important development. But guidance only changes practice when accreditation bodies, program operators and procurement requirements incorporate it. That transition will take time.
One objection any verifier or regulator raises early deserves a direct answer: if the measurement is automated, what stops the data itself from being manipulated in ways a remote reviewer would never see? The concern is legitimate and familiar from other regulated-measurement contexts. The honest answer is that digital MRV does not remove the integrity risk — it relocates it, from “did the auditor witness the activity” to “can the data stream be trusted, tamper-evidenced, and independently validated.” That is a manageable control problem with deliberate design: sensor calibration and chain-of-custody requirements, tamper-evident data capture, automated anomaly detection, and ingestion through controlled interfaces rather than self-reported feeds. The newer programs already do this in practice — Puro.earth builds a conservative safety margin into early monitoring steps, Verra’s pilot withholds a fifth of credits pending review, and Isometric runs automated checks on data arriving through its API. The World Bank’s dMRV guidance sets out system-evaluation criteria and a “hotspots” assessment aimed precisely at this risk. The objection is real, but it is an argument for designing the controls well — not for keeping the auditor on the plane.
CBAM: the compliance-side stress test
CBAM is now the largest live test of compliance-side first-mile MRV. Regulation (EU) 2023/956 established the mechanism. The definitive phase began on 1 January 2026, and the first annual CBAM declaration for 2026 imports is due by 30 September 2027. The Commission’s definitive-phase implementing acts set out verification principles, emissions-calculation methods, verifier accreditation and default-value rules.
The operational point is simple: many exporting installations are now being asked to provide installation-level emissions data in a form that EU importers can use for CBAM compliance. In many exporting jurisdictions, facilities producing CBAM-covered goods have not previously operated under verified facility-level MRV. The first-mile infrastructure is being built while the financial obligation is already live.
The definitive phase does not mean “verified data or no import” in all cases. Default values remain available where actual verified data is not available, but they are designed to be conservative and to create economic pressure toward verified actual data. This makes verified data a competitiveness issue, not only a reporting issue.
CBAM also shows how slow verification reform can be on the compliance side. Even in a regime designed for the mid-2020s, with many facilities equipped with modern telemetry, physical inspection remains central. Remote or virtual approaches may be allowed under defined low-risk conditions, but the default culture is still physical presence. That is precisely the asymmetry: digital measurement is increasingly possible, while formal recognition remains cautious.
Who pays for the first mile?
The cost issue cannot be avoided. Technology costs have fallen, but the economics remain difficult in many project categories. If the market demands stronger integrity, more frequent monitoring and better data, someone must pay for the first-mile infrastructure that produces it.
On the installation side, the cost is usually treated as a compliance cost. In many industrial settings, part of the monitoring infrastructure already exists for ETS, air-quality, operational or energy-management reasons. The marginal cost of CBAM-grade or ETS-grade reporting may still be material, but it is not always starting from zero.
On the project side, the economics are harder. A site visit, accredited verification and a verification report create a cost floor that does not scale down easily. Digital MRV can reduce some recurring costs, but sensors, connectivity, data platforms and system assurance can be expensive for small or distributed projects. Cookstoves are the clearest example: sensor-based usage tracking may improve integrity, but it can still be too costly at today’s credit prices unless deployed at very large scale or supported by buyers, donors or platform-based business models.
This creates a policy and market-design problem. If advanced MRV remains optional, high-integrity claims will remain uneven. If advanced MRV becomes mandatory without financing support or proportionality, smaller projects may be pushed out. The equilibrium must be explicit: higher integrity requires better MRV, better MRV costs money, and carbon prices must be high enough to support the level of assurance the market claims to want.
What needs to change
The first-mile problem will not be solved by another registry or another dashboard. It requires a coordinated shift in how evidence is generated, verified and recognized.
Program operators should treat high-quality digital data as primary evidence for suitable project types, not merely as supplementary information. Methodologies and VVB manuals need to make this operational, not exceptional.
Accreditation bodies should update witness practice so that verifiers using ISO 14064-5:2026-consistent remote techniques are not penalized for innovating within a risk-based assurance framework.
Regulators should avoid hard-coding physical visits where digital assurance can provide equivalent or better confidence. Physical inspection will remain necessary in many cases, but it should not be the default simply because it is familiar.
Buyers and financiers should recognize that integrity has a cost. If procurement demands better data, it should also support the MRV infrastructure that generates that data.
Technology providers should focus less on “end-to-end disruption” and more on auditability: calibration, data integrity, anomaly detection, controlled data interfaces and clear evidence trails for verifiers.
This is not an argument for removing professional judgment. It is the opposite. Digital MRV changes the verifier’s task. The question is no longer only whether the auditor saw the activity on site. It is whether the data stream is complete, calibrated, tamper-evident, consistent with other evidence and suitable for the claim being verified.
Conclusion
MRV credibility is no longer principally a question of what can be measured. In many sectors, much more can be measured than before. The harder question is what the system recognizes: in methodologies, in verification practice, in accreditation decisions and in regulation.
The first mile is where emissions reality becomes data. If that step is weak, every downstream system inherits the weakness. If that step is strong but the verification system does not recognize it, the benefit of digital MRV is lost.
The next frontier is therefore not only better monitoring. It is modernizing verification and accreditation so that credible digital evidence can scale. The M and the R have moved quickly. The V now has to catch up.
Source and version notes
The sources below are official or primary sources unless otherwise noted. Retrieval date: 3 June 2026.
Verra, “Verra Approves First Credits under DMRV Pilot for High-Frequency Issuances,” official announcement, February 2026; Foumbouni–Mitsamiouli solar farm, Verra Project 3788.
Gold Standard, “Gold Standard Drives Digital Transformation with Three New dMRV Pilots,” official announcement, February 2025; Decision Summary Q1 2025, publication date 15 February 2025, Pilot 1.0.
Berkeley Carbon Trading Project, Voluntary Registry Offsets Database, v2026-04, updated 26 May 2026.
ISO 14064-3:2019; ISO/IEC 17029:2019; ISO 14065:2020; ISO 14064-5:2026, “Guidance on activities and techniques used remotely in conducting verification and validation of greenhouse gas statements,” Edition 1, published February 2026.
UNFCCC Article 6.4 Supervisory Body, A6.4-STAN-AC-003, “Article 6.4 validation and verification standard for projects,” v03.0, entry into force 10 October 2025.
Regulation (EU) 2023/956 establishing CBAM, consolidated version; Commission Implementing Regulation (EU) 2025/2546 on verification principles; Commission Implementing Regulation (EU) 2025/2547 on emissions calculation methods; Commission Delegated Regulation (EU) 2025/2551 on accreditation requirements; Commission Implementing Regulation (EU) 2025/2621 on default values.
Puro.earth official document library: Puro Standard General Rules v4.4, May 2026; Validation and Verification Requirements v1.3, March 2026.
Cercarbono, validation and verification page; Rainbow (formerly Riverse), VVB requirements; World Bank, dMRV technical guidance note.
EU Monitoring and Reporting Regulation for the EU ETS third trading period, Regulation (EU) No 601/2012, including CEMS quality-assurance requirements based on EN 14181.
