Money Stuff: Stablecoins Are Growing Up

I have always thought of stablecoins as basically unregulated non-interest-paying crypto banks. The idea of a stablecoin is that there is a company, the stablecoin issuer, and you send the company dollars, and it gives you back tokens (“stablecoins”). Each token is supposed to be worth a dollar, and if you hand the issuer back a token it will hand you back a dollar. [1]  But mostly you hold onto the stablecoins and the issuer holds onto your dollars. You both arguably benefit from this:

• You can use the tokens like dollars on in crypto applications, where actual dollars are hard to use: The stablecoins live natively on various crypto blockchains, can be sent over the blockchain, can be used in smart contracts, etc., in ways that don’t really work with dollars held in bank accounts.
• The issuer has the dollars, which it can invest in super-safe short-dated US Treasury bills, or anything else, to make money. It doesn’t pay you any interest; any profit it makes from its investing is the issuer’s to keep.

But yesterday Circle Internet Group Inc., the big US-based issuer of the USDC stablecoin, filed publicly for an initial public offering. It has previously filed confidentially, but its prospectus is available now. It is an interesting read, and it has caused me to update my model of stablecoins in two respects:

1. Circle is not an unregulated bank; it is a tech front-end for US banks and asset managers. [2]  
2. I am not as sure as I once was about “non-interest-paying.”

On the first point. One way to do the stablecoin issuing model is to issue stablecoins, take dollars, and use the dollars to invest in whatever you want. If you are a big stablecoin issuer in even a moderately high interest-rate environment, “whatever you want” should definitely be exclusively US Treasury bills, because you can earn piles of money with no risk and continue living a really good life. But we have talked about how Tether, the biggest stablecoin issuer, has sometimes had more peculiar tastes. Never mind that, though. The point is that you have some money, you have to invest it in some stuff, you decide what the stuff will be and you go out and buy it.

That’s not quite how Circle invests. Circle does not go out and buy stuff with its depositors’ money: Circle takes that money and deposits it in (1) banks and (2) a Treasury money market fund run by BlackRock Inc. From the prospectus:

Prior to January 2023, USDC reserves were limited to cash balances held at banks, U.S. Treasury securities with maturities of 100 days or less, and the Circle Reserve Fund. Since January 2023, USDC reserves have been limited to cash balances held at banks and the Circle Reserve Fund. ...

As of December 31, 2024, approximately 85% of USDC reserves are held in the Circle Reserve Fund. … The Circle Reserve Fund is a government money market fund pursuant to Rule 2a-7 under the 1940 Act, holding a portfolio of U.S. Treasury securities with remaining maturities of three months or less, overnight U.S. Treasury repurchase agreements, and cash. …

The Circle Reserve Fund is managed by BlackRock. It is only available to Circle, and Circle is the only shareholder of the Circle Reserve Fund. … The Circle Reserve Fund invests at least 99.5% of its total assets in cash, U.S. Treasury bills, notes, and other obligations issued or guaranteed as to principal and interest by the U.S. Treasury, and repurchase agreements secured by such obligations or cash, with any securities generally maturing in 397 days or less and the portfolio having a dollar-weighted average maturity of 60 days or less and a dollar-weighted average life of 120 days or less.

That is: Before 2023, Circle invested some customer money directly in Treasury bills, but it has stopped. Now some of the money is in bank accounts, but most of it is in a money market account at BlackRock. BlackRock invests the money in Treasury bills, but BlackRock is BlackRock. It is a giant regulated US financial institution that runs money market funds. 

Tether is, arguably, a new kind of financial institution, a big company that convinced people to give it money that it can then invest however it wants. Circle is a fintech; it is a technology platform that takes money from customers and funnels it directly to regulated financial institutions, without doing direct investing of its own. This is probably more regulatorily soothing than the unregulated banking model. 

On the second point. BlackRock and the banks pay interest to Circle, which does not pay interest to holders of its stablecoin. (It does pass some of the interest on to Coinbase Global Inc., the crypto exchange, for distributing USDC.) Seems like a nice business. But Circle has other products:

Circle Tokenized Funds are regulated yield-bearing investments for collateral use in capital markets. We believe that certain major trading firms have moved, and will increasingly move away from, using stablecoins as collateral in favor of TMMFs [tokenized money market funds]. However, other TMMF issuers may be limited in their ability to provide significant instantaneous redeemability for equivalent underlying assets. We believe that the ability of a well-regulated stablecoin such as USDC to provide near-instantaneous redemption can offer a substantial advantage for this emerging trend of using TMMFs as collateral for margin trading on various digital asset trading platforms, maximizing capital efficiency in the flows between settlement assets and collateral and at the same time, reducing counterparty execution risk. Moreover, TMMF issuers typically operate on a single blockchain or have limited interoperability across different blockchain platforms. By integrating TMMFs with Circle’s existing infrastructure, we expect to enhance interoperability and enable investors to quickly transition between TMMFs and stablecoins, allowing traders to hold their assets in TMMFs while accruing yield and then be able to instantaneously convert into Circle stablecoins, which can then be used to purchase other digital assets on third party platforms.

I think the way to read that is roughly “institutional crypto trading firms are not going to hold all their cash in non-interest-bearing stablecoins, so we are offering them interest-bearing stablecoins.” (Technically: interest-bearing tokenized money market funds that are instantly convertible to stablecoins to do trades.) We have talked a few times about tokenized money market funds, which are in every relevant respect interest-bearing stablecoins. As we have discussed, there is some regulatory complexity here: Money market funds are securities, so you have to comply with US securities laws to issue interest-bearing stablecoins. But, you know, BlackRock can do that; that’s fine. Circle can probably do it, can combine the two elements of (1) operating a US regulated money market fund and (2) operating it as a useful stablecoin. [3]

I am sure that, with crypto roaring back and with regulators increasingly friendly to crypto, that will be a good business. But the old model of stablecoins is a ludicrous business: You take tens of billions of dollars from institutional investors, you invest it however you want, and you keep all the interest. That was great, but it probably can’t last.

Well here’s this:

“Low-key gonna break down ur investment plan rn,” the Arta assistant says, responding to a client’s query on his investment portfolio. “No cap, ur portfolio is fire!”

Yes. “No cap,” explains the Wall Street Journal, “is an assurance that the statement that preceded or followed it is indeed factual.” Fire means good. Low-key gonna tell u my reaction to this paragraph rn: No cap, it is not fire. But de gustibus non est disputandum and I am pretty old. Perhaps there are millions of people who would appreciate a robot low-key telling them that their portolios are fire, if that is indeed factual. I suppose that is Arta’s bet:

Artificial intelligence is coming to the world of investment advice, and it can speak in Gen Z slang. 

That is the pitch from Arta Finance, a wealth-management startup led by an ex-Google executive and backed by the former chief executive of Swiss-banking stalwart UBS. Arta is rolling out an AI assistant that can dispense financial advice in spoken conversations—and in any preferred tone and argot. Even for the 20-something millionaire set.  …

The AI tool won’t recommend any investments that don’t match customers’ stated appetite for taking risks. And it definitely won’t trade on its own without the users’ consent—it isn’t that kind of artificial intelligence. But it can walk through the pros and cons of specific stocks, point out cost-saving tax strategies and offer advice on how someone might tweak their investment strategy if they take a pay cut.

I am an absolute sucker for stories like this, stories of applying artificial intelligence to all of the work of financial advice that isn’t picking stocks. My view is that there are two approaches to AI in finance. One is the approach of quantitative hedge funds: You build deep learning models to analyze financial data and try to predict which stocks will go up, so you can buy them. The other is the large language model approach: You ask a chatbot what stocks to buy, and it ingests financial language and comes up with plausible language in response to your question. Intermediating stock picks through a language model seems weirdly indirect, but it has the advantage of explaining its reasoning and perhaps importing “tacit domain knowledge” from human-written financial research.

It also has the advantage of chatting with you. Recommending investments that will beat the market is a quite small part of the business of financial advising. (How could it be otherwise? Beating the market is hard.) A larger part is, you know, picking up the phone when you call, and giving you advice about taxes and charitable giving, and, when the market is down and you are panicking, reassuring you that volatility is a normal part of life and that you shouldn’t dump everything now. “No cap, ur portfolio is fire” can be a useful thing for a financial adviser to tell you, whether or not it is indeed factual. We talked last year about a retail financial advising chatbot named Bridget, and I wrote:

When you call your actual human broker for advice about what stocks to buy, you don’t expect her to tell you the stocks that will go up. You expect her to share research and earnings call transcripts with you, build themed portfolios, that sort of thing. Maybe she’ll give you a tip about a stock she likes, but you understand that it won’t necessarily work out. You understand that she’s a flawed human being answering the phone at your brokerage, and that if she could reliably pick the stocks that go up, she’d be doing something else.

That’s probably true of the chatbot too: If Bridget could reliably pick the stocks that go up, she’d be working at a hedge fund.

Similarly, an obviously viable use case for AI in finance is something like “translating technical information from domain experts into plain English for normal customers,” and it is only a slight further step to “translating technical information from domain experts into ornate slangy English for customers with weird senses of humor.”

I suppose you can combine the two approaches. Build a model that is really good at picking stocks and also chats with you in normal, or other, language. Here’s Arta’s website, which suggests it might do that:

It takes more than just a ChatGPT wrapper to make AI work for wealth. Arta combines the unique strengths of various AI methodologies - from agentic systems to function calling and RAG to non-generative AI approaches like machine learning and convex optimization. Arta AI has the ability to bring in the specific models that are appropriate for the task.

Unfortunately now I am imagining a quant analyst at Citadel building a machine learning model that (1) is very good at predicting which stocks will go up but (2) outputs its recommendations in Gen Z slang. “Your GameStop short is low-key sus, bestie, no cap, that stock is fire,” it tells Ken Griffin, and he closes the chat window and says “absolutely not.” 

Also this is only slightly related, but on the theme of “financial advice for young people,” last Thursday I wrote about Robinhood’s possible transition from “gambling app for 20somethings” to “retirement planning provider for 40somethings,” and I asked:

What about their kids? If you are graduating college today, do you think Robinhood is for old fogeys? Where do you put your first $100? What is the platform that you use for fun now, that will grow up with you, that will surprisingly be offering retirement planning in 10 years? Is it FanDuel? Is it Pump.fun? Am I a huge boomer for even suggesting those names? Is it something I’ve never heard of? Almost certainly.

A reader pointed out that the answers are probably Venmo and Cash App. 

One way to distinguish “private credit” from the thing that is not private credit — not “public credit,” really, but the broadly syndicated loan market — is that private credit lenders make relationship-based loans that they plan to hold to maturity, while syndicated loans trade in a pretty active market. This distinction has always seemed pretty fragile. If big financial firms are accumulating billions of dollars of financial assets, they’re just going to trade, come on. I wrote in October 2023 that “sometime within the next, like, three years, I am going to be writing about a story like ‘Company X is launching a marketplace to allow private credit lenders to trade loans.’” In the intervening year and a half I have written several times about companies working on those marketplaces. 

In credit, though, before you get all-to-all electronic trading platforms, you get bank traders’ pricing runs. Here’s a PitchBook story about private credit runs:

A list of bids and offers on private credit loans by a major Wall Street bank might suggest an evolutionary moment for the private credit market, long promoted by lenders as a buy-and-hold asset class, where loans do not trade.

While the existence of potential trading of some private credit unitranche loans is a far cry from the deep aftermarket liquidity found in the broadly syndicated loan market, a list sent out by JPMorgan and reviewed by LCD showed 38 private credit investments, with each loan including a bid and an offer. The table, entitled “JPM Secondary Private Credit,” was dated as of March 31, 2025. …

Banks and private credit firms other than JPMorgan send out similar lists of credits they’re willing to buy or sell, according to market sources. However, sources noted that some secondary trades are facilitated on a bespoke and exclusive basis, with GPs selling out parcels of loans they agented or buying out other lenders’ holdings in credits they want to consolidate their positions in.

Prices include Finastra at 99/100.5, Galway 98/100, Medallia 94/97, CoreWeave 102/107, “perhaps reflecting repayment expectations following the company’s recent IPO.” A bid/ask spread of 5 percentage points does seem kind of wide compared to public credit, but pretty tight compared to “hold to maturity.”

A couple of weeks ago I wrote a column with the title “You Need Regulators to Deregulate.” The idea was:

1. Before President Donald Trump took office, there were various federal rules, statutes and regulations. (I mentioned in particular the Foreign Corrupt Practices Act, a longstanding statute banning bribery, and also all of the rules of the US Consumer Financial Protection Bureau.)
2. Trump’s administration doesn’t like many of those rules and would like to get rid of them.
3. There are normal ways to get rid of rules: You ask Congress to repeal statutes, you have administrative agencies rescind rules or issue new ones, etc.
4. For various reasons — Congress wouldn’t repeal the bribery rules; the CFPB sent everyone home — Trump’s administration doesn’t want to get rid of the rules in the normal procedural ways.
5. Instead it has announced, or in some cases just hinted, that those rules won’t be enforced anymore.
6. They’re still valid and binding, though. If you violate those rules someone might sue you, or some future federal government (or this one for that matter) might bring a case against you. 
7. This is not a satisfactory situation to a lot of the more rule-abiding regulated entities — banks, for instance — who have asked the Trump administration to reopen the CFPB (which they dislike) so it can get to work actually repealing its rules.

The US Securities and Exchange Commission is in a slightly different situation. Roughly speaking:

1. In the Biden administration, the SEC issued a bunch of aggressive new rules that the financial industry hates.
2. The industry sued the SEC, saying that all of those rules are not valid.
3. In many cases courts were sympathetic to these objections, in some cases ruling against the SEC, which then appealed.
4. Then Trump was elected.

Now what? One possibility is that the SEC can just tell the courts “never mind” and the rules are gone. Another possibility is that those rules are on the books, the cases or appeals are still pending, and the only ways to officially get rid of the rules are (1) to follow the SEC’s own rulemaking process and repeal them or (2) to lose a final judgment in court.

In 2022, the SEC issued a 510-page proposal to require US public companies to make climate-related disclosures. We talked about it at the time; it was hugely controversial. It went through a long notice-and-comment rulemaking process, and last year the SEC issued a final rule. There were lawsuits, and appeals were consolidated at one appeals court. Last Thursday, the SEC “voted to end its defense of the rules requiring disclosure of climate-related risks and greenhouse gas emissions”:

SEC staff sent a letter to the court stating that the Commission withdraws its defense of the rules and that Commission counsel are no longer authorized to advance the arguments in the brief the Commission had filed. 

So … you probably don’t have to follow the rules? (Not legal advice.) But they are still technically the rules, though the SEC has stayed their effectiveness pending the court review. The court can still make a decision. If it decides “actually these rules are totally valid,” then it can just say that? I think? Even if the SEC is no longer authorized to argue that? I don’t know? SEC Commissioner Caroline Crenshaw objected:

The Administrative Procedure Act (APA) governs the process by which we make rules. The APA prescribes a careful, considered framework that applies both to the promulgation of new rules and the rescission of existing ones. There are no backdoors or shortcuts. But that is exactly what the Commission attempts today.

By its letter, we are apparently letting the Climate-Related Disclosures Rule stand but are withdrawing from its defense in court. This leaves other parties, including the court, in a strange and perhaps untenable situation. In effect, the majority of the Commission is crossing their fingers and rooting for the demise of this rule, while they eat popcorn on the sidelines. The court should not take the bait.

Rather, the SEC should do its job. It should defend its existing rule in litigation. If the agency chooses not to defend that rule, then it should ask the court to stay the litigation while the agency comes up with a rule that it is prepared to defend (be it by rescission or otherwise, but certainly in accordance with APA mandates). At the very least, if the court continues without the Commission’s participation, it should appoint counsel to do what the agency will not – vigorously advocate in the litigation on behalf of investors, issuers and the markets.

The Commission’s actions are inconsistent with the APA, historical practice, and they embody bad governance. We do not have license to wholesale abandon agency action simply because the now-constituted Commission would not have supported the rule when it passed. The new majority cannot now rewrite history to change the outcome of a properly held Commission vote.

I think that’s right? I mean, even six months ago, I would have bet that courts would eventually strike down the climate rules; I never really thought of these rules as binding rules. They were a sort of expressive statement by the Gary Gensler SEC that the current US federal courts would never allow to go into effect. Still they are, still, technically, legally enacted federal rules. The SEC can follow the usual legal processes to revoke them, but revoking rules is almost as hard and legalistic as enacting them. Or it can be like “ehhh the court is going to strike them down, and we don’t like them anymore anyway, so let’s not spend any more time on this,” which is probably a rational choice but not exactly a legal one.

Elsewhere in the legal realism of financial regulation:

Regulators of major US banks including JPMorgan Chase & Co., Morgan Stanley and Citigroup Inc. may be forced to fight for desks as the Office of the Comptroller of the Currency pulls staff back to a Times Square skyscraper — without enough space for them.

“Due to a workspace shortage at your location, you will not participate in a seat selection process to have an assigned workspace,” according to an email sent Friday to unionized staff and seen by Bloomberg News.

The “hoteling” system used to book workspace will be suspended “for an indefinite period of time,” staffers were told, and until additional space is found, places to work “will be available on a daily first-come, first-serve basis.”

Everyone’s assumption seems to be that the goal of government return-to-office mandates is to encourage voluntary attrition, and I suppose that having to fight for a desk would make me want to quit. We talked last month about an English court ruling finding that giving an employee a bad desk constitutes “unfair constructive dismissal,” but I don’t think US civil service law works that way.

Wall Street Trading Desks Warn S&P 500 Selloff Will Get Worse. Tesla Sales Slump to Lowest Since 2022 Amid Anti-Musk Backlash. Andreessen Horowitz in talks to help buy out TikTok’s Chinese owners. China Restricts Companies From Investing in US as Tensions Rise. “Finance industry “slippage” on 1.5C ‘will make a material difference’ to how hot and unlivable the planet gets.” Deutsche Bank’s asset manager fined €25mn over greenwashing scandal. A Big Coal Plant Was Just Imploded to Make Way for an AI Data Center. Goldman Is Opening Up Private Equity Deals to Rich Individuals. The Dark Money Behind Duke Basketball. “Javice’s attorneys had said she shouldn’t wear an ankle bracelet because her main source of income is teaching Pilates in South Florida.” “We investigated the potential impact of trade tariffs on the sales of European supercars, and in particular, Lamborghini.” “It would be the equivalent of buying SeaWorld and complaining it’s too wet.”

If you'd like to get Money Stuff in handy email form, right in your inbox, please subscribe at this link. Or you can subscribe to Money Stuff and other great Bloomberg newsletters here. Thanks!