Top NFT marketplace royalties: do platforms protect creators?

The result is less a creator economy than a routing contest.

The top NFT marketplace conversation is still crowded with confident claims: royalties are dead; royalties are fully enforced again; a collection has a 5% royalty, so its treasury receives 5% on every flip. Most of that is noise. A royalty field in metadata is not a cash flow. A marketplace toggle is not a protocol guarantee. And a Discord announcement is certainly not settlement data.

We have to separate intent from execution. In 2026, platforms do protect creator earnings in specific technical lanes. They do not protect them universally. That distinction is where the real market sits.

A royalty percentage is a preference until the contract and the transaction route make it a condition of transfer.

ERC-2981 solved discovery, not payment

The industry still talks about ERC-2981 as if it were a royalty lock. It is not. ERC-2981, created in September 2020, standardizes a way for a marketplace or application to ask a contract: who should receive a royalty, and what amount should be calculated for a given sale price?

That is useful plumbing. It gives marketplaces a common royaltyInfo() lookup instead of forcing every collection to invent its own format. It improves interoperability. It helps indexers, front ends, aggregators, and marketplace contracts recognize creator preferences without decoding a bespoke contract every time.

But the standard does not compel payment.

ERC-2981 has no magical “send funds to creator” clause that fires on every resale. It does not know whether a transfer is a marketplace sale, an OTC settlement, a wallet-to-wallet gift, a bundled trade, or a transaction routed through another protocol. The marketplace—or a contract architecture designed around the marketplace—has to honor the returned royalty data.

That gap matters because NFT traders are not sentimental infrastructure users. If one venue makes a 5% creator payment optional and another makes it unavoidable, liquidity notices. In a thin-floor collection, a few percentage points can change which bid wins, where listings migrate, and how quickly a seller capitulates.

The practical distinction looks like this:

The implication for artists is blunt. “My collection supports ERC-2981” is not an earnings thesis. It is a compatibility statement.

For collectors, the implication is less obvious but equally material. A collection with protected creator earnings may have structurally different liquidity than one that relies on voluntary payments. This is not inherently bullish or bearish. It is a market design choice. But pretending the two trade under the same conditions is lazy analysis.

OpenSea’s shift: enforcement became a contract question

OpenSea has moved furthest toward a more explicit split between optional creator earnings and enforceable creator earnings. That is progress—not because it makes every royalty mandatory, but because it stops disguising conditional systems as universal protection.

For collections created in OpenSea Studio after April 2, 2024 at 10:00 a.m. PT, OpenSea says creator earnings can be enforceable. The same is true for custom contracts compatible with ERC721-C or ERC1155-C. The mechanism is tied to Seaport v1.6 and Seaport Hooks, which can require creator-defined conditions before an NFT transfer proceeds.

In plain English: the transaction can be designed so that payment of creator earnings is part of the route, rather than a polite suggestion displayed in a checkout interface.

That is a meaningful upgrade from the old marketplace-era model, where a platform simply read metadata, added a royalty line item, and hoped its competitors did the same. The older model depended on collective good behavior in a market whose participants are professionally trained to optimize execution.

Seaport Hooks change the conversation from “Will this marketplace honor my settings?” to “Can this sale path execute without satisfying the collection’s transfer conditions?” It is not a philosophical breakthrough. It is a mechanism. Mechanisms are where value capture either survives or dies.

Still, the qualifiers matter.

OpenSea does not claim that all collections have mandatory creator earnings. Non-upgradeable custom contracts without the relevant compatibility remain in the optional category. And even enforceable logic is not a universal net thrown over every wallet movement. It cannot establish that every peer-to-peer arrangement, unsupported protocol, or off-market settlement pays the creator.

This is the part that tends to disappear in social posts. Traders see “enforced royalties” and infer complete protection. Creators see “supported royalties” and infer recurring revenue. Neither inference is safe without looking at the contract type, the transfer restrictions, and the marketplace path.

OpenSea also caps the creator-earnings percentage a collection owner can set at 10%. That ceiling matters less as a recommendation than as a guardrail. There is no universal correct royalty rate for NFT art or PFPs. A 10% secondary take may make sense for a one-of-one artist with a thin but committed collector base; it can be a liquidity grenade for a high-turnover PFP trying to attract market makers.

Then there is earnings matching. OpenSea may reduce the percentage charged on its own platform if the same collection is configured with a lower percentage elsewhere. Its documented example is simple: 5% on OpenSea can be matched down to 3% when another marketplace charges 3%.

The cultural premium of enforced royalties is real. But markets price the cheaper executable route first.

Enforcement is not the same as insulation. If liquidity finds a lower-friction venue, the collection’s economics are tested there.

The OpenSea fee is not the artist’s fee

This should not require repetition, yet it does: platform fees and creator earnings are different line items.

As of September 15, 2025, OpenSea’s NFT platform fee is 1.0%. That is what the marketplace charges for facilitating the transaction. Creator earnings are the amount directed to the collection’s specified recipient, subject to the collection setup and the sale route.

The confusion is not harmless. When sellers complain that a trade has “too much royalty,” they often lump together the marketplace fee, creator earnings, network gas, and sometimes aggregator routing costs into one emotional number. That is understandable at the wallet-confirmation screen. It is useless for analysis.

A seller comparing net proceeds needs to isolate:

• the marketplace’s own transaction fee;
• creator earnings configured for the collection and actually applied to that trade;
• network gas and any execution costs;
• whether an aggregator or routing layer changes the total;
• the bid or listing price available after those costs.

For primary drops, the economics are different again. OpenSea typically charges 10% of the NFT mint sale price on primary drops, leaving 90% to the creator payout address designated for the drop. That is a primary-sale commercial arrangement, not a statement about secondary royalties. Mixing the two is how otherwise competent people end up publishing nonsense charts.

The right question is not “Does OpenSea support royalties?” It does, with a more serious enforcement lane for eligible contracts. The right question is: What is enforceable for this contract, on this route, at this point in its trading life?

That is less tweetable. It is also the only question that pays.

Magic Eden’s Solana model: optional is actually optional

Magic Eden’s Solana marketplace offers a different, more visibly market-driven model. It charges a 2% marketplace fee on transactions. That fee belongs to Magic Eden’s transaction economics; it is separate from creator royalties.

For Solana collections that have not adopted MIP-1, buyers can choose to honor 100%, 50%, or 0% of the royalty amount. The default is full royalties. The operative word is not “default.” It is “choose.”

This is the cleanest illustration of why UI defaults should never be mistaken for enforcement. A full-royalty setting can represent community sentiment, social pressure, or a marketplace’s preferred behavior. But when a buyer can select zero, the royalty is voluntary at execution.

Magic Eden says royalties are paid immediately after a sale when the trader elects to honor them or when they are enforced in the collection metadata. Collections that have adopted MIP-1 do not offer that buyer-choice selector. Again, the distinction is technical, not ideological.

A non-MIP-1 collection can carry all the visual signals of creator support—royalty information, marketplace display, a community norm—and still give the buyer a 0% payment option. A collection using an enforcement-capable metadata path can create a materially different trading environment.

Here is the practical contrast:

The market argument for optional royalties has always been straightforward: reduce friction, improve price discovery, let the buyer and seller negotiate value without a forced take. On liquid, speculative collections, that argument has teeth. A trader operating on tight spreads does not care that a creator’s roadmap PDF described long-term alignment. They care about execution.

The creator argument is just as straightforward: without secondary income, artists and teams subsidize the infrastructure, attention, and cultural labor that make a collection tradeable in the first place. That argument also has teeth, especially for digital art where the resale market can be the only durable revenue stream after mint.

Neither side wins by pretending the other is irrational. The real question is what a collection chooses to optimize.

A high-velocity PFP with fragile floor liquidity may decide optional or lower royalties give it better market access. A generative art project with curator demand and limited supply may accept less turnover in exchange for protecting creator economics. A brand-heavy collection may use primary revenue and licensing rather than rely on secondary fees at all.

These are not merely fee settings. They are market-structure decisions.

Discord says “support artists.” The chain asks “which route?”

This is where the public narrative gets especially soft.

In Discord, a community can overwhelmingly support royalties. Mods can encourage full payments. Holders can shame zero-royalty sellers. Influencers can frame royalty avoidance as extraction. All of that may create a social norm, and social norms can have value—particularly in art-led collections where provenance, patronage, and collector reputation matter.

But social consensus is not settlement enforcement.

On-chain, the decisive questions are colder:

1. What contract standard and transfer rules does the collection use?

ERC-2981 provides royalty information, but it does not enforce payment. Compatibility with an enforcement-aware setup is a separate matter.

2. Which marketplace or protocol is processing the sale?

A platform can only enforce conditions within the routes it supports. A collection’s OpenSea configuration does not prove equivalent behavior everywhere else.

3. Is this a standard marketplace transaction or a different kind of transfer?

Wallet-to-wallet transfers, private deals, OTC arrangements, and unsupported protocols do not automatically inherit marketplace fee logic.

4. What did the buyer actually select or what did the transaction require?

On Magic Eden’s eligible non-MIP-1 Solana collections, the buyer-choice setting can be 100%, 50%, or 0%. The default tells us nothing about the final selection.

5. Are we looking at a royalty setting or an actual payout trail?

A displayed percentage is intent. A completed transaction with the relevant payment output is evidence.

This is not an argument against creator economics. It is an argument against decorative analytics.

The best platforms for creator earnings are not simply the ones with the most artist-friendly branding. They are the ones whose contract compatibility, marketplace protocol, and transaction flow reduce the distance between an announced royalty and an executed payment.

That still leaves a hard ceiling on what marketplace architecture can accomplish. No current model demonstrated here guarantees collection across every possible transfer route. The dream of royalties that follow an NFT with perfect universality runs into the same ugly reality as every other on-chain rule: composability creates escape hatches, and enforcement creates friction.

What creators should choose—and what traders should price

Creators often approach royalties backward. They start with a percentage: 5%, 7.5%, 10%. The number is the least interesting part.

Start instead with the intended market behavior. Is the work a long-hold art asset, a PFP expected to trade constantly, a membership pass, a game asset, or a collectible whose value rests primarily on provenance? Is the goal to preserve resale participation, maximize liquidity, or direct more economics to the primary mint?

Then choose infrastructure that matches that answer.

For a creator evaluating a top NFT marketplace, the useful order of operations is:

1. Map the contract before setting the royalty.

Determine whether the collection can use an enforceable route, rather than assuming an ERC-2981 implementation changes payment behavior.

2. Model net trading friction, not just the royalty line.

Add platform fees, creator earnings, gas behavior, and the likely routing preferences of the collection’s actual buyers.

3. Treat marketplace compatibility as part of the artwork’s commercial design.

Metadata, provenance, transfer permissions, and royalty logic are not administrative afterthoughts. They shape who can trade and where.

4. Watch execution data after launch.

If holders consistently seek routes that avoid creator earnings, the market has delivered a signal. It may be a signal about royalty level, but it may also be a signal about weak demand, poor utility, or a collector base that was never aligned with the project.

5. Do not market optional payments as protected payments.

Collectors are not stupid. Sophisticated ones will inspect the mechanism. Less sophisticated ones will learn when their first resale settles differently from the announcement thread.

For traders, the inverse is true. A collection with enforced creator earnings should be valued with its actual exit friction in mind. That does not mean avoiding it. It means pricing it honestly. If the project has genuine cultural premium, respected provenance, and a collector base willing to pay for the ecosystem, royalties may be part of what sustains its market identity.

But if a collection has no meaningful artistic or social gravity, forced costs do not manufacture one. They merely sharpen the bid-ask pain.

The future is not “royalties win” or “royalties die”

The royalty war was always misnamed. It was never about whether creators deserve revenue. Of course they do. It was about whether a voluntary convention could survive in a permissionless market built to route around costs.

The answer is now visible. Voluntary royalty systems survive where culture, community, and venue design make them worth honoring. They weaken where liquidity is mercenary and alternatives are easy. Contract-level enforcement can materially improve the creator’s position, but only within the boundaries of compatible infrastructure and transaction paths.

OpenSea’s Seaport-based approach is the more serious attempt to turn creator earnings into an execution condition for eligible collections. Magic Eden’s buyer-choice model on non-MIP-1 Solana collections makes the trade-off explicit: full, half, or none. Neither system supports the fantasy of universal collection.

My verdict is uncomplicated. Creators should stop treating a royalty percentage as a business model. Traders should stop treating every fee as evidence of rent-seeking. And marketplaces should stop using vague creator-support language where the answer is technical and binary: is payment optional, enforceable on this route, or absent?

That is the signal. Everything else is marketplace noise.