Carbon credits, retirement, and the on-chain proof — in plain technical terms.

1. What a carbon credit is — precisely

A carbon credit is a transferable instrument representing one metric tonne of CO2e either reduced, removed, or avoided relative to a counterfactual baseline, certified by an accredited Validation/Verification Body (VVB) against a published methodology, and assigned a unique serial number in a recognised registry. Once retired against a claim, the serial is burned — it cannot be re-sold, re-claimed, or transferred. That uniqueness is the whole product.

The three verbs are not interchangeable, and the distinction governs price and defensibility:

• Reduction credits represent emissions that occur at a lower rate than baseline — e.g. a cookstove project under VMR0006 displacing non-renewable biomass. The atmosphere still receives carbon, just less of it.
• Removal credits represent CO2 physically extracted from the atmosphere — afforestation under AR-ACM0003, biochar under VM0044, direct air capture with geological storage (DACCS), enhanced rock weathering. Removals are the only category compatible with strict residual-emissions Net-Zero claims under SBTi's revised guidance.
• Avoidance credits represent emissions prevented from occurring — REDD+ forest-protection under VM0007 being the canonical (and most contested) example.

Each major registry imposes its own structural rules. Verra (Verified Carbon Standard) operates the largest pool, with methodologies ranging across VMs and AMs; Gold Standard tilts toward community-benefit projects with SDG co-benefit tagging; ART/TREES handles jurisdictional REDD+ at sub-national scale; Plan Vivo covers smallholder agroforestry; Climate Action Reserve (CAR) and American Carbon Registry (ACR) dominate North American compliance-adjacent supply.

Underpinning all of this is ISO 14064-2 (project-level GHG quantification) and the GHG Protocol Project Accounting Standard. A credit that cannot trace its quantification chain back to those documents — through a methodology, through a VVB report, through a registry-issued serial — is not a credit. It is a claim.

2. The Irish demand picture in 2026

Irish carbon-credit demand in 2026 splits across five identifiable cohorts, and each behaves differently on price, vintage tolerance, and methodology preference.

CORSIA matters disproportionately given Ireland's aviation footprint — Aer Lingus and Ryanair both fall within scope, and the ICAO Technical Advisory Body has narrowed eligible programmes significantly. Pre-2021 vintages are excluded from Phase I units.

SBTi's Corporate Net-Zero Standard v2 (consultation closed 2024, finalisation expected) is pushing Irish corporates toward removal-heavy portfolios for residual emissions, which has compressed supply on durable removals and inflated pricing relative to avoidance.

IMPT sits inside this picture as transactional commercial demand: every hotel booking processed retires one tonne, paid from our commission rather than added to the booker's invoice. At Irish booking volumes that aggregates to a non-trivial retirement count per quarter, all sourced from VCS and Gold Standard project pools and burned on-chain with serial-number traceability.

3. Quality signals — how to tell a good credit from a bad one

Carbon-credit quality is not a single number. It is the joint distribution across seven independent axes, and a project failing on any one of them should be treated as non-investable regardless of how it scores elsewhere.

• Additionality — would the abatement have occurred in the absence of carbon finance? The financial-additionality test (project IRR without credits below regional hurdle rate) is the strongest; barrier-analysis additionality is weaker and has historically been gamed in renewable-energy methodologies. Most pre-2020 grid-connected wind credits in India and Turkey fail current additionality tests.
• Permanence — geological storage (DACCS, mineralisation) is effectively permanent on human timescales. Biological storage (forests, soils) is reversible by fire, pest, drought, or land-use change. Buffer pools mitigate but do not eliminate the risk; Verra's AFOLU buffer pool has been pressure-tested by California wildfires.
• Leakage — does the project displace emissions elsewhere? Avoided-deforestation projects without effective leakage-belt monitoring routinely under-account.
• Baseline conservatism — the counterfactual must be defensible. Dynamic baselines (updated against jurisdictional reference levels) outperform static crediting baselines.
• Vintage — credits older than ~6 years carry rising ICVCM exposure and SBTi-eligibility risk.
• Methodology robustness — VM0048 (Verra's revised REDD+ methodology, 2023) is materially stronger than the legacy VM0007/VM0009 stack it replaces.
• VVB independence — the validator and verifier should rotate; the same body validating and verifying every period is a governance flag.

The ICVCM Core Carbon Principles, published March 2023, provide the current best umbrella assessment. The CCP-Eligible label requires programme-level and category-level approval. As of late 2025, the majority of legacy REDD+ methodologies were rejected, while ARR (Afforestation/Reforestation/Revegetation), ozone-depleting substances destruction, and selected cookstove categories passed.

4. The IMPT mechanism, in detail

The retirement pipeline is deterministic and externally verifiable. Each booking generates an event with the following lifecycle:

1. Booking confirmation — the hotel booking is settled, generating a booking ID and triggering a commission entry on IMPT's ledger.
2. Commission allocation — a fixed fraction of IMPT's commission is allocated to the carbon-retirement treasury. The booker pays nothing additional; the cost sits on our P&L.
3. Treasury draw — the on-chain treasury contract draws against the inventory of pre-purchased credits held in tokenised form, each backed 1:1 by an off-chain registry holding under a custody agreement with the bridge operator.
4. Registry retirement — the off-chain agent executes the retirement instruction in the source registry (Verra or Gold Standard), recording the booking ID in the retirement-beneficiary field. The serial number is captured.
5. On-chain proof — the serial number, project ID, vintage, methodology, and registry retirement URL are written to the booking's on-chain record. The token is burned.
6. Audit trail — anyone holding the booking ID can resolve to the on-chain record, and from there to the registry's public retirement page, where the serial number is independently confirmed as retired.

The verification primitives the buyer can run themselves:

• Look up the booking ID against the IMPT public ledger.
• Read the serial number, project ID, and vintage from the on-chain retirement record.
• Open the Verra or Gold Standard registry directly, search the serial, confirm retirement status, beneficiary, and date.
• Cross-check the methodology number against the project's PDD (Project Design Document) on the registry.

The mechanism deliberately produces no intermediary trust assumption. If the registry says the serial is retired against booking 0xABCD, that is the proof.

5. EU CRCF and what it changes for Irish firms

The EU Carbon Removals and Carbon Farming Regulation (CRCF) entered into force in December 2024. It establishes the first EU-level certification framework for carbon removals, distinct from the ETS and the voluntary market, and it will materially reshape what Irish firms can claim by 2027 when implementing acts and the certification methodologies become binding.

The framework defines four activity classes:

• Permanent carbon removal — DACCS, BECCS, mineralisation. Storage horizon >> centuries.
• Temporary carbon storage in long-lived products — wood in construction, biochar in concrete. Horizon > 35 years.
• Carbon farming — soil-carbon and biomass sequestration on agricultural land.
• Soil emission reduction — peatland rewetting, reduced mineral fertiliser.

For Irish corporates the practical implications are:

• Reporting separation — CSRD-reporting firms will be expected to disclose CRCF-certified units distinct from voluntary VCS/Gold Standard credits. Conflating them in Scope 1/2/3 disclosure is a flag for assurance providers.
• Permanence-class matching — temporary removals cannot be claimed against permanent residual emissions.
• Registry interoperability — the EU Union Registry extension for CRCF units is in design; the expectation is full serial-level traceability comparable to the ETS allowance system.
• Voluntary-regulatory boundary — voluntary credits remain valid for non-regulatory claims, but firms with EU regulatory exposure will increasingly need a CRCF-shaped portfolio for compliance-relevant reporting.

Irish firms with any CSRD scope, public procurement exposure, or aviation operations should be auditing their existing credit holdings against CRCF compatibility now, not in 2027.

6. What we do not recommend

The negative recommendation set is shorter and harder than the positive one. Avoid:

• Pre-2018 vintages — the bulk of pre-2018 forestry credits failed ICVCM CCP category review. Even where the underlying project was sound, the methodology baselines have not aged well, and SBTi guidance is tightening against legacy vintages.
• Credits under ICVCM red-flag methodologies — including the legacy VM0007 and VM0009 REDD+ stack, large-hydro CDM transitional credits, and several industrial-gas methodologies. Track ICVCM's published assessments; do not buy from a category mid-review.
• Registries without serial-level public lookup — if you cannot resolve a serial number to a retirement page in a public browser, the credit is unverifiable. White-label "carbon offset" products that cite no registry serial fall in this category by default.
• Sales without retirement guarantee — "we will retire on your behalf,