The Federal Reserve’s proposed GENIUS Act rules address reserves, redemption, risk management, reporting and custody for supervised payment stablecoin issuers. Comments are due November 30. The implementation details will define the operational requirements.

ENISA says EU Member States are expected to provide at least one certified European Digital Identity Wallet by the end of 2026. The next challenge is institutional reliance: can a bank validate a credential’s issuer, meaning and assurance before using it for onboarding?

SEBI’s Demat 2.0 pilot for tokenized corporate bonds is worth watching. The key question is how tokenized ownership connects with investor records and existing securities workflows, particularly when records conflict or settlement fails.

Pontes enables wholesale tokenized asset transactions to settle in central bank money. This answers one crucial design question: what is the cash leg? The next challenge is coordinating asset ledgers, settlement systems and existing market operations reliably.

The FSB’s new note on global stablecoins highlights a structural problem: the arrangement can span multiple entities and jurisdictions. Risk monitoring has to follow how the system operates and where vulnerabilities can travel across borders.

Bank regulators are asking how banks control, audit and shut down agentic AI. The architectural test is containment: if an agent exceeds its scope, can the institution revoke access, preserve evidence and restore operations independently?

Separating institutional authority from transaction execution changes how banks evaluate blockchain networks. A bank should be able to assess an execution venue on its technical merits while requiring it to verify the institution’s authorization. Achieving that separation is an architectural challenge in its own right.

A transaction approval can be bounded to one decision. Duties such as operational resilience and ongoing conduct require attention over time. Regulatory architecture needs to represent both, with evidence of each determination and the period to which it applies.

A valid signature shows that a key signed. It does not establish whether the signer was entitled to commit the institution to that transaction. Custody architecture needs to account for the authority behind key use, especially when signing responsibilities are delegated.

Interoperability is often measured by whether assets move between networks. For institutions, the harder test is whether the authorization and restrictions surrounding an asset remain valid at the destination. Asset portability without governance continuity creates another integration problem.

In native fund tokenization, a central architectural question is which register legally establishes ownership. If the authoritative record moves on-chain, transfer controls and responsibility for correcting errors need to remain explicit. Changing the register changes the operating model.

An AI agent’s effective authority changes when its tools change. A policy approved at deployment may no longer describe what the agent can do after new integrations are added. Governance has to assess the permissions actually available when a consequential action occurs.

A token can move between execution environments in seconds. The institutional context around that token is harder to move. Authorization, restrictions, custody arrangements, and responsibility still have to remain coherent when the asset changes venue.

The harder problem in programmable finance is increasingly governance coordination. Execution is getting faster. Financial activity is becoming more distributed. The unresolved question is how institutional authority, policy, and legitimacy remain continuously verifiable across the systems where that activity occurs.

The self custody discussion is getting most of the attention. The deeper issue is whether an institution can demonstrate control around the asset regardless of where the key is held. Possession and authorized control are not the same thing.

We often define interoperability as the ability to move assets between networks. For institutions, the harder problem is preserving the meaning of the asset while it moves. The technology changes. The institution’s authority structure should not have to start over each time.

As execution spreads across Ethereum, Layer 2s, and application specific infrastructure, settlement becomes only one part of institutional control. A transaction can satisfy protocol rules and still lack institutional authorization. Execution and authority need to work together.

The SEC’s crypto custody proposal is interesting beyond the question of where assets are held. Institutional custody also involves authority, segregation of responsibility, transaction controls, and evidence. The key can sit in one place while institutional control depends on a much larger architecture.

A private key proves technical control. It does not prove institutional authority. As crypto custody enters regulated investment structures, the important question is how an institution defines, constrains, and verifies who is permitted to act over the asset.