NFT vs ERC-20: Why Carbonized Uses Both for Carbon Credits

TL;DR: Each verified credit has its own project, vintage and serial number, which is naturally a job for an ERC-721 (NFT). Trading is easier with interchangeable units, which is a job for ERC-20. Carbonized lets you hold the unique credit as an NFT and wrap it one-to-one into a pooled ERC-20 for trading, then unwrap it again. Retirement happens on the NFT.

"Should carbon credits be NFTs?" is a question that usually triggers either hype or eye-rolling. The honest answer is: it depends on which property you care about at which step. Carbonized uses both token standards, each for what it does well.

Why credits are not naturally fungible

Two credits may both be one tonne of CO2e, but they can differ in ways buyers care about:

• Methodology: Verra VCS, Gold Standard, or another programme
• Project type: removal, avoidance, nature-based, engineered
• Vintage: the year of the reduction
• Geography and co-benefits
• Serial number: the identifier in the issuing registry

A unique token per credit preserves all of that. It also gives you something valuable for integrity: a one-of-a-kind object cannot be accidentally merged with another. This is the same reason registries assign serials in the first place.

What ERC-721 gives you

ERC-721 is the non-fungible standard. Every token has a unique ID and exactly one owner.

• Metadata travels with the token. Carbonized stores the project, methodology, vintage, tonnes, serial and an oracle proof hash in the credit record.
• Uniqueness is enforceable. The contract hashes the canonical serial number (trimmed and upper-cased; the methodology label is not part of the key) and refuses to mint the same serial twice, a direct defence against double issuance. See double counting.
• Retirement is clean. Burning one specific token while storing a certificate tells a precise story: this credit, for this reason, by this wallet.

The cost of ERC-721 for trading is friction. A buyer who wants 500 tonnes would need to buy many individual items, each with its own price. Liquidity fragments into thousands of one-off listings.

What ERC-20 gives you

ERC-20 is the fungible standard: balances, not individual items. Any one unit equals any other.

• Divisibility. Trade 0.37 tonnes if you wish.
• Simple pricing. A single price per unit across a pool makes order books and automated market makers practical.
• Composability. Wallets, exchanges and other contracts already understand ERC-20.

The cost of pure ERC-20 is the loss of identity. If you pour all credits into one fungible token without discipline, a low-quality credit and a high-quality one become indistinguishable, and the market tends to price at the quality of the worst unit. That is the risk of any commodity pool.

How Carbonized combines them

Carbonized's flow is deliberately layered:

1. Mint: an authorised issuer mints a CarbonCreditNFT for a reduction it reports. The contract checks only that the serial is new and non-empty and that the oracle proof hash is non-zero; it does not independently verify the reduction.
2. Wrap: the owner calls wrapCredit(tokenId). The NFT is locked in the token contract, and Carbon Credit Token (CCT) is minted equal to the NFT's tonnes. Amounts use 18 decimals, so 1e18 units is one tonne.
3. Trade: CCT trades against a payment token on the marketplace through an order book and auctions.
4. Unwrap or retire: burning CCT via unwrapCredit returns an NFT; only the NFT can be retired onchain.

See the contract references for CarbonCreditNFT and CarbonCreditToken, and the architecture overview.

One pool per methodology

To avoid the "worst unit sets the price" trap, there is one CCT contract per methodology pool, for example VERRA_VCS and GOLD_STANDARD. A pool accepts only NFTs of its own methodology; wrapping a mismatched credit reverts. This keeps each pool's quality definition coherent, although inside a pool vintage and project differences are still averaged out. That is a real trade-off, and a more granular pool design, say by vintage band, is something worth exploring.

Fully backed, one to one

Each CCT is backed by tonnes actually locked in NFTs. The contract exposes totals, such as tonnes wrapped and supply, so anyone can check the 1:1 relationship. Carbonized's audit page reconciles event-derived totals with contract counters and shows a match or mismatch.

Design trade-offs worth knowing

Unwrapping returns an NFT, not necessarily your NFT

When you unwrap, you burn CCT and receive an NFT from the locked set. Depending on the function used, you may choose a specific token or take any available one. In either case, you may not get back the exact credit you deposited. That is the nature of pooled fungibility and a reason high-quality specific credits might prefer to stay as NFTs and be sold through an auction.

Rounding and dust

Because CCT is divisible but NFTs are whole items, holders can end up with fractions of a tonne that cannot be unwrapped alone. It is a small annoyance, but real.

Gas costs

Minting and wrapping individual NFTs costs more gas than moving ERC-20 balances. On an L2 or similar low-fee chain, this is manageable, which is a reason we target Robinhood Chain Testnet, though fees there say nothing about future mainnet conditions. See the launch post.

Retirement path

Only the NFT can be retired onchain, because the certificate belongs to a specific credit. If you hold CCT and want to claim an offset, you unwrap first. A pooled token that did not preserve provenance at retirement would be a weaker claim. The retirement post explains why.

Alternatives we considered

• ERC-1155 (multi-token). It can hold both fungible and non-fungible types in one contract and supports batching. It is a credible choice, and a future migration is possible, but ERC-721 plus ERC-20 has the widest tooling support today and clear semantics.
• Pure ERC-20 per vintage or project. It preserves fungibility within a narrow class but multiplies the number of contracts and fragments liquidity.
• Only NFTs plus a marketplace. This works for high-value, low-volume credits, but not for volume trading.

When each token is the right choice

Takeaway

NFTs and ERC-20s are tools, not ideologies. Credits are unique at issuance and retirement, and fungible in between, so Carbonized models both ends with an NFT and the middle with a wrapped token. The price of that flexibility is extra complexity and some pooling trade-offs, which we try to be open about.