Digital money rules tighten while AI wealth pitches wait for proof

On 8 October 2026, the 10-year US Treasury yield stood at 5.22%, and SOFR printed 3.87%. Those two rates sit at the centre of the digital-money question because they determine what a stablecoin issuer can earn on reserve assets, and how much it must pay to keep dollars credible.

The Federal Reserve Board has now requested public comment on two proposals related to establishing a regulatory framework for Board-supervised payment stablecoin issuers under the GENIUS Act. That is a concrete step toward defining who may issue dollar-linked tokens, what assets back them, and which supervisors can demand changes when redemption pressure arrives.

Stablecoins are becoming a balance-sheet business

The mechanism is direct. A payment stablecoin is a promise to exchange one digital dollar for one real dollar. If the issuer holds Treasury bills or overnight funds, higher short-term rates can turn that promise into a profitable float. But the same rates also raise the opportunity cost for holders and increase the pressure on the issuer to prove that reserves are liquid, valued correctly and available in stress.

The Fed’s proposals matter because they move that question from marketing language into supervisory language. Comment periods will shape capital, liquidity, disclosure and enforcement expectations for issuers that want a federal perimeter. Investors should watch whether the final rules treat stablecoins as narrow payment utilities, as bank-like deposit substitutes, or as something in between.

The useful way to read this convergence is to separate plumbing from product.

The European Central Bank is working on adjacent ground. The ECB published a speech by Piero Cipollone titled “Money in the digital age: digital euro, tokenisation and the role of central banks”. It also published remarks by Isabel Schnabel and Christine Lagarde that touch on blockchain settlement and AI risk.

The ECB’s work on a digital euro, tokenisation and AI risk sets rules for public money, market infrastructure and supervisory oversight. It does not amount to approval of any private wealth-management product that uses AI agents to select investments.

AI wealth management is a different risk stack

The private side of the story is moving faster than the public side. By one account, a former top HSBC executive has launched an AI-driven wealth-management platform powered by a team of AI agents. The same account says the founder has set her sights on a billion-dollar valuation.

Those are venture ambitions, not regulatory milestones. An AI wealth platform does not become safer because central banks are studying tokenisation, nor does a stablecoin framework make an automated portfolio manager more trustworthy. The risks sit in different places. Stablecoin regulation concerns payment promises, reserve assets and redemption. AI wealth management concerns suitability, model failure, data governance, conflicts, liability and the old problem of who pays when an automated decision loses money.

The valuation ambition is also a financing story, not a performance record. A billion-dollar target may help recruit talent and raise capital, but it tells an allocator nothing about drawdowns, client retention, model governance or the cost of human oversight.

The perimeter is being drawn around plumbing

The useful way to read this convergence is to separate plumbing from product. Central banks and the Fed are working on the plumbing: settlement rails, reserve assets, digital money and the treatment of new operational risks. Private AI platforms are building products that sit on top of that plumbing.

For investors, the near-term regulatory perimeter is likely to appear first around issuers, custodians and market infrastructure. That is where supervisors can measure reserves, audit code, demand capital and intervene before a run. Wealth advice is harder to supervise because the damage often appears client by client, through unsuitable allocations or opaque model decisions, rather than through a single balance sheet that can be examined on a Friday afternoon.

The open question is whether the two tracks will merge under one supervisory logic, or remain separate regimes that only collide during a failure. The next test will be whether stablecoin rules demand enough transparency to let investors see who holds the debt, and whether AI advisers disclose enough to let clients see who is on the other side of the trade.