15c3-3 for Tokenized Securities: A Clear Guide

SEC Rule 15c3-3 is the possession-and-control rule that requires a broker-dealer to hold customer securities in a way that keeps them safe and separate from the firm's own assets. For tokenized securities, the same obligation applies, but the firm must now demonstrate control over a token recorded on a blockchain rather than a book-entry position at a central depository. That means proving who holds the private keys, reconciling onchain balances against the firm's books, and tracking corporate actions on assets that can move between chains and smart contracts.

Key takeaways

• Rule 15c3-3 exists to protect customers if a broker-dealer fails. It forces firms to segregate customer securities and cash so those assets are not tied up in the firm's own liabilities.
• The rule is technology-neutral in principle. A tokenized security is still a security, so the possession-and-control obligation does not disappear because the record moved onchain.
• Onchain custody reframes control as key management. Demonstrating possession means proving the firm, or a qualified custodian acting for it, controls the keys that can move the token.
• Reconciliation is the hard part. Firms must match onchain balances against internal ledgers continuously, across multiple chains and contract addresses, with no single consolidated tape to lean on.
• Market structure is early. Regulators, custodians, and infrastructure providers are still defining what counts as adequate control for tokenized assets, and much of the plumbing is emerging rather than settled.

Why this matters now

Tokenized securities are moving from pilots to production. Tokenized money market funds and Treasury products from established asset managers now trade on public blockchains, and several regulated venues and alternative trading systems are experimenting with settling securities transactions onchain. As real assets land on public ledgers, the question stops being theoretical: if a US broker-dealer offers customers access to a tokenized security, how does it satisfy the same investor-protection rules that govern a share of stock held at a central depository?

The stakes are concrete. Rule 15c3-3 was written after the paperwork crisis of the late 1960s, when brokerages drowned in physical certificates and customer assets went missing. It underpins the trust that lets people hold securities through an intermediary without fearing that a bankruptcy will vaporize their holdings. Extending that protection to tokenized assets is a precondition for institutions to scale onchain. For a fuller map of how these rules translate to blockchains, see our guide to onchain financial market infrastructure.

What Rule 15c3-3 actually does

Rule 15c3-3 has two core parts. The first is the possession-and-control requirement: a broker-dealer must maintain physical possession or control of all fully paid and excess margin securities that belong to customers. In practice that has meant holding those securities at a good control location, most commonly a customer account at the Depository Trust Company, the central depository founded in 1973 that holds US equities in book-entry form.

The second part is the customer reserve requirement. A firm must calculate, on a regular schedule, how much of customers' cash it is effectively using in its own business, and deposit at least that amount into a special reserve bank account that creditors cannot touch. Together, these provisions mean that if the broker-dealer collapses, customer securities and cash are ring-fenced and returnable.

This is the function to respect before reinventing it. The traditional system is fast, reconciled daily, and battle-tested across decades of market stress. The US moved to T+1 settlement in 2024, compressing the window in which customer positions sit in transit. Any onchain equivalent has to match that reliability, not just its speed.

How possession and control work for a tokenized security

When a security exists as a token, the depository model does not apply cleanly. There is no single DTC record; there is a smart contract on one or more blockchains that tracks who holds the token. Satisfying 15c3-3 then breaks into a sequence of concrete steps.

1. Identify the asset precisely. A tokenized security is defined by its chain, its contract address, and its token standard. Firms need a token security master that pins down exactly which token represents which legal security, because the same fund can be issued on multiple chains.
2. Establish control over the keys. Possession translates to control of the private keys or the smart-contract permissions that can move the token. A qualified custodian, or the firm under an approved arrangement, must hold those keys in a way regulators accept as a good control location.
3. Segregate customer tokens. Customer holdings must be kept separate from the firm's proprietary tokens, typically in dedicated wallets or segregated contract positions that map one-to-one to customer accounts.
4. Reconcile onchain against the books. The firm reads the blockchain to confirm that the tokens it claims to control actually sit in the wallets it controls, then matches those balances against its internal customer ledger.
5. Track corporate actions. Dividends, redemptions, splits, and transfers change the state of the position. The firm must capture these events onchain and reflect them in customer records the same day.

Reading the chain accurately is where the real difficulty lives. Onchain venues are fragmented across many blockchains, decentralized exchanges, and tokenized trading systems, and there is no single consolidated price tape or universal asset identifier today. To learn how firms source and standardize that raw ledger data, see our explainer on what onchain data is and how blockchain data providers work.

Traditional custody versus tokenized custody

What firms gain, and what they must build

Done correctly, tokenized custody changes the economics of holding securities in concrete ways.

• Faster settlement: capital is not locked for a business day waiting to settle. Onchain transfers can reach finality in minutes, freeing collateral that would otherwise sit in transit.
• Continuous reconciliation: instead of end-of-day batch checks, a firm can read the current state of the chain at any moment and confirm that customer wallets hold what the books say they hold.
• Programmable corporate actions: redemptions and distributions can execute through the contract itself, reducing the manual processing that traditionally introduces breaks and delays.

The catch is that every one of these benefits depends on reading fragmented onchain markets correctly. This is where a normalized data layer becomes load-bearing. Allium provides enriched, normalized, and labeled onchain data that institutions use to read tokenized and onchain markets, standardizing raw records from many blockchains into consistent verticals. Allium does not operate a venue, custodian, clearing house, or transfer agent; it is the read layer that lets a firm reconcile what the chain actually shows against its own books.

For adjacent context on how a normalized onchain data foundation gets built and used in practice, see how a stablecoin payments data foundation gets built and how onchain data is used to study perpetual futures.

The instruments driving the question

Much of the early demand for 15c3-3 clarity comes from tokenized cash-equivalent products. Tokenized money market funds and Treasury products are the most mature category, because their value is stable and their appeal as onchain collateral is direct. Our guide to tokenized Treasuries walks through how these instruments work. Stablecoins sit in a related but distinct bucket, and understanding what a stablecoin is and the risks they carry helps separate a payment token from a tokenized security. For the underlying mechanics, our overview of asset tokenization covers how a legal claim becomes a token in the first place.

Risks and open questions

The honest status is that tokenized-securities market structure is early, and several questions remain genuinely unresolved.

• What counts as a good control location? Regulators are still defining when a wallet, a smart contract, or a qualified custodian arrangement satisfies possession and control. The answer is not fully settled.
• Key management is a single point of failure. Control of a token is control of the keys. Lost, stolen, or compromised keys have no depository fallback, which raises the bar for custody engineering.
• No universal identifier. Without a CUSIP-equivalent that spans chains, firms rely on chain-and-contract mappings that can drift as assets bridge between networks. A consolidated onchain tape and a shared identifier standard are emerging efforts, not shipped products.
• Reconciliation under chain reorganizations. Blockchains can reorganize recent blocks, and firms must decide how many confirmations constitute finality before treating a transfer as settled.
• Legal enforceability. The link between the token and the underlying legal security must hold in bankruptcy. If a court cannot connect the onchain record to the customer's legal claim, the protection 15c3-3 promises could weaken.

None of these is a reason to dismiss tokenized custody. They are the specific problems being worked through as the market matures, and the firms solving them well will be the ones that treat possession, control, and reconciliation as seriously onchain as the traditional system does today.

Frequently asked questions

Does SEC Rule 15c3-3 apply to tokenized securities?

Yes. A tokenized security is still a security, so a US broker-dealer holding one for customers must satisfy the same possession-and-control and customer-reserve requirements. What changes is how the firm demonstrates control, which shifts from a depository account to cryptographic control of the token onchain.

What does possession and control mean for a token on a blockchain?

It means the firm, or a qualified custodian acting for it, controls the private keys or smart-contract permissions that can move the token, keeps customer tokens segregated from the firm's own holdings, and can prove all of this by reconciling onchain balances against its internal customer ledger.

Why is reconciliation harder for tokenized securities?

Onchain markets are fragmented across many blockchains, decentralized exchanges, and tokenized trading systems, with no single consolidated price tape or universal asset identifier. Firms must read blockchain state directly, map each token to its correct legal security by chain and contract address, and match those balances against their books continuously.

Is there a DTC or CUSIP equivalent for tokenized securities?

Not yet in a settled form. Traditionally DTC holds US securities in book-entry form and CUSIP identifies them. For tokenized assets, the asset is defined by its chain and contract address, and a consolidated onchain tape and shared identifier standard are emerging efforts rather than live infrastructure.

How does onchain settlement compare to T+1?

US securities moved to T+1 settlement in 2024, meaning trades settle one business day after execution. Onchain transfers can reach finality in minutes, subject to each blockchain's confirmation and finality rules, which is why firms must decide how many confirmations they treat as settled.

What role does onchain data play in meeting 15c3-3 for tokenized assets?

Firms need accurate, normalized reads of blockchain state to confirm what tokens they control and to reconcile against their books. A data layer such as Allium provides enriched, standardized onchain data institutions use to read tokenized markets. It is a read layer, not a custodian, venue, or transfer agent.