Crypto wanted to change finance. Now banks, stock exchanges and fund managers are turning up with assets to put on the blockchain.
A slightly awkward reunion, given some of the things we've said about them.
The move is becoming easier to see: stablecoins for payments, tokens representing company shares, funds holding bonds and other assets, and AI agents helping run our finances.
If people hold dollar stablecoins, buy investments with them and give their agents dollar budgets, each activity makes the others easier. That could extend dollar dominance into the next financial system. How much room does it leave for other currencies?
Spain had the silver
I enjoyed reading Why Nations Fail, by Daron Acemoglu and James Robinson. Their argument about institutions is useful here: who holds power, who can participate, and whose interests the rules protect all affect whether wealth produces lasting prosperity.
In chapters 7 and 8, they contrast England's constraints on royal power with Spain's more absolutist path. Spain had wealth flowing in from the Americas, but the crown's control of trade helped concentrate the benefits. England gradually made more room for competing commercial interests. This is their interpretation of how institutions shaped the balance of power. Related research.
Philip II stopped payments to his lenders 4 times, although lending resumed. Having the silver did not make the finances look after themselves. Drelichman and Voth.
Today, the US has a different advantage: a currency people already want, increasingly distributed by private stablecoin companies. Other countries can adopt the same technology and still find their citizens choosing dollars.
The historical parallel is about who turns an advantage into lasting influence. If dollar-based services become the default across borders, how much financial power moves with them? And can the US maintain the trust that makes this work?
Your money, shares and investments in the same wallet
Bitcoin, ETH and other cryptocurrencies already live on-chain. The migration brings more of the financial world alongside them, in 4 connected parts.
Stablecoins. Tokens such as USDC aim to hold a value of $1, letting people send dollar-denominated payments between wallets. Their usefulness depends on the issuer's reserves and ability to redeem them.
Equities. DTCC has run live trades using tokenised assets, and the LSE is exploring tokenised equities with Payward. At Gryps.Finance, which I'm helping build, we approach this through the execution of equity-linked perpetuals: contracts giving price exposure without ownership of the shares. Buying the token and owning the company are not always the same thing. DTCC, LSEG, Gryps explanation.
Other real-world assets, or RWAs. Bonds and funds are following. Franklin Templeton's BENJI fund already records ownership on a public blockchain. Shares technically belong in this category too. Franklin Templeton.
Agents. The ambition is to delegate much of your financial admin: budgeting, paying bills, moving spare cash and managing investments within limits you set. Different products can do different parts of that today.
Imagine being paid, investing some of it and covering your bills through connected services, with an agent handling the steps. That is the appeal, even while access restrictions and incompatible systems make it messier in practice.
What happens when everyone's agent withdraws?
Crypto is agent-native. Blockchains and AI play well together because software can hold a wallet, sign transactions and interact with financial services through code. The interesting step is giving an agent permission to use those capabilities within limits you set.
Grok Bot's new finance connection is read-only. Meta's Muse uses connected accounts for financial guidance. Robinhood Agents, announced on 29 September and coming soon to eligible US customers, will use dedicated accounts with user-controlled trade approvals. These products show different levels of delegation; they don't all depend on crypto. Grok coverage, Plaid and Muse, Robinhood.
At Gryps.Finance, we're developing Gryps Agentic. Our public tools currently provide read-only market data; agentic execution is still a work in progress. Public Gryps tools.
Now imagine agents gaining permission to move deposits when a bank looks unsafe. Thousands could withdraw together. An agent-driven bank run is the risk: decisions that make sense individually can destabilise the institution everyone is leaving.
ECB-published research explored a related risk in simulated investment-fund withdrawals. Some reinforcement-learning agents produced extreme run-like behaviour; language-model agents were less prone to it. The study shows a possible mechanism, not an actual agent-driven bank failure. ECB research.
Your agent probably won't stay with a bank because it gave you a free railcard at university.
The dollar keeps appearing
The BIS estimated that about 98% of stablecoin value was dollar-denominated in May 2026. BIS analysis.
In TRM Labs' x402 payment dataset, 99.6% of settlement value was USDC. Most activity wasn't demonstrably agentic, so this doesn't tell us what all agents prefer. It does show a dollar-heavy payment system they could use. TRM Labs.
You hold dollar stablecoins, buy tokenised investments with them, receive dollar proceeds and give your agent a dollar budget. Each step makes the next dollar-based service more convenient. More buyers attract sellers, and developers build for both. The network effect compounds.
Issuers can also hold short-term US Treasuries under the GENIUS Act's reserve framework. Growth can support government-debt demand, although some purchases replace holdings elsewhere. It gives the US another funding channel, without making its debts disappear. Kansas City Fed.
For a movement with a fairly dim view of central banks, we've developed quite an appetite for dollars.
America has been good at distribution for a while
Hollywood, music, universities and global brands have helped make American culture familiar and desirable abroad. Joseph Nye described this attraction as part of American “soft power”, much of it created outside government. Nye on culture and brands.
I see a parallel in stablecoins: private companies making something American easier to access and use. Marketing and distribution reinforce an existing advantage. The reserves, redemption and underlying economy still have to work. A good logo will not rescue bad reserves.
You can dislike US foreign policy and still prefer dollars for your savings. That is quite a distribution advantage.
Europe, I'd quite like us to get on with it
The US backed private stablecoins and halted federal CBDC work through its January 2025 executive order. Europe is pursuing a digital euro, with potential issuance in 2029 dependent on legislation. US policy, ECB timetable.
What frustrates me is the gap between the euro's importance in everyday European life and its small presence in stablecoins. Europe has euro stablecoins and initiatives for tokenised bank deposits and wholesale settlement. The digital euro is a separate project for everyday payments. I'd like these efforts to produce services people choose before dollar-based habits become harder to change. ECB on euro digital finance.
I also want clear limits on what a central bank digital currency lets authorities see or control. The ECB says its design would offer offline cash-like privacy and would not be programmable money. Those protections matter. ECB FAQ.
China and its partners have explored alternative settlement through mBridge; Brazil's Pix shows domestic payments can improve without a blockchain. Neither automatically changes the currency people want to hold. mBridge, Pix.
For weaker-currency countries, easier dollar access can protect individual savers while drawing funds from local banks. What helps the customer can leave their central bank with less influence. BIS discussion.
Could countries choose crypto instead?
Could dependence on dollar stablecoins push countries towards Bitcoin, ETH or another cryptoasset?
Possibly, for some reserves or settlement arrangements. That is a scenario worth considering, although holding an asset, using its network and pricing an economy in it are very different commitments. A government could use Ethereum while continuing to settle in dollar stablecoins.
Bitcoin offers supply rules outside government issuance. ETH has a role in operating Ethereum. Neither automatically offers the price stability people want for wages and bills. More blockchain use does not settle which currency wins.
Belief and network effects matter here too. People accept money partly because they expect others to accept it tomorrow. Scarcity needs demand, and demand becomes more durable when there are useful things to do with the asset.
Memecoins make that shared belief rather obvious. The dog in the logo isn't necessarily the problem. Explaining to your landlord why this month's rent depends on its popularity might be.
Would a country accept crypto's volatility to reduce dependence on the dollar? Would its citizens make the same choice?
Who wins the currency contest?
My current view is that the dollar has the strongest starting position. Agents will work around the currencies their users earn, spend and invest in. If dollar services connect those activities most easily, automation could deepen that advantage.
That depends on continued trust in the dollar and access to those services. Better alternatives, restrictions or a loss of confidence could change the picture.
For the euro and other alternatives, the challenge is giving people a practical reason to switch. Better marketing wouldn't hurt either. A central bank consultation paper is a tough sell next to Hollywood.
What would make you give your agent a euro budget instead?