An institution should take a tokenization pilot to production only when four things are settled in advance: the legal treatment, the production custodian, the operating owner and the economics. In the US, the regulatory gates moved in 2026. Within two weeks the SEC granted a conditional exemption, expiring after five years, for trading tokenized stock and proposed custody rules for advisers and funds. In our view, the harder gate is now internal ownership.
Who decides this
The board decides whether digital assets become a business line and sets the risk appetite and budget for it. The head of digital assets owns the roadmap and is accountable for moving each pilot to production or stopping it. That person needs authority over the operations, risk, finance and technology teams who will run the production service. Without that authority, the roadmap is a list of hopes.
What must a pilot prove before it earns production?
A pilot earns production when it has met exit criteria written before it started, covering legal treatment, custody, operations and unit economics. A pilot that only proves the technology works has proved the least important thing.
Legal treatment comes first, and tokenization does not change it. The SEC’s March 2026 interpretation puts “digital securities” in their own category: financial instruments already defined as securities, formatted as tokens, with ownership records kept wholly or partly on-chain. They are securities whether issued on-chain or off. A tokenized bond or fund share brings every existing duty with it, plus new questions about the ledger.
The remaining criteria are operational. Can the production custodian hold the asset under the rules that apply to you? Does on-chain position data reconcile with the books of record, every day, without manual repair? Is there a cost per transaction or per holder that beats the current process at production volume?
Our view: write the exit criteria as a one-page memo signed by the head of digital assets and the chief risk officer before the pilot launches. Set a date by which the pilot either meets them or stops.
Which regulatory gates are open as of October 2026?
In the US, two gates opened in 2026, one statute is phasing in, and two items remain at an early or proposal stage. Plan production around what is final and treat proposals as optional upside.
| Gate | Status as of October 2026 | What it means for a production plan |
|---|---|---|
| How securities law applies to crypto assets | SEC interpretation, effective 23 March 2026; joined by the CFTC | Five asset categories; tokenized securities stay securities |
| Trading tokenized listed stock | SEC exemptive order, 17 September 2026; expires five years after publication | Tokenized securities venues may trade tokenized stock in permissioned pools, under conditions |
| Adviser and fund custody of crypto assets | SEC proposal, 1 October 2026; 60-day comment period | Custodian choice for funds and advisers may widen; not yet final |
| Payment stablecoins | GENIUS Act enacted 18 July 2025; stablecoin provisions take effect at the earlier of 18 months after enactment or 120 days after final regulations | A regulated cash leg for on-chain settlement is coming into force |
| Retail crypto commodity transactions (CFTC) | CFTC advance notice, 5 October 2026 | Early stage; comments sought, no rule proposed |
Two cautions apply. The interpretation is guidance, not a rule, and the SEC has said it may refine it. And this table is US-only. Our guide to tokenization strategy for a financial institution, linked below, covers how other regimes treat the same assets.
Where does custody sit when the pilot ends?
Pick the production custodian during the pilot, not after it, because custody is the gate most likely to send a working pilot back to the start. A pilot can run on a test environment that no regulator has examined.
The SEC’s custody proposal of 1 October 2026 shows how the US rules may change. It would set a tailored framework under the Investment Advisers Act and the Investment Company Act for registered advisers and regulated funds. It would allow state trust companies to act as custodians for client and fund crypto assets, and would permit self-custody in certain circumstances. It would also update auditing requirements for advisers and broker-dealer custody requirements for funds.
It is a proposal, open for comment for 60 days after publication in the Federal Register. A production plan that depends on it needs a fallback custodian that qualifies under today’s rules.
Our view: run the pilot on the custodian you expect to use in production, even at higher cost. Have it confirm in writing which assets, chains and wallet types it will support at scale.
Public chain or permissioned network?
The SEC’s tokenized-stock exemption points to a hybrid: public, auditable contracts with permissioned access to trading. That is a useful signal for any institution choosing infrastructure in 2026.
Under the innovation exemption of 17 September 2026, tokenized securities venues may trade tokenized NMS stock through permissioned automated market maker liquidity pools. Venues set the access standards for participants. The smart contracts must be auditable, public and deployed on a public, permissionless ledger.
The other conditions read like a production checklist. There are limits on the number of symbols and on volume traded. The venue must verify that tokenized shares carry the same rights as the traditional shares of the same class. An issuer gets written notice and a chance to object before an unaffiliated third party’s token of its stock is listed. Trading halts when the primary listing exchange halts. The venue must publish information about its operations and its affiliates’ trading.
The trade-off for an institution is reach against control. A public chain gives interoperability and outside verification; a private network gives control over who sees what. The SEC’s order suggests regulators can accept public infrastructure if access, rights and disclosure are controlled at the edges.
How should the roadmap be sequenced?
Sequence by dependency, not by ambition: the asset and its legal treatment first, then custody, then distribution, then the on-chain cash leg. Each stage needs a named owner and a stop rule.
In practice that is four stages. First, one asset type with a settled legal treatment and an existing client base. Second, production custody and daily reconciliation against the books of record. Third, distribution to a defined set of clients through channels compliance has approved. Fourth, on-chain settlement with a regulated cash token, once the stablecoin rules apply to the issuers you would use.
The fourth stage has a date to plan around. The GENIUS Act’s stablecoin provisions take effect 18 months after its July 2025 enactment, or 120 days after final regulations if that comes first. That puts the latest start in January 2027. Until the issuers you would use are operating under those rules, keep the cash leg off-chain and design the interface so it can switch.
The board’s part in sequencing is to hold the stop rules. Ask for a one-page report at each stage gate: criteria met, criteria missed, cost to date and the decision requested. A board that only hears about pilots at launch events cannot stop the ones that are not working.
Each stage should leave something behind that the next one needs: a classification memo, a custody agreement, a reconciliation report, a client-onboarding record. If a stage leaves nothing behind, it was a demonstration.
Our view: resist adding a second asset type until the first has run in production for a full reporting cycle. In our experience, the problems show up at month-end, not on launch day.
What goes wrong: the strategy nobody could execute
The common failure is a sound strategy that no one inside the institution was resourced to run. An innovation team builds the pilot with a vendor. It works. Then it reaches the operations, risk and finance teams, who were never asked to own it and have no budget for it.
The pilot sits in a sandbox while the vendor contract renews. A year later the board asks why the programme has produced press releases but no revenue. Nobody can say who was meant to take it to production, because the strategy never named them.
The fix is to write production ownership into the strategy before the pilot starts. Name the operating owner, budget the production run, and give the head of digital assets authority over the teams the service depends on.
How Alcyone approaches it
Alcyone builds roadmaps around ownership. Its engagements start with a thirty-minute call with a senior partner, then a fixed-scope, fixed-fee assessment in week two. A plan follows over three to six weeks, with a named owner for every part. The same partners stay through the launch window and report quarterly afterwards; how an Alcyone engagement runs sets out each stage. One published case is directly relevant: for a national banking association, Alcyone helped deliver a compliant, live sustainability NFT initiative in a regulated environment.
For institutions whose roadmap includes issuing a token backed by reserves or real-world assets, TrustSwap’s Panthea sets out its reserve and custody set-up for issuers. Panthea provides the technology layer; reserves are held in segregated accounts at Blockfort AG, a Swiss-regulated financial intermediary and VQF member. Any custody arrangement, this one included, should be checked against the custody rules that apply to the institution before it is chosen for production.