CLARITY Is Not About "Legalizing Tokens"—Here's Why the Industry Still Needs It

Every time I see the U.S. CLARITY Act described as a law that “legalizes tokens,” I think the conversation has already drifted. The point of a serious market-structure law cannot be to make a word legal. It has to be to tell the difference between a real right, a record of use, an ordinary payment arrangement, and a thing that has been packaged to look like something else.

That distinction matters because useful technology has spent too long trapped in the same sentence as vague promotion and financial speculation. A digital object can prove access, record a service allowance, or attest to somebody’s eligibility. Those are concrete jobs, quite different from creating an impression through a story told around an object.

First, the status matters

H.R. 3633 is not U.S. law as I write this. The official record lists a Senate-reported version with an amendment dated June 1, 2026, after the Senate Banking Committee advanced it by a 15-to-9 vote. ^[1] That is the formal version I am discussing here. I am not treating later discussion text circulating online as if it were the same bill or a finished result.

What the bill proposes is a framework for the offer and sale of digital commodities, and for related intermediaries, across the SEC and CFTC. The practical question is not “may I put this on a ledger?” It is which regulator, registration, disclosure, anti-fraud, and anti-money-laundering responsibilities may attach to a particular activity. ^[2]

The bill does not create a general exemption from other law. Depending on the facts and on any express preemption or limitation, a particular activity may still implicate contracts, consumer protection, tax, privacy, sanctions, and other laws. This is not legal advice. It is a technology and product reading of why the boundary itself is important.

A token is not the right it records

One account at an airline can hold miles, elite status, upgrade certificates, tickets, and travel credit. They can all show up in the same app, but they are not interchangeable. This is an everyday way to see the problem; it is not an assertion that the Senate substitute itself uses these examples.

The table is not a legal classification chart. It is a product test. A restaurant card that grants a free meal after six visits is not the same object as a points program tied to dollars spent. A coupon that only one person may use this month is not the same object as an account credit that can be applied across services. When something goes wrong, the system can only reconcile the account if its rules have been clear from the start.

There are two different layers here. A right is what the holder can claim, such as a seat, a service upgrade, or a period of storage. A record says who received, used, transferred, or revoked that right, and when. A verifiable credential is a third thing: a checkable statement by an issuer about eligibility, for example that someone reached a membership tier. A credential is not the ticket. The ticket is not a payment. A token may simply be a way to identify, carry, or transfer one of these objects.

That is why the legal language needs more than the word “utility.” The formal Senate text does not create a universal legal category called a utility token. It proposes separate concepts and conditions, including digital asset, digital commodity, and network token. A proposed network token is not decided by a project declaring it one. The text ties it to use of a distributed-ledger system and names disqualifying financial rights, including debt, equity, liquidation rights, and expected interest, dividends, or other payments from a person. ^[2]

The House-passed version of H.R. 3633 used non-primarily-speculative loyalty points, rights, licenses, and tickets as examples of objects with significance beyond their existence as digital assets. The Senate-reported substitute discussed above is a different text. ^[3] That legislative history does not give every such object a legal result. It does show why a label and a data format cannot end the inquiry. Who issued it? What does the holder receive? Can it move? When does it expire? Can it be refunded? What does the issuer owe? Those are the questions that describe the substance.

An agent ledger should not turn every event into a payment

AI agents make the distinction more visible. This is why they belong in a CLARITY article: a product should not flatten many different rights, service units, and usage records into one vague “token.” A service may meter storage by GB-month, network traffic, GPU minutes, API calls, human-review capacity, or access to particular data. The delivery terms of a third-party service are a separate fact the system may need to record. Those units have different durations and conditions. At the end of the period, a billing system can settle them in currency, prepaid credit, or another contractual method. It does not follow that every usage record should produce its own payment.

An agent might act for a person when it accesses data. It might consume a particular quota during an inference. A third-party service may be passed through under specific delivery terms. A human or another agent may approve a step. These are business events that need to be attributable and checkable. They are not all separate transfers of money.

An ordinary database is often the right way to record them inside one company. I am not arguing that every meter needs a blockchain, or that every allowance should be a token. Signatures, verifiable credentials, and shared ledgers become interesting when several parties need to check the same facts across systems, or when audit requirements become stronger. A ledger cannot prove that an outside service was actually delivered. It can show who made a statement, which evidence was cited, when a rule changed, and where accountability should be traced.

Blockchain earns a place here when business objects need a clear origin, a rule set, and a record that several parties can inspect. If a token represents storage access, an event ticket, a revocable access credential, or a metered service allowance, its meaning comes from the business system’s ability to perform, not from its name.

Clarity has to survive a new label

The strongest objection is obvious. If an investment-like arrangement can be renamed a membership, a community benefit, or a use right, “clarity” becomes a new loophole. That is a fair objection. The case for this bill cannot rest on the word utility. It has to rest on whether the rules can deal with “new label, same substance.”

Under the proposal, the SEC (the “Commission” in the draft) may issue rules to prevent willful evasion. In adopting them, it may consider the totality of the facts and circumstances, including whether an arrangement’s principal purpose is to willfully circumvent the requirements while satisfying their literal terms and defeating their purpose; form, labels, and written documentation are not dispositive. ^[4] Its proposed nonfungible-token safe harbor is qualified in the same way. A safe harbor is a limited rule under which the offer, sale, resale, transfer, or conveyance of a qualifying NFT would not be deemed an offer, sale, or distribution of a security or investment contract, unless the transaction in substance has all the elements of an investment contract and subject to the stated exclusions. Membership rights, event tickets, and access credentials are only examples of possible non-investment uses. The text also excludes a mass-minted series of items with substantially similar or nearly identical traits that are marketed or sold interchangeably, a fractionalized interest, and a beneficial or economic claim on an NFT or an asset it represents. ^[4]

This does not prove the current draft has solved the problem. It leaves a harder question for product teams: what real right or record does this object represent? Who owes performance? Can it be transferred, divided, or revoked? What happens at expiry, when a refund is due, or when a dispute arises? Where do the metering facts come from, and who can verify them? If those answers are missing, a better token name will not make the business clearer.

I hope CLARITY eventually leaves less room to substitute a label for substance. It cannot create a right that a business never offered, and new technology should not make real use rights, qualifications, or service records suspect by default. Describe the object for what it is. That is more useful to technology and commerce than another token slogan.

References