For example, the critique that decentralized networks should be responsible for preventing illicit activity because they “... operate like eBay” is incorrect. eBay is a corporation. Platforms like Ethereum or Uniswap are not. They are code written by one group of people that is run by another, more diverse group. Holding these disparate parties responsible for stopping illicit flows is akin to holding Microsoft responsible for criminal uses of Excel. Not only would this not work, it would erode liberties and raise First Amendment issues.

The Op-Ed is also factually incorrect about stablecoin impact on banks (“..Letting exchanges provide pecuniary incentives to stablecoin users could draw deposits out of small banks and thus reduce lending..). This is a false Big Bank lobby talking point. The academic research doesn’t support it. Community bank depositors skew older and often have social ties to their bank. Does the Journal really fear a 65-year-old farmer from the Midwest converting his savings to digital dollars at a crypto exchange just to get discounted trading fees?

The Clarity Act embraces innovation while protecting Americans. That’s why certain incumbents oppose it. It’s also why the rest of us should support it.

Also if you didn’t know better, you’d think the fact that Wells Fargo is now joining the likes of J.P. Morgan and Citibank to offer “an internal solution” that makes “payments and transfers faster and more efficient” only further proves that it’s a great idea.

Also, if you didn’t know better, you’d find this development an amazing twist of irony: the technology originally invented to disintermediate organizations like the big money center banks, either via cryptocurrencies or stablecoins, both of which live on actual public networks like Ethereum, and one of which the banks have been fighting a scorched earth battle in DC to curtail, that same technology will now be used by banks to do better banking!

Call it the Deus Ex Machina of blockchain, yet another instance in the long history of technological disruption where stale incumbents with hefty profit margins were able to use the technology invented to disrupt them to entrench their position even further. Similar to when local newspapers used the web to improve classified ads, forestalling Craigslist, or when Blockbuster video used the internet to improve the rental store experience, making it better than streaming.

Oops.

I sit in the uncomfortable position of knowing better. Partly because I’ve spent over a decade studying (and teaching) where blockchain makes sense and where it doesn’t, and partly because I worked on some of these exact projects (in one of these exact banks) back when I too didn’t know better. And I’m here to tell you that headlines like the one about “tokenized deposits” at Wells Fargo are at best empty, and at worst dangerous.

But I don’t expect you to take my word for it, or to rely on historical analogies. I expect you to use your common sense.

Banks, as we all know, have a real-time payment problem. They struggle to allow their clients to send payments quickly, especially on nights and weekends, even when the payment is to a different client of the same bank, or more embarrassingly, to a different account held by the same client at the same bank, just in a different jurisdiction.

Blockchain, as the proponents often declare, is a better way to move money. That’s because blockchains empower something called a token, and tokens have magical properties, like real-time transfers, programmability, and atomic swaps. Blockchains already do this in decentralized fashion. You can use a network like Ethereum or Solana to send dollars to anyone anywhere, or swap it for a different currency.

But decentralization has drawbacks, as the bankers like to point out. They include limited throughput and illicit use. Ergo, the best solution is for a bank to deploy an internal blockchain that it fully controls, tokenize customer deposits on it, and unleash the magic. This is the story that everyone from Jamie Dimon to the Suit Simps now shill on a daily basis.

There are two problems with this narrative. The first is that there is nothing magical about a blockchain token. It’s a very specific kind of database entry with unique properties, some desirable, others not. The second is that Wells Fargo (and JPM and Citi) could have rolled out the same exact internal real-time payment solution 20 years ago, using Excel.

Don’t believe me? Recall that a bank is nothing more than a series of balance sheets. It has assets (central bank reserves, loans, etc) and liabilities (customer deposits, corporate debt, etc). Each balance sheet is stored on a ledger. That ledger used to be a giant notebook but is now an electronic database—and has been for decades. Payments are just debits and credits on said ledger. If we have the same bank, and I pay you $100, then all it has to do is debit my balance by $100 and credit yours. Voilà! You’ve been paid. And you know what’s a really good technology for storing rows of data and making changes across it? A spreadsheet.

To be fair, GSIBs like Wells Fargo are a lot more complicated than that. They have different balance sheets in different countries, each with its own liquidity, capital, and regulatory constraints. They also have a hodgepodge of internal bookkeeping systems—the result of countless mergers and spinoffs—that don’t talk to each other very well. Reconciliation is a nightmare: different banking hours across jurisdictions, different programming languages, different data formats, etc.

If you didn’t know any better, you’d assume blockchain is a proper fix. After all, Ethereum runs around the clock and handles payments across the world as swiftly as ones across the table.

So why not deploy it inside a bank, as JPM apparently has, with its tokenized deposit solution that has an unpronounceable name? It’s literally a private & permissioned version of Ethereum running inside a bank!

In practice, this is a terrible idea. By permissioning the validator set to a few nodes operated by one company on a single IT stack, the bank is eliminating all of the features that make blockchain useful. Kinexxys does not offer immutability, guaranteed settlement, or censorship resistance. If you want a more detailed explanation as to why, you can read this. But for now, try answering a few simple questions:

1. What happens if North Korea starts using this solution? Will Chase let it, because blockchains are “censorship resistant”?

2. What happens if a bug results in a $10 transaction being accidentally processed as a $10 billion one? Will Chase let the error stand, because blockchains are “immutable”?

3. What happens if a hacker uses advanced AI to penetrate this system and drain ordinary people’s bank accounts to pay itself? Will Chase be OK with this exploit in the way Ethereum and Solana were not that long ago?

Obviously not. Among other reasons, their regulators would have a fit. One could argue Chase has a legal obligation to override Kinexxxys anytime something goes wrong. But then why bother with a fake blockchain in the first place? Why deal with the added complexity of consensus and cryptography if the bank doesn’t want the features they were combined to enable? Why not just use a modern database?

The permissionless validator set of any blockchain is similar to the cement mix in concrete. Stripping it out doesn’t lead to a different, but equally useful, building material. It leads to wet sand. Being permissionless is why blockchains have to use cryptography and consensus.

Still not convinced? Then riddle me this: Payment instructions are just bits of data: The world runs on data. How is it that Gmail and WhatsApp can allow billions of users all over the world to send each other 10mb video files in near-real time, and have been doing so for years, all using vanilla database tech, but Citi needs a fake blockchain to transmit a few kilobytes from one of its internal systems to another?

Or why is it that some banks (like Silvergate) were able to offer 24/7 payments many years ago without a fake blockchain, but Chase couldn’t?

Or how is it that India—a country not exactly known for its infrastructure—was able to roll out UPI for over a billion people, letting them make real-time payments across thousands of banks and FinTechs, and Brazil did the same thing with Pix, but Wells Fargo can’t offer the same feature to a tiny subset of just its customers until Q3?

These questions practically answer themselves. Banks don’t need blockchain to do better banking, at all. All of the features their fake blockchains hope to offer could have been offered long ago using existing database technology. But the banks chose not to. First because doing so would have been expensive in the short run, and second because doing so would diminished revenues in the long run. In banking, payment delays generate profits.

That’s why these banks fear yield-bearing stablecoins so much: a payment instrument that’s more useful than a checking account but pays more interest than a savings account is better than anything they can ever offer—even on a proprietary blockchain. Some bank executives have the audacity to claim tokenized deposits on fake blockchains are superior to stablecoins on real ones because they can pay interest. But this begs a simple question: why doesn’t the same bank pay interest on non-tokenized deposits?

Technology has nothing to do it.

Everything the largest banks plan to do on “DLT” could have been done using simpler solutions long ago. The question you should be asking is why did they wait until now to do it, and why must they do it using novel tech with many downsides?

Even in an enterprise setting, blockchains are slow and complex. They lack the tooling, support documents, and employee know-how modern database tech offers. They are also fragmented, particularly in an enterprise setting (there is only one Ethereum, but by my count at least 4 different ways to deploy a private instance).

Given the size and importance of banks like Chase, Wells Fargo and Citi, the smallest mistake could easily be catastrophic, as we’ve seen before. Regulators have strong opinions about what kind of core systems banks can use, and for good reason. Alas, today’s regulators also lack the sophistication needed to supervise so-called enterprise blockchains.

From a risk-reward point of view, banks deploying a stripped down version of novel tech, one that offers none of the benefits but all of the downsides, is a dangerous strategy. I would strongly advise against it.

• Winston Churchill (sort of)

I spend a lot of my time disagreeing with other people in the crypto industry, some of whom are friends. Our main disagreement is over just how much decentralization matters in core protocol design. They think it’s just one of several important features, I think it’s the only one that really matters. They think scaling is more important, I think it’s tangential. They think business development and partnerships are required for success, I think not. They think financial resources help protocols succeed, I think too much money guarantees they fail. They think permissioned networks can work, I chuckle.

Most of all, they think my views are too idealistic, and by extension, unrealistic. This is actually our greatest disagreement. My perspective isn’t that of a naif wishing on a kumbaya future.

I am a cynic who has spent a lot of time studying the past and how human systems evolve. I’ve also seen first hand the lengths to which powerful organizations go to protect their power and profit. My views are closer to Machiavellian. If you understand how the real world works, then it is the people who fall for vapid press releases about tokenization on corporate databases who are the idealists.

To believe what they do, one must also believe for-profit companies care more about innovation than their own bottom line. Or that the innovator’s dilemma doesn’t apply to platform technologies. Or that successful executives making seven-figure salaries based on their knowledge of the status quo are eager to see it change. I don’t. I believe in the laws of corporate inertia, and that only the most decentralized crypto systems are capable of reaching escape velocity. Everything else will be coopted and corrupted to the point of uselessness.

To believe in crypto is to believe in the power of incentives. Any blockchain that attracts millions of users and settles trillions in value will forever incentivize its own corruption. It would be foolish for the largest companies (or even governments) to not try to hijack it. Ignoring it might prove existential for some of them. That’s why they declared Bitcoin a scam ten years ago, or why they are telling you tokenization only makes sense on their terms today. This too is Machiavellian. First, try to stop it. If that doesn’t work, then coopt it. The only crypto systems that have a chance of surviving this threat are the ones that go out of their way to remain open and neutral from day one.

We often think about protocol security through the lens of external attacks, like a 51% re-org. We should be just as concerned about an internal takeover, perhaps even more so now that the oldest protocols are fairly hardened. Internal takeover is the history of almost every major TradFi exchange, settlement system, or even social media platform in operation today. It’s how Visa and Mastercard went from replicas of today’s non-profit tokenization consortia networks to money-making machines. It’s also how Google went from arguing against advertising as a business model for search to becoming the greatest advertising company ever. It is the arc of Enshittification, the natural consequence of the S curve that prominent VCs used to believe in.

For a layer-1 blockchain, the risk of takeover is even greater than any card network, clearinghouse, or social media platform. That’s because the TAM of a programmable and omni-asset settlement system is greater than most incumbent networks combined. A general purpose L1 can do payments, securities settlement, social, gaming, art, tickets, identity, and more. There is so much to enshittify.

Viewed through this lens, it’s the people who believe in permissioned networks—databases that can be unraveled by the push of a button—who are naive. Same goes for any supposedly permissionless Layer-1 with a concentrated validation set, or allegedly public Layer-2 with no proofs and a single sequencer. To believe in such systems is to believe in individuals as incorruptible, institutions as never evil, and governments as always restrained.

More tangibly, it’s to believe that Visa wants Mastercard to succeed.

Today, the takeover scenarios I speak of are not hypothetical. Take for example the leading purveyor of databases controlled by a button, a company whose CEO is on a proud mission to make incumbents and intermediaries great again. In a recent interview, he eloquently explained his belief that closed enterprise networks running Proof of Authority are fairer to open ones running Proof of Stake. His logic? Joining Ethereum consensus costs money (~$60k at today’s prices) but joining his network only requires would-be participants to “demonstrate value” to existing ones.

Well, it so happens that Visa is already a participant in this network and Mastercard is not. How does a company demonstrate value to its largest competitor? Or better yet, what if Visa conspires with Mastercard to both join this network but never let another competitor in, forever enshrining their duopoly status at the top of Western payments? How would a FinTech hoping to totally transform payments “demonstrate value” to this trillion-dollar behemoth?

By asking nicely?

If you think I’m being too harsh, then you haven’t studied the history of payment and clearing systems. But don’t take my word for it, just ask the smaller banks and credit unions in America how they feel about The Clearing House. Or any bank that isn’t a shareholder of EWS how it feels about Zelle. Or Robinhood about the NSCC, particularly during the memestock boom. Or Custodia about the Fed, or FinTechs about FedNow.

Then put yourself in the shoes of the CEO of a profitable payment firm with high take rates and gross margins. You got to where you are because you understand the importance of owning the network—it’s practically tattooed on your brain. Before crypto, all settlement systems were either run by incumbents or by governments (who are influenced by incumbents). Now there’s something new called a public permissionless blockchain, and some very smart people tell you it’s a settlement system that nobody controls but anyone can use. That includes your biggest competitor, or any startup that views your profit margin as their opportunity.

Pop quiz hot shot, what do you do? Do you embrace it?

Or do you look for some kind of hybrid alternative, one that supposedly provides some of the benefits of a blockchain but lets you maintain power and pricing? And maybe afterward instruct your PR people to come up with a compelling story about regulation and responsibility and illicit use?

The question answers itself. Viewed through this lens, the takeover scenarios I present aren’t all that Machiavellian. They are BAU. Competitive companies take every edge they can find. Owning, or at least controlling, the means of settlement is the ultimate edge. Of course they’ll try to take over any network that lets them, then use false attacks and lawfare to sabotage the networks that don’t.

None of this is going to work in the long run, mind you. Not because corporations are bad at playing this game, but because false-decentralization is objectively inferior to the status quo. It’s neither as efficient as what TradFi runs today nor as secure as real decentralization. On enterprise networks, the cryptography is bloat and the consensus a charade. Fake decentralization is only effective in VC pitches and on conference panels. It will fail in the real world.

My Machiavellian lens makes me wonder whether the banks and brokers playing this game already know this. If they do, then their embrace of fake crypto is a clever sleight of hand, designed to slow progress and sway lawmakers. The strategy is understandable on a human level. All of these companies are run by older people closer to the end of their careers than the beginning. They have reputations to maintain and Hamptons lifestyles to support.

Their delay tactics will only work for so long. The world will eventually find its way to the most decentralized systems, in the same way that water eventually flows to the lowest elevation. The centralized world’s profit margins, much of which derive from the delays and frictions of the old ways, are somebody else’s opportunity. This too is Machiavellian. A fully decentralized settlement system is a powerful weapon to wield against your competition, especially if you don’t have the tech & business model baggage that they do. Throw in fading trust in existing institutions, and the process accelerates.

Water eventually finds its way to the lowest point, assets eventually find their way to the safest infrastructure. That’s the Nash equilibrium of the world that we live in. Best to be a realist, as I am. A decentralized system like Ethereum has many flaws—fending off capture is expensive and complicated. But it’s still better than the enterprise and corporate alternatives being bandied about today. A lot of idealists are going to learn this lesson the hard way.

With the PBOC out, the ECB now holds pole position as the last remaining major central bank that takes this topic seriously enough to try something. But the ECB has also painted itself into an impossible corner due to its obsession with protecting incumbent banks, so that something is unlikely to get very far or be adopted.

Part of me is disappointed. Central bank digital currencies were an awesome subject to study and teach, regardless of one’s position on them. Few ideas evoke thornier questions of technology, economics, and politics. Even if dead, their story arc still makes for an excellent teaching moment. From their emergence out of nowhere, to their Utopian (or Orwellian) promise, to their apparent inevitability, to their quiet demise. There is a lot this story can teach us about the current state of money and power in different countries.

I have a theory as to why they never took off, but need to explain the history first.

The Reaction

The vague notion of a digital currency issued by a central bank has been around for decades, arguably since the birth of digital. It was energized by the emergence of Bitcoin, but not enough for any central bank to take action. Back then most would have told you that their currencies were already digital, on account of their electronic payment systems.

That all changed thanks to the one-two punch of Facebook announcing its ill-fated stablecoin project called Libra and the pandemic exposing the limitations of existing payment solutions. Seemingly overnight, everyone was obsessed with the urgent need to issue a CBDC, either to head off the rise of corporate currencies, or to make it easier to distribute stimulus.

Importantly, both of these motivations were muddled with confusion. Libra was just a stablecoin. Today, we see those as complementary to fiat money, not a substitute. We also understand that emergency aid could be distributed via other types of modern payments, like FinTech apps, prepaid cards, and of course stablecoins. But Facebook was a deeply unpopular company back then, not everyone trusted FinTechs, and nobody understood blockchain.

Despite this confusion, or maybe because of it, CBDCs jumped to the forefront of policy debates. Different countries had different motivations for pursuing one, but fear drove the agenda everywhere. First and foremost was a global fear of Facebook taking over payments (except in China, where Alibaba and Tencent already had). Beyond that, the UK feared fragmented digital payments breaking the singleness of money. The EU feared the ever-growing dominance of American card networks. Multiple EM nations feared capital flight via some blockchain-based alternative, be it a stablecoin or Bitcoin.

In the US, there was growing concern that China might use a digital currency to take on the dollar. Conservatives cared about this marginally more than liberals, but there was xenophobia on both sides. So congressional hearings were held on the need for America to issue a Digital Dollar, a non-profit was founded by former government officials, and even our new Fed Chair weighed in.

China has always been a pioneer on this topic. It started looking into a CBDC as far back as 2014. But I always thought the fear that a Digital Yuan would somehow threaten the dollar was misplaced—politics and economics matter a lot more in currency competition than form factor, and there are good reasons why the Yuan is rarely used as a reserve currency. But this was a fearful time in the West, technologically and geopolitically, so it was tempting for otherwise intelligent people to think a Digital Dollar would be the proper response.

By the time we got to the latter days of the pandemic, seemingly every country was looking into a CBDC. The topic dominated academic conferences and the agenda of global organizations like the BIS. The momentum was further fueled by yet another crypto bubble and bust, and lingering concerns about private-sector solutions like Tether. Also in the background was Russia’s invasion of Ukraine and the resulting sanctions. The world of money seemed ripe for disruption, and governments didn’t want to be left behind.

And yet, nothing actually happened outside of China. Issuing a general purpose CBDC is one of the most transformative things any government could do, financially and economically. There are many technical and operational problems that need to be solved, and thorny questions to which there are no easy answers. It’s all unprecedented and there’s a lot that could go wrong. That’s not the type of thing governments usually sprint to. Looking back, it is remarkable how eager so many were to even try. Entrepreneurs can move fast and break things, technocrats cannot.

Ironically, even though all of this hype failed to lead to the creation of a CBDC in most places, it did inspire a wicked backlash against the very idea of one, particularly in America.

The Counterreaction

The easiest way of understanding a CBDC is as a form of digital cash. Technically speaking, it’s a direct liability of the central bank. Practically, you can think of it as a dollar bill that exists on a database, as opposed to on paper. Cash is seldom controversial. If anything, it’s the preferred way to transact by people who mistrust corporations and governments. But digitization flips this script. Whereas physical cash is private, a CBDC could be surveilled. Whereas cash is universally accessible, digital cash can be censored. A government who issues it could exercise novel control over users.

The mere possibility of these features tends to set off alarm bells. A CBDC could be designed to be private and censorship-resistant, but governments seldom prioritize such features, and ordinary people often fear new technologies being used to take away their freedoms. Viewed through that context, it’s not surprising that the sudden urge to embrace CBDCs led to an equally powerful urge to oppose one.

America never seriously considered issuing a CBDC. The Fed never got past the technical research stage (and only at a regional bank, not the board) and the rest of the government showed little interest. You could almost see former Fed chair Jerome Powell fight the urge to roll his eyes whenever he was asked about it. But that didn’t stop some Americans from freaking out over a hypothetical one. The revolt began during the Biden administration, when a nominee for comptroller of the currency was shelved due to her vaguely favorable views on this topic. It continued in the second Trump administration with anti-CBDC executive orders and state laws. Even now, with the subject clearly in the rearview mirror, there is a bipartisan housing bill awaiting Trump’s signature that flat out bans a central bank digital currency. If signed, an abstract product that never was will never be.

As someone who has spent years trying to wrap my arms around this complicated topic, I’ve been genuinely impressed with the grassroots opposition to it. It didn’t occur to me just how polarizing CBDCs were until a couple of years ago when I heard an ad on my local sports talk radio station warning people about the dangers of one (and encouraging them to buy gold). Part of me gets this fear—in my last book I described a general-purpose CBDC as a “monetary Death Star.” But another part just chuckles. The Fed isn’t qualified to build a CBDC, even if it wanted to (and Congress allowed it to). Heck, it took years to build FedNow, and that’s just a simple payment system, one that hasn’t gotten much adoption. A CBDC is a lot more complicated and a helluva lot more impactful.

Complexity is one reason this topic is now mostly dead. All of the public policy decisions that most government have made towards their private financial systems is another.

The public-private distinction that isn’t

Unlike America, China was very serious about rolling out a CBDC. It also put real resources behind getting the e-CNY adopted, even borrowing a page from the crypto industry and giving away free money. But the whole thing landed with a thud, garnering little organic adoption. One possible explanation is that Chinese payments are already digital thanks to widely-popular chat-based solutions like Alipay and WeChat Pay. Taking on these products with a government-issued alternative was always a tall order. Students of mine who tried the e-CNY reported a clunkier product with less features and more censorship. Most of them didn’t see the point.

From a purely academic point of view, the primary reason for ordinary people to switch from a private-sector payment solution to the e-CNY is safety: Alibaba and Tencent could both fail. They could commit fraud or make mistakes, leading to unbacked user balances. Every concern that skeptics and regulators have had about stablecoins also apply to FinTechs. Stables and payment apps are private forms of money. The PBoC however issues a public form of money, and public money is always safer. Central banks can’t fail in the traditional sense.

This isn’t just my opinion. It’s basic monetary theory and one of the bedrock assumptions that fueled both the initial interest in CBDCs and the hesitation by even the most bullish central banks to consider a widely-accessible one that could be held in unlimited amounts. Most policymakers would consider such a product too safe and a competitive threat to bank deposits. Its very existence could accelerate bank runs as depositors would have a safe and convenient alternative to run to.

All of which begs a simple question: Why didn’t Chinese users—even the ones targeted with giveaway promotions—eagerly embrace the safer option once it was provided? Do they not know what’s good for them financially?

That would be an ego-satisfying conclusion for any academic to make. It would also be the wrong one. If you peel the financial onion one layer further, it becomes apparent that the safety distinction between a privately issued fintech balance and public e-CNY balance is not that stark. Not in the modern era anyway. Tencent and Alibaba were forced to move their reserves to the PBoC years ago. They were also barred from directly paying interest on user balances. This makes their user balances ultra safe. Indeed, in academic parlance this particular model of a private issuer backing user balances with central bank reserves is known as a “Synthetic CBDC”. It combines the safety of central bank money with a thin layer of private sector intermediation.

Put more bluntly, it’s impossible to imagine a situation where the Chinese government lets either company fail in a fashion that compromises user balances. That would be catastrophic. Modern governments love bailouts and the Chinese government has been rescuing its failing state-owned banks for a long time. Alipay and WeChat Pay are clearly too big to fail. That being the case, the rational decision for Chinese citizens is to stick to the private-sector solution that offers greater features and less direct surveillance.

Alas, if the biggest distinction between a CBDC and private sector alternative in any country is counterparty risk, but the central bank is willing to do “whatever it takes” to rescue banks and FinTechs should anything go wrong, then this is a distinction without a difference. The biggest differentiator between central bank money and commercial bank money is no longer relevant. That’s not just a Chinese phenomenon but a global one. In our current bailout-happy world, the liabilities of banks, FinTechs, and stablecoin issuers are effectively as safe as that of their respective central bank.

Like turtles, our existing financial system is already CBDCs all the way down. We just didn’t realize it.

Everyone is too big to fail

To me, this new reality was cemented in 2023 during the regional banking crisis. The decision to rescue all SVB depositors proved that we had crossed some kind of Rubicon. That the U.S. government wouldn’t even entertain the idea of a small haircut for VCs foolish enough to keep $20m in an uninsured bank account showed the appetite for any loss in a bank failure was now zero. This wasn’t just an American phenomenon. In Europe the Swiss government literally changed its laws so it could bail out Credit Suisse. And I understand why. Decades of government intervention into ever smaller financial crisis has made the entire system more fragile. Bailouts beget more bailouts.

Hell, there’s even an international organization that publishes an official list of Global Systemically Important Banks, firms deemed to be “too big to fail” before any crisis. Claude estimates they account for almost half of all global deposits, around $50 trillion worth. But don’t count out the firms just outside of this list as being too small to be rescued. What the folks who make such lists understand better than anyone is the risk of a small org failing cascading to the big boys. That was the justification for saving SVB and Signature Bank. It’s the justification that will be used to rescue ever smaller firms going forward.

Looking at the financial system through this lens is how I arrived at my prior argument that we already live in a world of implicit CBDCs. As I stated back in 2023, it’s worth asking whether we’d be better off with the explicit variety.

“Societies that accept greater government intervention in the private banking system have already embraced implicit CBDCs. Financially, they might be better off switching to the explicit variety. They would have less moral hazard while enjoying other benefits such as greater economic inclusion and less friction in payments.”

Looking back, my construction was only half right. There is indeed more moral hazard than ever. But you don’t need an official (explicit) CBDC to expand inclusion or improve payments. Stablecoins on decentralized blockchains also fulfill those functions. They’ll never be as anonymous or universally accessible as cash, but neither would most CBDCs. I also expect market forces to push the most successful stablecoins to be mostly hands off.

Which brings us back to Europe, where the ECB recently achieved a major legislative milestone for the digital euro. After years of political wrangling, the prospect of a CBDC became more likely after getting voted out of committee inside the European Parliament. I won’t pretend to understand EU politics, but friends who do tell me this is an important step. There is still more politicking to be done, and presumably a lot more infrastructure to be built, but if we are optimistic then Europe might have a digital euro sometime in the next five years.

This still doesn’t negate my argument that CBDCs, as we understood them, are effectively dead. The constraints the ECB has imposed on itself for the digital euro make it far from a monetary Death Star. For example, even though they call it a form of cash, it can only be acquired and held with the aid of a bank, FinTech, or some other kind of intermediary. There is no non-custodial ownership option—to use a crypto term—and it’s likely every unit will have to live inside a KYC-perimeter. This type of intermediation would be required to enforce the digital euros biggest restrictions: individual holders will have strict balance limits (likely €3000) and commercial users will have strict time limits (likely 24 hours). It’s also possible the intermediaries that furnish the digital euro will have to do constant AML checks.

I’m not sure what to call this design, but it’s a far cry from digital cash. Physical cash can be acquired from anywhere (not just a bank), held anonymously in unlimited amounts, and transacted privately without the involvement or even knowledge of an intermediary. Physical cash also circulates abroad as freely as it does at home. The digital euro will not offer these freedoms. On its website, the ECB explains why: to protect user privacy (but only from the central bank) and to protect bank deposits (from any competition). Alas, not only are these desires in conflict with each other—governments can easily spy on people by way of a private bank—together they are likely to impede adoption.

The ECB has stated all along that one of the reasons it wants to issue a CBDC is to diminish the footprint of foreign payment providers like Visa and Mastercard in Europe. But the restrictions their leaders are imposing on their product makes me think they will fall short of that goal and see the digital euro go the way of the e-CNY, landing with a thud.

Card swipe fees are capped in Europe, so cost can’t be the reason people and merchants switch to an alternative. All of the restrictions imposed on holding and using the digital euro will incentivize people not to. Like the e-CNY, it will be a clunkier alternative, running on government sponsored infrastructure. It would be one thing if the digital euro was perceived as being safer than a bank or FinTech (or stablecoin) alternative, but the European Central Bank has trained a generation of depositors not to think that way. It would be another thing if the digital euro paid interest, and paid more than the measly amounts Eurozone banks currently pay their depositors, but this will never happen. The ECB has made it clear all along that the digital euro will never pay interest.

This final restriction is rather ironic, and yet another reason the desire to protect banks from disruption has bastardized the CBDC design space. Today, the ECB pays interest on the excess reserves banks keep there. A digital euro and and euro bank reserves are identical products, financially speaking. Both are direct liabilities against the same balance sheet. But going forward, who holds that liability will determine whether it pays interest. If a bank holds it, then it shall. If you do, then it won’t.

Not only that, but so long as the European banks that benefit from this governmental largess pass on a fraction of that interest back to you by way of your bank deposit, then you have an incentive not to adopt the digital euro. An incentive that grows with the rate of inflation and prevailing interest rates. An incentive finance by the ECB.

Putting it all together: the European Central Banks wants to issue a digital euro, but one with more restrictions than the private-sector solutions it will compete with. The ECB will pay people not to use it, and keep bailing out the banks Europeans use to make payments today. All of which begs a simple question: why even bother? Why not just build a better bank-to-bank payment system and call it a day?

In Conclusion

And so, an exciting topic that came in like a lion must now go out like a lamb. Five years ago, many of us thought different societies would not tolerate a CBDC because it would be too radical from what we have today. The politicians ramming through anti-CBDC legislation in places like the U.S. are still focused on this concern. But today, something closer to the opposite has occurred. In any country where the government is focused on bailing out banks during the worst of times, and protecting them from competition during the best, a CBDC isn’t differentiated enough to be adopted.

A decade-plus in the crypto industry has taught me to never say never. We are living in an era of rapid change and various political and economic pendulums swinging violently back and forth, so it’s possible that things happen in the years to come that force at least some governments to reconsider. But a lot would have to happen for any government to issue a disruptive CBDC, the kind we contemplated and debated five years ago. And the one thing said government would have to overcome would be the obsession with protecting banks.

In the meantime, there will be discussions about “wholesale CBDCs” and other types of incremental improvements of our existing payment stack—all of which I support. There is no reason why central banks shouldn’t upgrade their existing payment systems to be 24/7 or accessible via APIs. America should definitely do this, though we’d have to call it something different so as not to trigger the fearful. It will be beneficial to our economy, regardless of what we call it. But it won’t be as disruptive as what we once imagined from a true-CBDC.

That opportunity now lies with stablecoins.

But there is something happening in the crypto industry today that makes me feel like it’s not accurate enough, something more troubling than people just being wrong. Ignorance is one thing, and you can argue some amount of it is healthy out on the technological frontier. What ails us now is a deeper malaise, one closer to cowardice. The problem today isn’t that people don’t know anything. It’s that in crypto, nobody believes in anything.

Sure, there are still Bitcoin maxis posting debt charts and waiting for the long-awaited collapse of the supercycle. But most of them sold out to Saylor, and their belief system is in crisis now that he might be a forced seller. If they actually believed in anything, it would be his buying they would have found problematic. I don’t know about you, but endless financial engineering by a has-been IT company isn’t what I find most interesting about censorship-resistant non-sovereign money. But it was for the maxis, because “number go up” is the only thing they actually believed in, and Strategy was good for that, at least for a while. Now the number is going down, and there’s a quantum threat to deal with. So what’s a disenchanted fanatic who bought too much at $100k to do? Beg the government to announce a Strategic Reserve, of course. Because that’s what Bitcoin really needs to succeed. The government.

This malaise isn’t specific to any coin or community. It’s everywhere. There are ETH Maxis who went all in on the lame Ultrasound Money narrative, only to see their favorite coin flail over the past five years. Now they don’t know what to do. Some are obsessing over small changes to the issuance curve, as if that’s the reason pension funds aren’t loading up on ETH. Others are going around trotting the oldest (and dumbest) critique against crypto: that the networks might be useful, but the coins aren’t valuable. Anyone who believes that doesn’t believe in economic security. They just believes in…databases.

The biggest tell that some sycophant doesn’t actually believe in anything is his tendency to go all in on whatever meta happens to be working today and backfilling the intellectual justification. You know the type, the guy who is all about Hyperliquid today and was all about Solana a year ago. In 2024 he tried to convince you that memecoins are the future. Now he’s obsessed with protocol revenues and privacy. “The institutions are coming for on-chain RWAs” he says, “but they can’t use a public blockchain because that would violate securities laws.” If you ask him which securities laws he’ll look at you dumbfounded for a while, then change the subject back to Hyperliquid.

To be clear, there is some truth to all of this. Hyperliquid is good at what it does and we do need better solutions for on-chain privacy, partly to cater to institutions. But Hyperliquid has its drawbacks, and the tension between transparency and privacy is a challenge for institutions, one that will take time to sort out. What bothers me about these fanboys, most of whom are on their third protocol marriage, is their inability to grapple with these nuances. It’s not that they changed their minds—we should all be doing that as we learn—it’s that they never learn from their mistakes. They claim some protocol is perfect, later declare it useless, then tell you some other project is perfect. Their belief system perennially consists of a price chart and the talking points spoonfed to them by an insider using them as exit liquidity. And we know how that ends. Just look at Zcash, another interesting project hijacked by salesmen and suckers.

This crisis of faith is apparent across the industry, but it’s most acute among the people currently obsessed with institutional adoption. You know this type too: The guy who hadn’t heard of the DTCC until six months ago but now retweets all of their vapid claims. Or the one who can’t stop talking about vaults. Or the ones who think Stellar is still a thing. For the record, I’m all for institutional adoption. I was writing about it long before most of these knuckleheads were in crypto. But I never thought it would be limited exclusively to legacy institutions and only on their terms. It takes a certain lack of imagination to think that’s how any disruptive technology plays out, never mind one invented to disintermediate incumbents.

Alas, there is now a strong contingent within the crypto industry who lives and dies by the throwaway quotes of mid-level corporate executives doing innovation theater. Last year I started referring to them as Suit Simps, mostly as a joke. But now I feel sorry for them. They remind me of the nerds back in high school who got overly excited anytime a pretty girl smiled at them. Just watch them the next time Jamie Dimon makes some fantastical (but rubbish) claim about their internal database with an unpronounceable name. They’ll start drooling.

It’s one thing for randos on Twitter or retail investors to behave this way. But the crisis of confidence now runs so deep that it has even infected some of our best thought leaders. In 2018, they were writing seminal blog posts on why decentralization matters. In 2026, they are throwing money at corporate databases run by card networks and trading firms. Then they turn off their brains and regurgitate the unverified claims of how much adoption these databases supposedly have. It takes the best founders in crypto—founders these same VCs swore by—years to bootstrap minimal adoption on real blockchains, but somehow magically an archaic TradFi firm that runs on COBOL has gone from zero to trillions in a year.

The claim is so preposterous that it’s an insult to our intelligence—said firm was still using fax machines until yesterday—but that doesn’t matter. Powerful financial firms, the ones who’ve been trying to kill crypto with lies and lobbying forever, have now declared permissioned databases the only acceptable path forward, and the brightest minds in crypto VC now outsource their thinking to them. But only for things like payments and capital markets, mind you! These same VCs still believe in Web3 and DeSo that empowers DeSci so we can sort out DePIN and build decentralized systems for AI inference. That kind of decentralization they still believe in. But disintermediating stock trading? That’s a bridge too far.

It’s tempting for me to classify these silly people as lacking integrity, but that would be unfair. Integrity can only flow from some kind of belief system, which I’m starting to realize many people in crypto never had. Some just jumped on the bandwagon when the going was good. Others sort of believed but are now buckling under the dual pressures of prices being down and attention being elsewhere. And some are just idiots.

I for one remain as confident as ever that we are in the early innings of a major transformation of multiple trust frameworks, starting with finance. I’ve been wrong about plenty, and have had to evolve my mental models repeatedly. There’s still a lot that I don’t know, and interestingly enough, the number of questions on that list has grown over the past year (I have yet to figure out why). But to me, that’s what it means to believe in something. You start with what you don’t know, then work your way backwards to what you are reasonably sure of. All the archetypes I mentioned above did the opposite. They were sure of everything, until things got hard, and now they believe in nothing. Either that or they were just chasing the easy money all along, much like how Wile E. Coyote chased the Road Runner, and have finally gone over the cliff.

Now is a good time to decide which camp you’d like to be in. Either believe in something or just move on.

Let’s start with the title, which accurately describes stablecoins—blockchain tokens that use a centuries-old structure to represent fiat currency—as a private form of money, but falsely concludes this is in and of itself makes them risky. Most dollars in circulation are private forms of money. They include bank accounts and money market funds. Per Federal Reserve data, aggregate money supply in the U.S. is over $22 trillion, of which only $5 trillion is the monetary base. Credit creates money, so wherever there is lending in the private sector, there is private money.

American consumers and firms conduct their most important economic activities almost entirely in private forms of money. They get paid via direct deposit to their bank account (~$19 trillion) and save in money market funds (~$8 trillion). If private forms of money are dangerous, why even waste the ink needed to criticize the $300 billion stablecoin market? It’s a drop in the supposedly dangerous bucket.

It would be one thing if stablecoins were unregulated and poised to stay that way. That would certainly be cause for concern. But the Genius Act passed a year ago, and multiple federal agencies have already released hundreds of pages in proposed rule-making. The pending Clarity act might regulate them further still. When all is said and done, stablecoins will arguably be the most regulated form of private money out there.

Mr Ip’s response? “... no legislation can fully remove risk that is intrinsic to the design of stablecoins.” This is an odd standard to hold stablecoins to since no legislation can fully remove the risk of any financial instrument. The endless laws and regulations passed after the 2008 financial crisis didn’t stop the failures of Silicon Valley Bank and First Republic, among the largest bank failures in U.S. history.

What makes stablecoins unique is the straitjacket regulators will now place on them. Under Genius, the issuers of payment stablecoins are narrow banks that can only invest their reserves in the safest investments. To his credit, Mr. Ip acknowledges as much in his column. But he doesn’t grapple with the way this basic fact refutes the rest of his argument.

Unlike traditional banks, stablecoin issuers can’t participate in fractional-reserve lending or maturity transformation. Unlike money market funds, their liabilities will be fully visible in real-time on public blockchains. This unprecedented transparency is important for safety and soundness. Why focus on the safer and more transparent innovation, one set to compete with older and clearly more dangerous forms of private money?

The rest of Ip’s column is a potpourri of non-sequiturs and false analogies. Stablecoin issuers are said to have an incentive to “reach for yield”, as if every other issuer of private money doesn’t. Once again, the important thing here is that stablecoin issuers won’t be able to reach as far as other intermediaries, like banks. New regulations are said to only apply to American firms, but that’s true for all regulations, including bank ones.

The concept of the singleness of money is said to be violated, but this too is more problematic for other deposit products, even FinTech accounts. The homogenous nature of stablecoin reserves (assets) naturally leads to more homogenous coins (liabilities). It is always telling when critics argue that stablecoins are risky because they don’t enjoy (costly) perks like access to lender-of-last-resort facilities or bailouts. If those services are beneficial, a rational observer would argue they should also be offered to stablecoin issuers, doing so will cost taxpayers far less than offering them to banks.

Ip quotes Fed Governor Michael Barr as stating the Genius Act has “loopholes” about reserve composition. First, we should note that Barr was the Vice Chair for Supervision under whose watch the regional banking crisis occurred. Second, and in his defense, Barr made that quote last year, when Genius had just passed but no rulemaking had occurred. All potential loopholes will be closed by the regulatory rulemaking process which involves the Federal Reserve, OCC, and FDIC. Does Greg Ip consider our banking regulators incompetent? If so he should really stop worrying about stablecoins and focus on banking.

Ip also shares a misleading stat from the Chainalysis 2026 Crypto Crime Report (” Stablecoins account for 84% of illicit crypto activity such as sanctions evasion and money laundering”) without mentioning the very next bullet point from the same report, which states illicit activity accounts for less than 1% of all crypto volume. This is like sharing a stat on the percentage of cars involved in a police chase and using that to argue against people driving.

The most disingenuous part of Greg Ip’s complaint is his comparison of stablecoins to free banking. This oft recycled concern is nonsensical. Free banking, as the name implies, was a form of unregulated banking, at least at the Federal level. The Genius Act introduces a new Federal licensing and regulatory regime for any stablecoin issuer with assets over $10 billion. Section 5 of the law states that issuers “...shall be licensed, regulated, examined, and supervised exclusively by the Comptroller.” That’s the Office of the Comptroller of the Currency, the very federal agency established by the National Bank Act of 1863 to end the free banking era.

This column isn’t the first time Greg Ip has waxed poetic rather than analytical on the dangers of a cryptocurrency. In 2021, he compared Bitcoin to Fentanyl, as if a digital currency that nobody has to use and a highly addictive substance that kills scores of Americans are equivalent. That argument didn’t age well now that so many of the world’s largest and most respected institutions have embraced Bitcoin, so Ip has found a new boogeyman. Too bad stablecoins are one of the safest and most useful financial innovations in a long time. You don’t need to know anything about crypto to see that. You just need to understand the basics of money and banking.

In crypto, as in life, price action dictates opinions, not the other way around. This isn’t the first bear market that’s hijacked people’s beliefs around what works, and it won’t be the last. So as the Suits look ascendant, and podcasters crash out, I thought it helpful to share my basic framework for what ends up capturing the value in different kinds of networks.

In a decentralized settlement system, the native coin captures substantial value

This is the closest thing we have to a tautology in our industry. The relationship between network effects and coin value has to be true for a decentralized settlement system to remain secure. The mechanics of Proof of Work and Proof of Stake break down otherwise. If you don’t believe in this relationship, you don’t really believe in crypto.

Note that I said substantial value, not all value, or even most. History will reveal whether some version of the Fat Protocol thesis is accurate. But it’s reasonable to believe that you can’t have trillions of dollars secured by a Layer-1 blockchain with only billions in economic security, there’d be too many attack vectors. At the same time, you don’t need a trillion dollars in economic security to protect a trillion dollars in assets, that would be overkill. Economic incentives are only one part of what makes a decentralized system secure.

For the first decade of crypto, there was a popular belief that blockchain was valuable, but Bitcoin wasn’t. That turned out backwards, as countless enterprise networks failed to get anywhere while Bitcoin became a trillion-dollar asset. In this second decade, it’s popular to argue that a smart contract network like Ethereum will be useful, but ETH the asset will not be. This too will be proven farcical.

It’s not that the asset issuers, dApps, or even Layer-2s that rely on Ethereum won’t capture value, they will. It’s just that Ethereum, as the settlement system, has a stronger moat than any of those. Bootstrapping a decentralized settlement system is hard, building on top of such a system is easier.

That which has the strongest moat is most likely to capture substantial value. Anyone who argues otherwise is overthinking it, and likely overinterpreting current price action.

In a centralized settlement system, the equity holders capture substantial value.

This category includes TradFi exchanges, card networks, clearinghouses, and so on. There is no debate as to who captures the value here, it is always the equity holders because they are the only game in town.

If the centralized system is run by the government, then value flows to the equity holders of the firms given exclusive access; governments don’t run open systems. Think: RTGS systems, ACH, RTP, etc.

If the centralized system is operated by the private sector, value can flow either to the equity holders of the system itself, or the equity holders of the firms given exclusive access to it.

The first category includes card networks and TradFi exchanges. The shareholders of Visa and the NASDAQ capture substantial value.

The second category includes clearinghouses and ancillary enablers of settlement activity where the network operator is structured more as a governance layer. The shareholders of the DTCC and SWIFT do not directly capture much value, but the shareholders of the brokers who can access DTCC’s services, and the shareholders of the banks that use SWIFT to coordinate payments, do. And guess what? The largest shareholders of both systems are said banks and brokers.

The “private, but not profit-making” nature of these systems is part accident of history, part political cover. DTCC’s dominance is enshrined in law, it’s a government-protected monopoly. It would be problematic if it also made a lot of money. Passing value to the next layer of shareholders is a clever way of protecting the existence of a highly profitable cabal.

Given their entrenched position and their protected profits, the biggest threat such cabals face is from decentralized systems that perform similar (or superior) services without the exclusive access. In TradFi, incumbent banks and brokers constantly machinate behind the scenes to keep startups out of the systems they control. They can’t do that in DeFi.

That’s why they’ve lobbied so hard to have the government deem decentralized systems out of bounds for regulated activity. That strategy is no longer working, so they’ve resorted to launching fake alternatives that they fully control.

In any permissioned database, the equity holders capture all the value.

So-called “enterprise or permissioned blockchains” are not a thing. They are just corporate-controlled databases that incorporate mostly useless cryptography to pretend to be of the crypto world. In reality they are no different from legacy TradFi systems; they offer none of the features of an actual blockchain.

Like legacy centralized settlement systems, the value capture from these networks goes to the equity holders. It remains to be seen whether this happens directly or indirectly. Some are structured as non-profits and run by a foundation, indicating indirect value capture. Others have raised hundreds of millions of dollars from VCs, indicating a more direct form of monetization.

Regardless, every permissioned database has a single button that controls who gets access to it, that’s why the cryptography is performative. And whoever controls that button dictates value capture. To argue otherwise is to defy the basic properties of power. It’s also ignorant of history: most of the world’s card networks and exchanges began as non-profits and utilities. Today they are massive profit centers. Why? Because there was a button.

I am skeptical such networks will ever get enough traction for value capture to even be possible. Their fake decentralization makes them chronically unsecure; they are neither regulated by a government nor by cryptoeconomic security. TradFi executives know better than anyone how dangerous it is to put your business on a network controlled by an unregulated button. The most likely outcome for the permissioned approach is a bunch of fragmented systems, controlled by competitors, with little adoption.

In a permissioned database with a coin, the equity holders capture even more value, as the coin is purely extractive.

Permissionless networks like Bitcoin, Ethereum and Solana need a coin. They can’t be secure or censorship resistant without one. A permissioned database doesn’t have this problem. Security is provided by the controller of the button (for better or worse) and censorship is the point.

It makes no sense for centralized databases to have a coin; that’s why the first generation of enterprise networks never had one. They even advertised the lack of such as a perk; “all the benefits of Ethereum, none of the hassles or regulatory uncertainty of ETH.”

Alas, crypto’s volatility proved to be a feature as quality coins soared in value, and regulatory clarity is coming, so the newest crop of corporate networks masquerading as chains also feature a coin.

You could argue that’s their main innovation over their predecessors. It’s almost like the purveyors of these products studied the various grifts the crypto industry ran on Main Street and decided to run the biggest one—that of an unnecessary coin—on Wall Street. It would be highly clever if it wasn’t immoral.

Not only does a permissioned network not need a coin for security, it also doesn’t need it to charge for usage. Permissionless systems have to charge for blockspace because it is scarce—scaling decentralized networks is hard. They charge the required fees in their own coin (as opposed to an off-chain payment or a stablecoin) to preserve censorship resistance.

Corporate databases can scale endlessly. And since they are not public, they don’t even need to charge a fee. They can still charge one to extract value, but it would make more sense to charge in fiat money, like the TradFi systems they emulate always have.

Some permissioned systems claim to need a coin to incentivise adoption, but this makes no sense. If the network solves a problem for its participants—the ostensible reason this network was created—why do you have to also pay them to adopt it? Amazon doesn’t pay companies to use AWS, it charges them for it, AWS is useful. Actual blockchains pay miners and validators because they are mercenaries, they do work that is valuable for others (users, dApps, etc) and wouldn’t show up to work if they weren’t paid. Corporate databases where the validators are the same firms who rely on its service don’t have this problem.

(It’s also worth noting that even on permissionless systems, paying for adoption, as opposed to security, has never worked).

A pointless coin that is paid to entities who would be doing an activity anyway has no value capture mechanism. Like the users of a social media platform, it is the value capture mechanism. In time, it will get dumped on unsuspecting buyers who mistake it for a real cryptocurrency. This extraction is compounded by the fact that in any permissioned system, a coin can only be launched unfairly—the controller of the button also decides who gets to earn the coin on day 1 (usually itself).

I am obviously skeptical such networks will ever get real adoption (though they’ll advertise plenty of fake adoption to promote their coin). But even if one did, the coin would also have a major supply problem. Not only is the demand for the coin of a database contrived, but the real supply is infinite.

In any permissionless system, the decentralization of the network makes it hard to change the inflation schedule. That’s why people trust Bitcoin’s 21 million supply cap. But in a permissioned setting, the controller of the button can change the inflation schedule whenever it wants. If the network advertises privacy, they can do so privately.

Endless shades of grey

There are many different designs that this mental model doesn’t fully capture, such as the utility and governance coins of decentralized applications, L1 networks with permissionless but highly concentrated validators, and L2s. I am still learning myself how to best model token value capture for all of these. But as a general rule of thumb, I always ask two questions:

1. Does the project even need a coin to function? If it doesn’t, then the coin won’t capture substantial value.

2. Is there a proximate equity holder who also holds a claim on the project’s success? If there is, coin holders won’t capture substantial value.

Let’s run through a few examples of what I mean:

• Bitcoin can’t exist without a coin, and there’s no equity-enabled entity that has any major say over how the network operates. Thus BTC is the primary way to capture value from the underlying network’s success.

• Ethereum can’t exist without a coin, and there’s a weak Foundation that goes out of its way to not be captured. ETH is the primary way to capture the value of the network’s success.

• Hyperliquid launched without a coin so it can work without one. It has since issued one to decentralize the network, but the validator set is highly concentrated. Also noteworthy: there is a powerful founder, a for-profit, and a Foundation. The project is highly profitable, but I believe these factors nevertheless limit its upside. I am a fan, so hope to see them decentralize further. Doing so would add value to their coin.

• Base doesn’t have a token and does well without one. All the value capture goes to Coinbase shareholders. If they do issue a token, they’d have to change the design drastically for it to capture value.

• The value capture of any L2 token, even ones tied to a more decentralized design, will always be throttled by the fact that some of the security is derived from an L1.

• Polymarket has been a massive success without a coin. That in and of itself raises serious questions about the value-capture properties of any future coin.

• DeFi lending protocols benefit from having a governance token more than decentralized exchanges; credit systems require an underwriter and equity layer to scale, trading systems do not. Ceteris paribus, this dynamic means the coins of lending protocols have a higher ceiling than trading ones.

• There is a long tail of supposedly permissionless Layer-1 systems that launched with a for-profit arm, one that simultaneously raised money against its own equity and the coin. Almost all of them were dead on arrival.

• Databases work fine without a coin.

In conclusion

It’s still early, and much remains to be learned as to what kinds of coins capture value in the long run. There have been long stretches where the above rubric would have cost traders money, and crypto bull cycles often feature the most useless coins pumping the hardest before collapsing. Today, XRP and BNB remain among the more valuable assets in crypto, despite failing my litmus test.

So I’d understand if anyone wants to dismiss everything I just said. But I stand by it. If applied over the past decade, it would have mostly picked sustainable winners, and perhaps more importantly, not blown up.

We know this because they ran the same playbook on Bitcoin. Tokenization is a greater threat, so the fight is going to be even nastier, particularly now that some kind of legislative or regulatory clarity seems likely. The pushback has already begun. For example here’s a quote in yesterday’s Bloomberg article hinting at new SEC exemptions for tokenized securities:

Securities industry insiders such as Citadel Securities and SIFMA have pushed back, warning that broad exemptions for tokenized stocks could weaken know-your-customer, anti-money laundering and other investor protections.

AML/KYC is a failed compliance approach. There is near unanimous consensus in the academic literature that it doesn’t work, particularly when put through a cost-benefit analysis. In my experience, most corporate executives and government officials readily admit this when asked in private. It’s hard not to when trillions of dollars get washed through the regulated financial system every year.

Being critical of AML/KYC doesn’t make someone pro illicit activity. It makes them honest. This particular emperor really has no clothes, and my hunch is that most intelligent people agree with me. That’s probably why this was one of my most viral tweets.

Nevertheless, the approach perpetuates itself because it provides a powerful moat to incumbents and gives ineffectual politicians someone to yell at. As a general rule of thumb, the more a corporate exec uses AML/KYC to criticize crypto, the more his own firm has egregiously violated it. Case in point: Jeffrey Epstein and Bernie Madoff’s banker has been really worried about Bitcoin’s use in illicit activity. And the trade groups that represent the TradFi firms where most of the laundering happens worry about stablecoins.

Given the existential threat public blockchains pose to some of these firms, I anticipate them to really start ratcheting up their fear-mongering over this issue. They will likely also use it to campaign for permissioned databases masquerading as chains, ones that they fully control. If left unchecked, they might use it to stall progress altogether. We’ve seen what the banks are willing to do to protect their moats in the stablecoin yield debate. Tokenization of real-world assets is just as big of a threat.

So going forward, it’s important for us to be aware of the dynamics in this debate and argue against FUD with facts and first principles. The current approach to stopping illicit activity in banking and markets has failed. The transparency and programmability of decentralized systems is a great opportunity to invent something batter. But we’ll never get there if the Suits use fear to stall progress.

The expression “crypto loves to speedrun the lessons of financial history” is stated so often that it has become a cliche. But there’s one lesson from TradFi that the crypto industry refuses to learn—to its detriment. We go through cycles of almost learning it, only to run away, usually because something breaks.

The lesson? The design principle that systems with an authority in charge of final settlement can never be treated as autonomous. Not only can the authority revert past transactions, it can censor future ones. That combined ability makes the authority liable for all transactions, even the ones it doesn’t touch.

Call it the “if you break it, you own it” principle of financial systems.

The world we know

This principle informs the design of every major payment and clearing system in the world today. The level of authorized intervention (or lack thereof) both determines who can use this system and how those users behave.

To wit:

• Credit card transactions are reversible. This is convenient for consumers, but an added risk for merchants. That’s why credit cards are seldom used for important economic activity.

• Wire transfers cannot be reversed, even after obvious mistakes or even a hack. This is why wires feature prominently in important economic activity.

• ACH payments offer limited reversibility. This is why Coinbase makes you wait days before letting you withdraw coins purchased with an ACH deposit. If you pay with a wire you can withdraw immediately.

• Security transactions settled through a clearinghouse like the NSCC are mutable inside the netting period (e.g., T+1) but considered “final and irrevocable” after. This is why Robinhood was forced to pull the plug on meme stocks in 2021.

• AML/KYC/CFT frameworks make banks liable for every payment that they intermediate. This is why banks don’t want poor customers and routinely debank clients they deem risky.

• Payments made with physical forms of money are instantly final and irrevocable. That’s why physical money has historically been popular with drug dealers, the mafia, and central banks.

I could go on, but you get the picture. Rational users of any financial infrastructure start by asking a simple question: who has the power to screw me, and how? They then adjust their behavior accordingly. Coinbase understands that it can be wronged by an ACH dispute, so it makes you wait. Similarly, Coinbase also knows it can be wronged by a chain reorg, so it makes you wait. Totally different technology, but the same risk, so the same behavior. Importantly, Coinbase’s customers base their behavior on how they expect Coinbase to act.

This design principle shows up everywhere once you know how to spot it. It can be found in the rules of credit card schemes, securities markets, clearinghouses, crypto exchanges, and even the universal commercial code. It also spelled out in the opening paragraph of the Bitcoin white paper:

Completely non-reversible transactions are not really possible, since financial institutions cannot avoid mediating disputes. The cost of mediation increases transaction costs, limiting the minimum practical transaction size and cutting off the possibility for small casual transactions, and there is a broader cost in the loss of ability to make non-reversible payments for nonreversible services. With the possibility of reversal, the need for trust spreads.

In TradFi, this sort of risk is usually dealt with via censorship. As a general rule of thumb, the more instant & irreversible a system is, the more the powers that be restrict access to it. Thus FedWire—the most important payment system in the world—is only available to the most elite financial institutions. If the Fed is going to honor every payment, then it needs to at least control who can request one in the first place. Censorship also gives it someone to blame whenever something goes wrong. Every AML or sanctions violation that’s conducted in dollars ultimately clears through the Fed, but the blame sits with the banks. Unlike the Fed, they can block payments all day. The law mandates that they do.

Bitcoin was designed to break this paradigm. It was the first (and still only) autonomous payment system that does not censor. Anyone can use Bitcoin, to do anything, and every transaction goes through, for better or for worse. Cash is the only other payment instrument that works this way. That’s why Satoshi called Bitcoin peer-to-peer electronic cash. Only a decentralized system can be both irreversible and censorship-resistant.

DeFi needs neutrality

Offering guaranteed outcomes while remaining open to the public is about the only thing crypto does better than TradFi. It makes many sacrifices to get there. But the juice is worth the squeeze, because this combination also allows for other, more sophisticated, types of economic activity, like DeFi. Smart contracts and atomic swaps are powerful tools for diminishing the types of risks TradFi has been grappling with forever, such as counterparty risk.

But the contracts are only smart, and the swaps only atomic, if the underlying blockchain is as predictable as gravity. Violating the rules of neutrality erodes those features, even if done sparingly. Once you introduce the ability to intervene, the world will conspire to force you to use it. That’s the principle of Chekhov’s Fork. This is also why permissioned systems aren’t blockchains and offer none of the desirable features.

The debates around blockchain neutrality used to be mostly hypothetical. There wasn’t much controversy when Sui reverted the chain or various others intervened to stop the Balancer exploit, mostly because nobody cares about those chains. But the past few months has seen the $200b Drift exploit on Solana and the even larger KelpDAO bridge attack on Ethereum and several of its L2s. The stakes of the debate are now higher.

The Drift exploit involved a lot of USDC, but Circle chose to do nothing, even though it technically could. I think they did the right thing. Intervening to stop one hack makes Circle de facto responsible for all hacks. This wouldn’t just be a problem for Circle, it would be a problem for all USDC users. Who is to decide what constitutes an exploit anyway? Circle can set up some kind of internal adjudication system, but it would likely move slower than the hackers, and be prone to abuse. That’s what happens with credit card chargebacks.

Circle not staying on the sidelines would also be a problem for DeFi. Most of the benefits of DeFi go out the window if stablecoin issuers start freezing and seizing arbitrarily. This is why I believe the most neutral stablecoin will eventually take massive market share. Imagine a perp DEX (like Drift) with the stablecoin collateral suddenly being frozen. Or a DeFi lending pool in Aave, for that matter. The results would be more catastrophic than any hack.

A surprising number of people within the crypto industry wanted Circle to intervene. This hack was too big they said, or North Korea too evil. But these are all emotional arguments, not logical ones. We heard similar defenses of bank bailouts during the 2008 financial crisis, and we know how that turned out. It’s bad enough if a system is designed from genesis to have an intervention mechanism. It’s a lot more corrosive if the bailouts are doled out in arbitrary fashion.

If you break it, you own it, and breaking its own rules is exactly what the Arbitrum security council did in the wake of the Kelp/Aave exploit, freezing the funds wrongfully taken in that exploit. A lot of people celebrating this decision, until the security council got sued by the victims of an unrelated incident involving North Korea. This came as a surprise to the pro-bailout camp, but it shouldn’t have. Once the Arbitrum security council demonstrated a willingness to play favorites, it opened itself up to endless liability. And now that it has intervened once, users have to discount the possibility of it intervening again. Note that this corrosive logic also works the other way. If I was part of the security council, I’d be equally worried about being sued for not intervening in the next hack.

Tellingly, the more centralized layer-2 chains that were also impacted by this particular exploit did nothing. Coinbase is already walking a fine line with Base. I strongly suspect their attorneys advised the Base team to not do anything, lest they arm their critics, regulators, and litigants with ammo arguing that Base is just an extension of its centralized service.

So where do we go from here?

Any debate about blockchain neutrality inevitably returns me to Bitcoin and Ethereum. Bitcoin intervened once to fix an inflation bug, but is now so decentralized it may not even be able to respond to the quantum threat. Ethereum forked in the wake of The DAO hack, but has regained respectability by refusing to intervene after the Parity bug and more recent Bybit exploit. Bitcoin and Ethereum, as platforms, host the most value in crypto. Nothing else even comes close. One explanation for this success is that they are the oldest chains. Another explanation is that they are the most neutral.

This debate is far from settled, and I suspect the current surge in corporate databases masquerading as chains will lead to more “experts” calling the possibility of intervention and a lack of neutrality as some kind of virtue. I also suspect they’ll all be humbled by the unintended consequences of their actions. Eventually they’ll learn the key principle Satoshi understood all along.

If you break it, or even if you try to fix it, you own it.

RE: Notice of Proposed Rulemaking — Implementing the GENIUS Act for the Issuance of Stablecoins by Entities Subject to the Jurisdiction of the OCC; Docket ID OCC-2025-0372

Dear Office of Chief Counsel:

I. Introduction

I am an adjunct assistant professor of finance at Columbia Business School, where I have lectured on blockchain and cryptocurrency for the past 6 years. I have written several books on the subject, co-authored numerous relevant papers, and consulted with both the largest banks and FinTechs on related topics. I am also a member of the New York Department of Financial Services’ Virtual Currency Advisory Board. I have been studying the topic of stablecoins closely for the past 8 years, from long before they became a mainstream issue. In particular, I have closely studied their possible intersection with capital markets and banking.

I submit these comments to express strong concern about one provision of the proposed rule: proposed § 15.10(c)(4), which prohibits permitted payment stablecoin issuers from paying any form of interest or yield to stablecoin holders, and which extends that prohibition through a rebuttable presumption to arrangements involving affiliates and related third parties. I urge the OCC to withdraw or substantially revise this provision.

My objections proceed on three independent grounds. First, the proposed rule does not track the statute. The GENIUS Act’s text on yield reflects a deliberate, narrowly drawn legislative choice made after extended public debate. The OCC’s proposed rule’s expansion of that prohibition through a rebuttable presumption covering third-party arrangements is not a reasonable implementation of the Act; it rewrites it. Second, the ostensible motivation for the proposed rule is based on an unexplained and unjustified form of discrimination, that of placing the needs of borrowers ahead of savers. Yields or rewards associated with stablecoins, even if accumulated through third parties, are a benefit to American savers—be they individuals, corporations, or even municipalities. The practice provides them with additional ways of earning income and increasing their purchasing power. The OCC’s proposed rulemaking harms this demographic, even though it is larger than the borrowers such rules are meant to protect. Third, even for borrowers, the policy rationale animating this overreach is empirically unfounded. The case for extending the yield prohibition to third-party arrangements rests on banking-industry arguments about deposit flight and threats to credit creation that I have examined and found to be without merit. The OCC should not embed those arguments in federal regulation.

II. The OCC’s Extended Yield Prohibition Exceeds the Authority Granted by the GENIUS Act

A. The Statute’s Text Reflects a Deliberate and Narrow Congressional Choice

Section 4(a)(11) of the GENIUS Act (12 U.S.C. § 5903(a)(11)) prohibits a licensed payment stablecoin issuer from paying any holder of a payment stablecoin any form of interest or yield solely in connection with the holding, use, or retention of the stablecoin. That is the prohibition Congress enacted. It is directed at the issuer. It targets direct payments. It says nothing about third parties.

This was not a drafting oversight. The yield question was among the most contested issues in the entire GENIUS Act legislative process. For months before the GENIUS Act’s enactment in July 2025, the stablecoin yield debate received Congressional attention and generated sustained and visible lobbying from both the banking industry and the digital asset industry. The question of whether yield could flow to stablecoin holders through distribution partners, FinTech intermediaries, or affiliated entities was not abstract or forgotten; it was ever-present. In fact the American Bankers Association sent a letter with recommendations on amending the GENIUS Act to all members of the House on the very day the chamber voted on it, and the very first bullet point in that letter called for the statute to “strengthen the prohibition on interest and rewards payments to mute deposit substitution by expanding the prohibition to affiliates of the issuer.”[1]

Congress decided not to. It enacted a prohibition that covers direct payments by issuers and went no further. When a legislature passes laws following extensive public debate on a specific and well-understood question, and the resulting statutory text addresses that question narrowly, an administrative agency may not infer a broader prohibition from silence. The Supreme Court has been clear that agencies must identify clear congressional authorization before claiming authority over questions of major economic and policy significance.[2] Whether third-party yield arrangements involving stablecoins should be prohibited is precisely such a question. It received exhaustive legislative attention, and Congress chose not to prohibit them.

The OCC’s proposed rule does not reflect this history. It does not explain why Congress’s silence on third-party arrangements should be read as an implicit prohibition rather than an implicit permission. The most natural reading of a statute that prohibits issuers from making direct yield payments while saying nothing about third-party arrangements is that Congress drew the line at the issuer. That is the most faithful reading of the GENIUS Act in its broader context—one that concluded less than a year ago.

B. The Rebuttable Presumption Mechanism Compounds the Statutory Problem

Even if the OCC has the authority to address third-party yield arrangements, the rebuttable presumption in proposed § 15.10(c)(4)(i) would be an impermissible exercise of that authority. The presumption deems a broad category of ordinary commercial arrangements — payments to affiliates, payments by related third parties, white-label arrangements — to constitute prohibited yield payments by the issuer, subject to rebuttal only through a process requiring the issuer to submit written materials satisfying the OCC’s judgment that the arrangement is not prohibited.

This inverts the statutory structure. The GENIUS Act establishes a specific prohibition on identified conduct. The OCC’s rule presumes that a much wider category of conduct is likewise prohibited, unless the issuer can prove otherwise. The burden Congress placed on the government to identify prohibited conduct is effectively shifted to issuers to justify conduct the GENIUS Act does not prohibit. That is not implementation; it is reversal.

Practically, this will have a chilling effect on the full range of competitive arrangements that Congress deliberately left open: merchants offering economic benefits to stablecoin users, fintech companies building yield products on stablecoin infrastructure, distribution partners seeking to attract consumers through economic incentives, use of stablecoins by a wide range of financial firms as a loss leader to offer other useful services. All of these parties would face uncertainty about whether their arrangements might be deemed to trigger the presumption. The cost of that uncertainty falls disproportionately on new entrants and smaller operators who lack the resources to navigate case-by-case analysis of a complex rule as proposed.

I recognize that the OCC has anti-evasion authority under section 4(h)(1) of the Act (12 U.S.C. § 5903(h)(1)). But anti-evasion authority permits an agency to prevent bad-faith circumvention of statutory prohibitions that exist. It does not authorize an agency to extend a statutory prohibition beyond its terms by labeling the extension “evasion prevention.” Congress prohibited something specific. The OCC should not use its anti-evasion authority to prohibit something broad. That is not evasion prevention; it is a substantive policy expansion that requires independent congressional authorization.

III. Extended Yield Prohibition Harms Savers, an Important Demographic That the OCC Ignores

All yield prohibitions are based on the assumption that competition for deposits harms borrowers by raising the cost of credit. But credit is a two-sided market, with banks acting as intermediaries. A bank account is the primary form of savings for the vast majority of American consumers and businesses. When competition forces banks to pay more for deposits, savers benefit. Barring such competition via rulemaking harms savers. While I understand the motivation, the OCC’s disposition on the yield issue shows a bias towards definitively harming savers in the hope of potentially helping borrowers. (I say potentially, because there is no guarantee that banks will pass on the savings from cheap deposits to borrowers, and as I argue in the subsequent section, they often do not). Savers are both a larger and a broader demographic category than borrowers. Why is the OCC proposing rules to harm the majority to benefit a minority?

To put the disparity in perspective, I estimate the percentage of small businesses who save money at an insured depository institution to be greater than 95%. It is borderline impossible in the year 2026 to run a business without a bank account. And yet, per the Federal Reserve’s most recent Small Business Credit Survey (SBCS), only 37% of small businesses applied for any kind of small business credit in the most recent year for which data is available, and of that cohort, only three quarters were approved.[3]

A similar disparity exists for consumers. FDIC data from 2023 shows that over 80% of Americans are “fully banked.”[4] If third-party yield-sharing arrangements for payment stablecoins were to put upward pressure on bank remuneration, then over 100 million American households, consisting of over 250 million individuals, would benefit. What’s more, older Americans and retirees who statistically are more likely to rely on interest income to make ends meet would benefit. Not everyone is a borrower, but almost everyone is a saver.

The OCC should not be proposing rules that potentially harm the economics of most Americans, particularly seniors. Just because this demographic does not have a powerful trade group or lobby organization does not mean it can be ignored. Savers are people too, and third-party yield-sharing arrangements for stablecoins can only benefit them.

IV. The Empirical Case for the Extended Prohibition Rests on Arguments That Are Not Supported by Evidence

The policy rationale underlying the OCC’s extended prohibition is ultimately traceable to the banking industry’s campaign to restrict yield on stablecoins — a campaign predicated on claims about deposit flight, threats to credit creation, and risk to community banks. I have examined each of these claims in detail. In each case, the claims fail on the merits. The OCC should not codify them as regulatory policy without confronting the evidence.

A. There Is No Established Empirical Basis for the Deposit Flight Thesis

The banking lobby’s central argument is that yield-bearing stablecoins will cause consumers to withdraw deposits en masse from the banking system, destabilizing banks and threatening financial stability. This claim is not grounded in evidence.

The first problem with the deposit flight thesis is that it ignores how stablecoin reserves work. Per the text of the GENIUS Act and the OCC’s own proposed rulemaking, licensed issuers must back each unit of their coin with a mix of short-dated Treasury bonds, bank deposits, and reverse repo arrangements. Stablecoin growth therefore generates demand for bank deposits on the reserve side, even as it may compete for deposits on the liability side. The net effect on aggregate bank deposits depends on a mix of variables, including the magnitude of these flows and the relative composition of each issuer’s reserves.

As a general rule of thumb, it is difficult to drain the banking system of deposits via other forms of private money. As evidence, consider the growth of money market funds, which per the Investment Company Institute, reached $7 trillion in deposits in 2025.[5] Like payment stablecoins, money markets were originally feared to drain deposits from the banking system. But despite their impressive growth over the past twenty years, bank deposits are also at an all-time high. Breaking down the mechanics of how these funds work explains why. Operating a money market fund requires bank deposits. So does purchasing or divesting from the shares of one. The brokers who offer money market fund investments require bank accounts. So do the bond traders who trade with them.

The issuance and use of payment stablecoins would similarly require everyone involved to have a bank account. T-bills are bought and sold for bank deposits. Repo and reverse-repo are conducted with bank deposits. Issuers will have to use bank deposits to interact with authorized participants who can request the minting or burning of their stablecoins. Adoption of payment stablecoins might change the types of deposits that sit inside the banking system at the margin, thus altering the economics of some banks, but the overall deposit base is unlikely to shrink materially.

It is also possible adoption of payment stablecoins increases bank deposits. Stablecoins live on global and borderless decentralized networks. The data shows most adoption and usage today comes from abroad.[6] It is likely that a significant portion of this demand is not substitution away from dollar banking, but rather net new dollar adoption; people selling other currencies or assets denominated in other currencies for digital dollars. When combined with the multiple ways stablecoin issuers must rely on bank accounts, this means adoption of payment stablecoins leads to more deposits in the dollar banking system. Third-party yield arrangements would only accentuate this mechanism.

There is no coherent empirical basis for the claim that growth in yield-related stablecoin adoption will, on net, reduce aggregate U.S. bank deposits. It may just as plausibly increase them, as yield-sharing arrangements increase demand for payment stablecoins abroad, requiring even greater deposits at domestic banks.

B. Any Competitive Impact on Deposits Would Affect Bank Profits, Not Bank Lending

Even accepting—for the sake of argument—that third-party yield arrangements can reduce bank deposits at the margin, the correct policy response is competition, not prohibition. Banks can pay more for deposits. They currently choose not to.

The national average deposit yield in the United States is barely above 1%, per the latest FDIC data.[7] This despite the fact that SOFR—the most relevant benchmark for overnight rates—is 3.66%.[8] Deposits at banks and credit unions are currently quite cheap, relative to the market. Not surprisingly, this dynamic has enabled high net-interest income (NII) in aggregate. Per FDIC data, U.S. banks earned $194 billion in net interest income in Q4 2025 alone.[9] The annual figure is greater than $750 billion. One globally systemically important banks (GSIB) reported $95 billion in NII.[10] All of these numbers are at or near all-time highs.

Ceteris paribus, such a high NII is only possible when deposits are cheap, but bank-based credit is not. As the data makes clear, cheap deposits are currently being funneled into bank profits, not cheap credit. The banking industry’s reluctance to compete on deposit pricing is not a sign of financial fragility; it reflects strong market power, something that competition from stablecoins might partially erode.

A modest compression of NII resulting from more competitive deposit markets would affect bank profits, not bank lending. Banks make credit decisions based on risk-adjusted lending spreads, not on the absolute level of funding costs. A bank facing higher deposit costs does not mechanically reduce its loan book; it first looks at other sources of funding, and then evaluates whether the risk-adjusted spread on new lending still justifies origination, which it typically will at any interest rate level consistent with a functioning economy. The banking industry’s conflation of profit compression with credit contraction is a rhetorical move, not an economic argument. The OCC should not accept it without scrutiny.

C. The Empirical Record Does Not Show That Cheap Deposits Produce Cheap Credit

The banking industry’s argument depends on a causal chain: cheap deposits enable cheap credit, so anything that raises deposit costs will raise borrowing costs and harm credit availability. I have examined this claim quantitatively and found it to be unsupported.

If cheap deposits reliably produced cheap credit, the extended period of near-zero deposit costs that followed the 2008 financial crisis and persisted through the early 2020s should have produced historically low rates for the types of credit that rely heavily on bank deposits. The data shows that they have not. Indeed in certain cases something closer to the opposite has happened.

Take credit cards, which are the most widely available form of credit to ordinary Americans and small businesses. Credit cards can only be issued by a bank, and the vast majority of outstanding credit card debt sits on bank balance sheets.[11] There is over $1 trillion in credit card debt outstanding.[12] But despite the current era of cheap deposits, credit card interest rates hover near all-time highs.[13] Today, the cost of deposits for banks is significantly cheaper than it was in the late 90s, but somehow, what they charge when they lend some of those deposits to credit card holders is significantly more.[14] The question of exactly how banks decide the interest rate for credit card debt is a complicated one, but the one thing I can say for sure, by looking at both the profitability of banks’ card franchises and the cost to borrowers, is that cheap deposits do not translate to cheap borrowing costs.[15]

A similar phenomenon can be observed for other popular forms of credit, like auto loans. Federal Reserve data shows the average cost of an auto loan was 7.22% in November 2025, the last period for which data is available.[16] That number has come down from its post-pandemic all-time high but is still significantly higher than the mean for the past 20 years. Pre-pandemic, banks were paying depositors almost twice as much for their savings the last time car loans were this expensive. When it comes to auto loans, cheap deposits have not translated to cheap credit.

This analysis can be conducted across banking, including for small business loans.[17] The empirical pattern has a straightforward explanation: retail and small business lending rates are determined primarily by competitive dynamics in lending markets, the cost of risk, and the cost of loan origination — not by the marginal cost of deposit funding. Banks are not price-takers in many lending markets, and their pricing power in those markets does not disappear when their funding costs rise.

Just as importantly, banks have a myriad of options of where they deploy customer deposits. They do not have to lend those funds out in the traditional sense. They can also park them as reserves with the Federal Reserve or invest them in government and agency securities. U.S. banks currently hold $3 trillion in reserve balances at the Federal Reserve.[18] They do this for a myriad of reasons, including the fulfillment of capital requirements and to make payments, but are also paid for the privilege — the Federal Reserve currently pays 3.65% for reserve balances.[19] That yield exists for complicated reasons of monetary policy but is nevertheless a significant source of income for banks.

Alas, reserve balances are money that has been taken out of the credit creation process. They do not empower small businesses or lead to job creation. They are also risk-free and have minimal operating costs. Their existence diminishes the argument that competition for deposits impedes lending. Even if banks did lose some deposits to payment stablecoins, they could simply fill the gap by reducing reserve balances, with minimal impact on the availability or cost of credit.

U.S. banks also hold trillions of dollars in government debt. Per Federal Reserve data, the largest 25 banks hold over $3 trillion dollars’ worth. Payment stablecoin issuers also hold government debt. The reserve restrictions of the GENIUS Act, combined with the economics of issuing a stablecoin in today’s rate environment, leads to a dynamic whereby T-bills and reverse-repo are likely to account for the majority of any issuer’s reserves. This currently holds true for the largest domestic stablecoin issuer.[20] Given this reality, any impact from the substitution from bank deposits to payment stablecoins is likely to be muted as far as American borrowers are concerned. Banks might hold less government debt in aggregate and stablecoin issuers might hold more, but the credit markets are unlikely to feel the difference.

If stablecoin demand continues to come mostly from abroad, overall borrowing costs may even go down, even while aggregate bank deposits shrink marginally. Banks account for a minority of debt outstanding in the U.S., thanks to our robust capital markets and the presence of large non-bank lenders such as insurance companies and private credit funds. All other forms of credit are benchmarked to government securities. If stablecoins do lead to net-new demand for dollars as many expect, they may bring down borrowing costs across the board. Third-party yield-enhancement would only amplify this phenomenon.

The implicit promise embedded in the deposit-flight argument — that protecting cheap bank deposits will result in cheaper mortgages, auto loans, and credit for small businesses — is not delivered in practice. The mechanical reasons why are clear when we look under the hood of modern banking. The OCC should not write regulations that protect bank funding costs on the theory that those costs will be passed through to borrowers when the evidence shows they are not.

D. The Academic Research Shows That Credit Is Unlikely To Be Severely Impacted Even If Banks Face Severe Competition

In an important paper analyzing the potentially destabilizing impact of a central bank digital currency (CBDC), Whited, Wu, and Xiao find severe deposit outflows from banks do not have to result in severe credit contraction.[21] A CBDC is a far greater competitive threat to bank deposits than even a yield-bearing stablecoin due to the perceived safety of holding a digital liability of the central bank. (On this point, the banking industry and I empirically agree, given their vocal support for proposed anti-CBDC legislation).[22] And yet, the model presented in this paper shows that even severe outflows from depository institutions does not have to severely curtail credit from them. Banks have a myriad of sources of funding, of which deposits are only one.

Looking at the issue from the opposite lens, a model produced by the White House Council of Economic Advisors (CEA) finds that banning all interest arrangements for stablecoins hardly improves credit availability.[23] In the baseline example, barring yield would only increase bank-created credit availability by a negligible 0.02%. The model also considers the impact on community banks with assets below $10 billion and finds that barring interest only increases their lending by 0.026%.

The CEA even examines the dynamic in the most extreme scenario, a highly unlikely one where stablecoins grow to be 6 times larger than they are today, stablecoin issuers make the economically irrational decision not to invest in yield-bearing reserve assets like T-bills, and the Federal Reserve abandons its current approach to monetary policy. Even in this extreme scenario, credit creation would only increase by single digit amounts.

This analytical research confirms the empirical observation that traditional deposits are hard to substitute away from, banks are flexible, and victimizing savers is more likely to benefit bank shareholders than bank borrowers. The OCC should not be proposing rules with such weak economic underpinning and strong adverse consequences.

E. The Real Structural Threat to Community Banking Is GSIB Concentration, Not Stablecoins

The banking industry has been careful to frame the yield prohibition as a protection for community and regional banks, whose deposit franchises are more relationship-dependent and who would be least able to compete with yield-bearing stablecoins. I take community banking seriously as a policy matter, and I do not dismiss concerns about systemic risk to smaller institutions. But the framing is strategically misleading.

The structural threat to community banks over the past two decades has not been technology firms or stablecoin issuers. It has been the continuing concentration of deposits at the largest U.S. financial institutions.[24] The GSIBs have grown their share of total U.S. deposits dramatically, particularly following the 2008 financial crisis and the 2023 regional banking crisis. They benefit from scale advantages in technology and operations, national brand recognition, and the implicit too-big-to-fail taxpayer subsidy that lowers their cost of funds relative to smaller institutions. These are direct, ongoing, and measurable competitive advantages that are eroding the community banking sector and provoking a response of greater consolidation.[25] Multiple GSIBs have announced public plans to increase their share of deposits.[26] That increase is likely to come at the expense of community banks.

Stablecoins, by contrast, are an indirect and attenuated competitive force for community banks. Community banks’ deposit customers are overwhelmingly domestic retail and small business depositors, most of whom are not currently using stablecoins and are unlikely to migrate to yield-bearing stablecoins in the near term even if permitted to. Community bank customers also skew older, which is relevant when talking about cutting edge digital technology. The community bank deposit franchise is far more immediately threatened by large banks’ ability to offer nationwide digital banking products, investment services, and other integrated financial services than by a stablecoin product.

If the OCC is genuinely concerned about the viability of community banks, the appropriate policy focus is on the structural imbalances that concentrate market power in the largest institutions. Restricting third-party yield arrangements for stablecoin issuers does nothing to address GSIB concentration. It protects the aggregate deposit franchise of the banking system as a whole, with the largest banks being the principal beneficiaries. The community banking argument is being used instrumentally to advance the interests of the institutions whose profitability is most at stake.

V. Conclusion and Recommendation

The OCC’s proposed § 15.10(c)(4), and in particular the rebuttable presumption in proposed § 15.10(c)(4)(i), exceeds the authority Congress granted to it in the GENIUS Act and rests on a policy rationale that does not survive empirical scrutiny. I respectfully urge the OCC to revise the proposed rule to confine the yield prohibition to what the statute actually requires: a prohibition on direct payments of interest or yield by permitted payment stablecoin issuers to their stablecoin holders.

The OCC should not extend that prohibition to third-party arrangements through a presumption mechanism that Congress did not authorize. The GENIUS Act was the product of sustained legislative deliberation on precisely the questions this rulemaking addresses. Congress heard from the banking industry at length. Congress considered and declined to prohibit third-party yield arrangements. The OCC’s implementing regulations should faithfully reflect what Congress enacted, not correct what the OCC or the banking lobby believes Congress should have enacted.

Economically, all arguments that stablecoins of any kind, including ones that feature a third-party reward mechanism, threaten the availability or cost of credit are weak. They are not supported by the empirical data on how bank deposit franchises impact credit today. They are also not supported by the academic literature. They are further compromised by the likelihood of offshore adoption, the banking needs of stablecoin issuers, and our robust capital markets. Regulatory rulemaking should be based on the strongest empirical analysis and economic arguments, not industry talking points.

As final considerations, it should be noted that the proposed rules would only apply to domestically licensed and regulated stablecoin issuers within the jurisdiction of the OCC. They will not apply to foreign issuers of dollar stablecoins that only serve non-Americans. The banning of third-party yield arrangements here thus leads to a perverse dynamic where foreigners have access to higher quality dollar products than Americans, even though most of the rewards enjoyed by said foreigners would be financed by American taxpayers, one way or another.

It should also be noted that all of the analysis presented on either side of this debate assumes an interest rate environment similar to the one we have today. But things change and the macro environment is always unpredictable. Barring all stablecoin yield arrangements can be severely harmful to ordinary Americans and businesses if we ever return to a high-inflation and high-interest rate environment like the 1970s. Yield in such a macro environment is not a “nice to have” for users of payment stablecoins. It would be a necessity. Without it, users of payment stablecoins would suffer a crushing inflation tax and see their purchasing power diminish daily. Rulemaking should consider all possible economic environments. If laws and regulations allow them, and market forces deem them useful, then third-party yield arrangements could help Americans protect themselves in certain interest rate environments.

I appreciate the OCC’s attention to these comments and welcome the opportunity to provide further information or engage with the agency’s staff on these issues.

Respectfully submitted,

Omid Malekan

Adjunct Professor, Columbia Business School

Author, Re-Architecting Trust (2022)

New York, NY

cc: Prat Vallabhaneni

Partner, White & Case LLP

1221 Avenue of the Americas

New York, NY 10020

Rachel Rodman

Partner, White & Case LLP

701 Thirteenth Street, NW

Washington, DC 20005

[1] https://www.aba.com/advocacy/policy-analysis/letter-to-house-members-re-genius-act

[2] See West Virginia v. EPA, 597 U.S. 697, 723 (2022)

[3] https://www.federalreserve.gov/publications/2025-march-consumer-community-context.htm

[4] https://www.fdic.gov/household-survey/2023-fdic-national-survey-unbanked-and-underbanked-households-report

[5] https://www.ici.org/news-release/money-market-funds-hit-seven-trillion

[6] https://www.brookings.edu/articles/the-rise-of-stablecoins-and-implications-for-treasury-markets/; https://www.mckinsey.com/industries/financial-services/our-insights/stablecoins-in-payments-what-the-raw-transaction-numbers-miss.

[7] https://www.fdic.gov/national-rates-and-rate-caps

[8] https://fred.stlouisfed.org/series/SOFR30DAYAVG

[9] https://fred.stlouisfed.org/series/QBPQYNTIY

[10] https://www.alphaspread.com/security/nyse/jpm/financials/income-statement/net-interest income?utm_source=chatgpt.com

[11] https://fred.stlouisfed.org/series/CCLACBW027SBOG

[12] https://fred.stlouisfed.org/series/CCLACBW027SBOG

[13] https://fred.stlouisfed.org/series/TERMCBCCALLNS

[14] https://www.bankrate.com/banking/cds/historical-cd-interest-rates/

[15] https://www.consumerfinance.gov/about-us/blog/credit-card-interest-rate-margins-at-all-time-high/

[16] https://fred.stlouisfed.org/series/RIFLPBCIANM60NM

[17] https://www.nerdwallet.com/business/loans/learn/rates-fees

[18] https://fred.stlouisfed.org/series/CASACBM027SBOG

[19] https://fred.stlouisfed.org/series/IORB

[20] https://www.circle.com/transparency

[21] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4112644

[22] https://emmer.house.gov/media-center/press-releases/congressman-emmer-announces-stakeholder-support-for-the-anti-cbdc-surveillance-state-act

[23] https://www.whitehouse.gov/research/2026/04/effects-of-stablecoin-yield-prohibition-on-bank-lending/

[24] https://banks.data.fdic.gov/bankfind-suite/SOD/summaryTables?displayResults=&endDate=2025&instType=&institutionType=banks&institutionTypeTimeSeries=&lastYear=2025&locations=&pageNumber=1&reportType=top50CommercialBanks&resultLimit=25&sortField=STNAME&sortOrder=ASC&startDate=2024&totalsType=

[25] https://www.wsj.com/finance/banking/regional-bank-mergers-lending-4f35ab11?utm_source=chatgpt.com

[26] https://thefinancialbrand.com/news/banking-trends-strategies/5-key-points-from-jpmorgan-chase-2024-investor-day-178422; https://www.wellsfargo.com/assets/pdf/about/investor-relations/earnings/second-quarter-2025-earnings.pdf; https://www.investing.com/news/stock-market-news/bank-of-america-stock-falls-after-2025-investor-day-mediumterm-targets-4333583

Neither do most of the insiders I correspond with regularly. Among them are investors, developers, engineers, cryptographers, and entrepreneurs—people who have dedicated their lives or careers to Bitcoin and blockchain.

I don’t recall the last time I discussed Satoshi’s identity with any of them. It has never seemed important.

And for the record, I did my first Bitcoin transaction over a dozen years ago (alas, only for a friend). I’ve written books about it, taught it to masters students for years, and explained it to billionaires, princes, and Fortune 100 CEOs.

I’ve never voluntarily speculated on Satoshi’s identity in those conversations. I’ve only mentioned that not knowing aids Bitcoin’s mythical lore and origin story, both of which are needed to be accepted as money. Not having Satoshi around also aids decentralization (maybe to a fault).

Tellingly, Satoshi’s identity almost never comes up organically on Crypto Twitter or at academic conferences. We talk about the quantum threat and economic security and block sizes and ordinals, but we never talk about the creator. It’s not important.

Meanwhile, as the virality of the John Carreyrou piece indicates, there are lots of outsiders who care. For them, figuring out Satoshi’s identity is very important, maybe the most important piece of knowledge when it comes to Bitcoin.

This dichotomy is telling. If you work in crypto, you likely believe in two things:

1. Bitcoin solves important problems

2. The fact that it solves important problems means it, or something like it, was inevitable.

Looking back, the same seems true for other inventions. We don’t debate who invented the automobile or television, and there are no statues to Tim Berners-Lee as far as I know. That’s probably because these things feel inevitable in retrospect.

The people who really care about Satoshi’s identity might find Bitcoin interesting as a concept, but are undecided on its importance as a solution. The human drama is interesting to them because the use cases are not.

To be fair, Bitcoin’s drama is enhanced by the fact that new types of money are rare, it has appreciated in value, and the inventor holds (or held) a sizable sum. But this too is revealing. If you believe in crypto then you believe in the power of incentives. The incentive to “touch that money” is massive. So Satoshi either lost the keys or died before transferring them.

Either outcome leads me to another important conclusion: Satoshi didn’t think Bitcoin was a big deal and wasn’t sure it would amount to much. He was just a tinkerer trying to build something cool, not a Monetary Prophet who came down the mountain with the white paper engraved in stone. (If you don’t believe me, read its first paragraph, then read this email).

That mix of curiosity and humility is something he shared with many other great inventors throughout history. He didn’t just conjure a workable solution, he stumbled onto it. If he hadn’t, then someone else would have, eventually.

And that’s why the smart people who really understand the elegance, novelty, and utility of Bitcoin—not the haters, but also not the pathological sycophants and evangelists—don’t care who invented it.

Great ideas almost invent themselves.

Why do blockchains have native coins at all?

Native coins provide two important functions on real blockchains such as Bitcoin, Ethereum, and Solana.

First, they provide an incentive mechanism that helps secure a permissionless network where the nodes who update the ledger, be they miners or validators, are anonymous and free to come and go. Introducing a native currency to deliver this incentive was arguably Satoshi’s greatest contribution to the older domain of digital currencies.

Permissionless networks have a Sybil problem. How do you prevent one participant from pretending to be many and “stuffing the ballot” in consensus?

In Bitcoin, Sybil resistance comes by way of Proof of Work. Stuffing the ballot has an energy and hardware cost. The only reason miners are willing to spend money to do this costly work is because they get rewarded in the native coin.

Why the native coin and not some other asset, like a stablecoin? Because the native coin has the strongest inflation protection and property rights. If you pay miners in stablecoins, they have to both worry about the purchasing power of that coin and the freeze & seize capabilities of the issuer.

Bitcoin miners spend fiat money to acquire bitcoin because it’s one of the more reliable assets out there: fixed inflation & strong property rights, provided by a decentralized protocol that is hard to change. And since miner revenues are denominated in bitcoins, they have an incentive to act honestly. Bitcoin’s success increases their profit.

Proof of Stake networks like Ethereum and Solana use the native coin for an even more direct form of Sybil Resistance. Validators have to buy it and stake it to propose & attest to blocks. Their “skin in the game” is purely financial. On Solana, this means cost of capital. On Ethereum, it means cost of capital and a risk of slashing.

Once again, it wouldn’t make sense to do this with any other asset, because every other asset has more risk. And once again, validators on a PoS chain have a strong incentive to perpetuate its security by acting honestly. They are the largest natural holders of the native coin so they benefit the most from the network’s success.

The second reason permissionless blockchains have a native coin is for their fee mechanisms. Charging fees in the native coin, as opposed to any other asset, has several benefits:

• It creates circular alignment for miners and validators. Transaction fees, which are their other source of revenues, also enjoy strong property rights.

• It provides censorship-resistance to users. If fees could be paid in any other coin, then whoever controls that coin (like the issuer of a stablecoin) can deny access to some users.

• It allows PoS chains to offer native yield.

• It enables an auction mechanism for block space, a scarce commodity in any permissionless setting.

Putting all of this together creates an impressive flywheel of security and value creation. Miners and stakers want to earn the native coin more than any other asset, so they behave honestly. Users in turn would rather pay fees in the native coin than any other asset, so they are willing to hold more of it.

Thus: in any permissionless setting, the value prop of the native coin is intrinsically tied to both the security of the blockchain and its network effects. You could even argue that owning the native coin the best way of monetizing those network effects.

Throw in the programmatic inflation, and the fact that total transparency allows anyone to confirm the existing supply, and you end up with a highly desirable asset.

Most people think Satoshi invented a decentralized infrastructure to enable a new kind of currency. I think he invented a new kind of currency to secure decentralized infrastructure. To me, this order of operations is even more bullish.

Permissioned databases don’t need any of this

Permissioned databases don’t need a native currency at all. Usability and security for these networks emanate from an authority, as opposed to incentives.

In a permissioned setting, Sybil Resistance is achieved by way of a dictator. A single entity decides who gets to participate in consensus. Those who satisfy the needs of the dictator are rewarded with continued participation. Those who run afoul are de-permissioned and banished.

The purveyors of corporate databases masquerading as blockchains act like this is some new and clever design, but it’s not, it predates Bitcoin by decades. Byzantine Fault Tolerance was first proposed in 1982. There was no mention of a coin back then because none was needed. Block proposing in most BFT chains is done via Proof of Authority rotation, a revealing name if there ever was one.

Permissioned databases also don’t need a native coin to charge fees. They can avoid fees altogether, or use fiat money. Permissionless blockchains need to charge fees because they are also public: anyone can do anything, including spam & DDoS attacks. That’s one of the downsides of censorship-resistance.

Corporate databases have built-in censorship, so the whole thing is moot. The known (and likely regulated) entities who help run this database aren’t going to let anyone participate. They might claim to be public today, but it’s only a matter of time until money launderers and terrorists start using their network, and the authorities come knocking. The executives who run these platforms won’t go to jail over a principle, they’ll censor aggressively.

A system built on censorship can still charge fees but it doesn’t have to, it can just KYC everyone and kick out undesirable users. That’s how TradFi works, and permissioned systems are built like TradFi.

What about incentives? Some permissioned networks claim to use a native coin to incentivize validation & usage. This is performative, and mostly pointless. If the dictator really wants to get others to join, it can pay them directly, or give them equity in the platform. Indeed, in any permissioned setting, contractual obligations trump database entries.

Besides, being paid incentives in any coin is only effective to the extent somebody else is willing to buy it. Incentives only create the supply, there must be a different source of demand.

Bitcoins aren’t valuable because miners get paid in them, they are valuable because other people want to hold bitcoins as a form of money and need some to pay transaction fees. ETH and SOL are valuable because some people want to stake them, and other people want to use them to pay gas fees.

The native coin of a permissioned database has no clear source of demand. It’s also not a good way of monetizing the platform’s success because the companies involved have equity. On a platform controlled by a corporation, equity will always trump a coin.

And that’s before you realize the coin’s serious inflation problem.

The real inflation rate for the native coin of a permissioned database is infinity

Bitcoins are valuable in part because their inflation is controlled by a decentralized protocol, one that is hard to change. The cryptocurrency’s 21m hard cap—arguably its more desirable feature on a memetic basis—is only true so long as the underlying network is decentralized.

If I announced a competing OmidCoin, one with an even more limited hard cap of 2 million coins, but with only five miners, all of whom are my friends, the market wouldn’t treat it as being even more valuable than bitcoins. It would treat it as a joke, because there is no mechanism to hold me accountable to that supply cap. I can change it whenever I want by messaging my buddies.

A permissioned network run by a corporation or consortium works (or more accurately, doesn’t work) in the same way. All promises of supply limits are only as trusted as the dictator in charge. Why? Because the permissioning authority could always order the other validators to change the inflation rate and threaten to de-permission the ones who don’t.

This means a single entity can print ever larger amounts of this coin at will. It can double the supply tomorrow and quadruple it next week. There is no physical mechanism to stop it, even if the network has two-hundred validators, or two-thousand.

A more optimistic observer might argue that, just because an authority can do this doesn’t mean it would want to. Maybe the authority means well, or its executives wear really nice suits. That’s a cute theory. Too bad it flies in the face of thousands of years of monetary history. Roman emperors and contemporary central bankers meant well, too.

The whole point of Bitcoin (and everything else in crypto that’s come after it) is the sober realization that power does corrupt, incentives do matter, and trustless solutions are more likely to stand the test of time.

Besides, if the permissioning authority didn’t have ill intent, it would have either launched an actual blockchain or just built a corporate clearinghouse.

The native coin of a corporate database is a solution looking for a problem. But that’s not even the worst of it, because most permissioned systems advertise some kind of privacy. That might be useful for certain kinds of financial settlement, but has a dark side: it means there’s no outside transparency.

Lack of transparency on a permissioned system means the dictator can inflate the supply of the native coin without anyone knowing.

This is a design element so stark that it’s worth repeating: Unless a permissioned network allows anyone to run an independent node—one that allows unaffiliated parties to see every transaction and verify the entire state—then the permissioning entity can print endless amounts of the native coin without anyone knowing.

It could be doing so right now.

That’s not true for Bitcoin. I run my own node and can confirm the supply whenever I want. Same for Ethereum, thanks to the node we run at school.

Alas, this verification is impossible on any closed-system that advertises privacy. The only entity that knows the total supply of the native coin is the one with the greatest incentive to inflate it, then lie about it.

How convenient.

In conclusion

Permissioned databases don’t need a native coin. They can still issue one—greed is a more powerful force than even gravity—but they can’t guarantee the property rights of whoever holds it, provide an organic source of demand for it, or even prove its total supply.

When you put all of this together, you end up with a digital asset that has less fundamental value than a memecoin.

I don’t like memecoins, but can say a few positive things about them. If they exist on any permissionless system, then their holders enjoy hard-to-violate property rights. If they are issued via a dApp like Pump Fun, then their supply is fixed and knowable.

Most of all, the memecoins who admit to being one don’t have a point. In that sense, they are authentic. They might be an expression of nihilism, but not outright deception.

The native coin of a permissioned network can’t even claim that, which is why on a long enough timeline, it will be worthless.

This is not investment advice. It’s common sense.

Why am I writing this?

Working in an emerging field requires an open mind. Things change, so must you. Seven years ago, I would have told you that permissioned or “enterprise” blockchains—networks where a single entity controls who gets to participate in validation and consensus— were a good idea. If nothing else, they let institutions “build the right muscles” before using permissionless chains, where anyone could be a validator. I even worked on a few prototypes inside a large bank.

But I was wrong in a fundamental way—and not only because most attempts at building an enterprise or corporate chain have either been shut down or ended in disaster. I was wrong because a permissioned network where validation is gatekept by an authority has little to do with the technology that underlies Bitcoin, Ethereum, and the other decentralized platforms that move trillions of dollars in value, value that would not and could not exist without these networks.

The defenders of the permissioned approach keep defending it in the way college students defend communism: “It’s a good idea, it just hasn’t been implemented correctly.” That’s their right. I on the other hand can’t deny the crucial nuances I’ve come to appreciate after years of studying and teaching this subject. Nor can I just stand by as a new crop of purveyors pretend like their networks have anything in common with the real crypto that is revolutionizing finance and beyond.

So going forward, I will no longer refer to any permissioned network as a blockchain, and vow to correct those who do. It’s fine to call these “corporate databases” or “enterprise platforms” or “TradFi infrastructure.” But if the consensus isn’t permissionless, the network isn’t a blockchain.

This isn’t a semantic argument—the term “blockchain” didn’t appear until years after the invention of Bitcoin. My goal is to use the resurgent interest in corporate databases from TradFi firms, FinTechs, and even crypto VCs as a teaching moment. There are specific reasons why permissionless systems use hash functions and public-key cryptography to record state or trigger transactions, in a manner that leads to novel features like immutability, tokenization, smart contracts, and atomicity. Those reasons don’t exist for a permissioned chain. Neither do the features.

Just as a professor simply holding a hammer doesn’t make him a carpenter, a corporate database doing pointless encryption doesn’t make it a blockchain.

Back to basics: why hash a block in the first place?

Blockchains get their name from the way they use cryptographic hash functions to bind one batch of transactions to another. Proof-of-work chains also use hashing for the mining tournament, but all permissionless chains use them to create an easily verifiable sequence of past events.

[Note: if you’ve never studied how hash functions work, now would be a good time to get up to speed]

By hashing the content of every block, but also throwing in the hash of the prior block, blockchains turn the data in all blocks into an ordered commitment. The miner or validator who proposes the latest block doesn’t simply say “here are the latest transactions, which I have vetted”, it says “here are the newly vetted transactions, along with a commitment that they come after the proper sequence of prior transactions, the sequence that we all agreed upon a block ago, and two blocks ago…all the way back to the beginning of time.”

In my class I use the crude analogy of making a daily smoothie with different ingredients, but always throwing in a little bit of yesterday’s smoothie while making today’s. The resulting concoction could only result from every ingredient ever used, in that order.

Permissionless networks need this mechanism because their decentralized nature means there is never an “official” record of past events. How could there be if there is nobody in charge? Nor is there ever a single “correct state” for the present. In a permissionless setting, there is only ever a generally agreed upon history, one that leads to a very specific current state.

Hashing makes it easier for all honest participants to continuously confirm that they agree on both the past and the present.

We often cheat and say that blockchains are “append-only” or “immutable” ledgers. But that is not technically true. I run my own Bitcoin node, and I can easily change the data that it holds. It’s just a little computer with a database, one that I control. If I wanted to, I can always change the transactions that resulted in my current Bitcoin balance from a fraction of one coin to 10. Nothing physically prevents me from doing this—certainly not the SHA256 hash function.

What hashing ensures is that the rest of the network could instantly recognize that something is off, and that my node is no longer in consensus. It’s not that my edited balance is wrong per se, it’s just that the coins I just gifted myself are no longer bitcoins. They are their own thing, on a database that is only trusted by me.

Thus, the immutability of permissionless blockchains is a purely social construct. Node operators can change the past, but other node operators will find out immediately and treat those renegade nodes as outside consensus.

In the absence of an authority, the history that everyone—except Omid and his 10 whatevercoins—agree on is the truth. This elegant use of soft power and digital peer pressure is one reason some of us find these systems almost magical. They are self-reinforcing in the same way that money is.

The closest thing we have to their design in a non-financial context is language. There is no “official” version of the English language. There’s just the version that most people speak. You are free to make up new words, but soon you’ll be speaking a different language.

You’d be out of consensus.

Corporate and consortium databases don’t need any of this

Imagine a hypothetical internal “blockchain” deployed inside a single bank. We can call it Conexxit. Its purpose is to speed up crossborder branch to branch payments within this bank. It has 3 nodes running some simple BFT-style consensus mechanism. It has “blocks”, and each one is tied to the prior one with a hash. The people who built this database justified its deployment by telling their bosses that it was a blockchain, so the record of entries would be immutable. Once an entry goes in it can’t be changed.

Now imagine that a software bug leads to a $100 payment being accidentally recorded as a $1 billion one. Other than having too many zeros, this transaction is valid, so it passes through consensus and gets recorded by all 3 nodes.

What do you think happens next?

Does the bank just go on pretending like the undeserving recipient now has a billion dollars? Does it say “We know this is a mistake, but Kenny in IT told us blockchains are immutable and ‘a single source of truth’, so there’s nothing we could do, we’ll just have to eat the loss”?

Of course not. The bank will either ignore what the blockchain says, or order Kenny to change the buggy entry in all 3 nodes. Either solution is easy to implement, even if the network uses the most hardcore hash functions. One corporation controls all the nodes.

So what did hashing achieve for Conexxit? Nothing. This so-called chain turned out not to be one at all. It was mutable all along and there was no real commitment involved. I would go one step further and argue that hashing is a net negative for such a ledger. It adds pointless computation and complexity.

This is the reality of every permissioned database, even if it has many nodes and runs consensus. In a decentralized system, the official record of transactions lives inside thousands of independent nodes who constantly compare hashes to make sure they are speaking the same language. Unless they all collude or are all corrupted, the commitment sticks. But in a centralized setting, a single authority could order all the nodes to change a prior entry. And the nodes that don’t want to? They can be kicked out.

It matters not whether the permissioning entity wants this to happen. Nor does it matter if they promise to not exercise that power. So long as one exists, it can just be compelled to by a government. Kenny in IT might be the staunchest believer in immutability on earth. But his bosses, who answer to bank regulators, are not going to risk fines and a consent order to stick an ideal.

To be fair, it took even me a long time to realize why hash functions are useful in a permissionless setting. It turns out many other smart people didn’t realize this either. Years ago, during the last bubble of enterprise networks masquerading as chains, Amazon rolled out a new kind of database on AWS. They gave it the futuristic name Quantum Ledger Database (QLDB) marketed it as an “immutable database” that used hashing to create an “append only” ledger. In blockchain parlance, it was a single-node ledger with no consensus.

It got some press at the time and was marketed for activities where an audit trail was important, like finance. I myself likely cited it as an example of why “blockchains are better infrastructure.” But it never went anywhere and was quietly shut down. Why? Because people realized that QLDB was only immutable to the extent you trusted Amazon to not re-write a past entry, then rehash everything to make it seem like it was there all along. But if you are trusting Amazon to be honest, why not just have them sign an affidavit saying they haven’t changed anything? Why even bother with the hashing?

Hashing is only useful for data integrity if you commit the hash to an external location that you do not control and cannot change. That’s why serious blockchains optimize for having as many independent nodes as possible. It’s why I posted a SHA256 hash of my last book not on my website, but in the Bitcoin blockchain. It’s also why the guys who literally came up with the idea of using a chain of hashes as an audit trail and launched the first blockchain committed to the New York Times classified section.

Permissioned networks have no endogenous property rights

Hashing without an independent commitment is pointless. This futility doesn’t just apply to a three-node database inside a bank. It also applies to a shared network run by a consortium.

To see why, imagine a hypothetical consortium network founded by a software company and joined by a dozen TradFi intermediaries. We can call it the Bundesland Network, or BS. Consensus is permissioned and restricted to the twelve intermediaries running a simple BFT variant. There is a Foundation that runs governance, and it works with the software company to permission and de-permission validators. BS is often touted for its enterprise-grade control and privacy.

This network also uses hash functions, supposedly for commitment and immutability. But like the professor wielding the hammer, the claim is make believe. The permissioning authority could always order all the nodes to change a prior transaction then calculate new hashes. Any node who doesn’t comply with that order is de-permissioned from consensus and kicked off the network.

This method of violating the supposed immutability of BS chain would work, every single time. Not only that, but the changing of a prior entry would have unanimous participation from all network operators, for the simple reason that the ones who didn’t want it to happen have been kicked off.

It would help a little bit if BS chain allowed independent (non-validating) node operators in the way Bitcoin and Ethereum do. Those would at least be able to see that something has changed. If the change impacted one of their own balances, they would see it change in real-time, and understand the past has been modified. But they couldn’t do anything about it. For Bitcoin and Ethereum, independent nodes are the final arbiters of truth. What they believe has happened will always be the official history. But on BS chain, a single entity decides.

The BS network is not a blockchain. “Truth” for that database is what a company or foundation decides, backwards and forwards. The permissioning entity could even decide to halt the chain if it felt like it, nobody has the technical power to stop it. Even if it didn’t want to halt the chain, it can be compelled to by an outside authority, like a government. This is why there’s no such thing as tokenization on a corporate network.

It’s a conclusion so stark that I need to make it again: corporate networks might have database entries that represent financial assets, but those are not tokens. This too is not a semantic argument, it’s a simple observation of the unavoidable fact that permissionless systems give asset owners property rights that corporate networks do not. That’s what a token is to me, and why it’s special.

The other way to think about this is that in a permissionless network, the property rights of user assets are endogenous to the system. There’s no one in god’s green earth who can easily deprive you of your Bitcoins by manipulating consensus, not without spending many billions of dollars. But in a corporate network, a single entity has the power to destroy your assets, freeze them, seize them, or undo the past to make it look like you never owned them in the first place.

The property rights of user assets in a permissioned network are exogenous to the system.

And if there are no tokens, there are no atomic swaps. The permissioning entity can order all nodes to reject a single leg of a two-part transaction. There are no smart contracts on corporate network either. There is just code that acts on database entires, and it serves at the pleasure of a company or Foundation. Whether that code executes properly, or executes at all, is for them to decide.

The simple act of permissioning throws all of the features that make real blockchains interesting out, returning us to the same features TradFi offers all along, with a bit of window dressing. The lack of endogenous property rights even applies to a BS Coin, where the network ever to issue one, even though permissioned consensus doesn’t require a coin at all. The company or Foundation has all the power to inflate the supply, reverse past transactions, or steal your coins.

And since BS chain advertises itself as offering total privacy, you may not even know it until it’s too late.

What about reputation and legal recourse?

Defenders of consortium-databases-masquerading-as-blockchains will often respond to my critiques by arguing the rules of such platforms could be written to prevent historical revision. They also argue that kicking members out would be bad for business.

This might be true, but is irrelevant to my argument. Contractual obligations and business interests are TradFi things that apply to every important database. They have nothing to do with cryptography or consensus. The whole point of real blockchains was to achieve reliable outcomes in the absence of such.

Contractual obligations and business interests also don’t hold up in the face of governmental pressure. Imagine if SWIFT decided to migrate its entire messaging system onto a blockchain and declare it “immutable” and “censorship-resistant.” Would that exempt it from enforcing sanctions? Of course not. SWIFT is still in charge and can do whatever it wants, and governments know this. This is yet another reason permissioned database are dumb.

Permissionless networks provide a sort of plausible deniability in the face of governmental pressure.

Yes, OFAC could pressure US-based Bitcoin miners or Ethereum validators to censor certain users. But doing so wouldn’t prevent those transactions from landing eventually. It would just harm the economics of American miners and validators, which is a good reason for regulators not to disadvantage them.

Corporate networks don’t have this out. Not only that, the very presence of a permissioning authority creates a “regulatory throat to choke” for whatever rules a governement—any government, anywhere—wants to impose on ALL validators. Might the U.S. government someday do this? Might the Chinese? Or Brazilian or Swedish? The answer is always yes.

There’s no government that would conclude “we have extensive rules on how TradFi databases are supposed to operate, but this one uses encryption, so those rules don’t apply.” Virtually all financial regulations in effect today were written with a centralized intermediary in mind, and that’s what a corporate database is, a centralized intermediary whose demands and dictates are respected by a bunch of cronies, lest they be de-permissioned and kicked out.

If you follow this logic all the way through, another important blockchain feature disappears.

Corporate networks will never be truly public, as far as users are concerned. In time, they’ll have to KYC everyone.

Imagine a hypothetical corporate database optimized for fast agentic payments built by a leading FinTech. We can call it Rhythm. Validation is restricted to the highly regulated financial firms who join its consortium, but usage is public—allegedly. The marketing docs claim that anyone could install a wallet and start using the network. This supposed “public, but permissioned” design is not new. It was originally proposed by Libra long ago.

Now imagine that news breaks that Hamas, the Lazarus Group, and Vladimir Putin have found Rhythm and are actively using it. What do you think happens when governments show up and start demanding censorship? Will the licensed and regulated entities that run this network risk everything to keep the network public? Will the FinTech that founded the whole thing and its licensed and regulated partners say “I’m sorry, we’d love to follow your demands, but our marketing docs say this network is public, so Vlad and Kim are allowed even though we have the power to censor anyone”?

I don’t think so. They certainly didn’t with Libra.

But that’s not all, because regulators are smart enough to know that so long as Rhythm remains public and reliant on only cryptography identity, sanctions evasion and money laundering are likely.

So they’ll demand the elimination of any non-custodial functionality and for every user to be KYC’ed. And the corporations who run validation will have no choice but to comply, particularly after their new hire Kenny in IT points out the unfortunate reality that due to the permissioned nature of the network, and the fact that the validators themselves are KYC’ed, then regulators (and Federal prosecutors) could always pin a single problematic transaction to a specific corporation. Here a Fintech processed a payment to a terror group. There a card company financed North Korea.

Once again, if all it took for a payment system to avoid expensive KYC was to employ elliptic-curve cryptography, then every FinTech and card network would have done so long ago.

So what’s the cryptography, Kenneth?

It took me years to realize this, but decentralized systems like Bitcoin and Ethereum only work due to a highly specific mix of elements: Hashing, public-key cryptography, permissionless consensus, independent nodes, and economic incentives.

Most databases, including the most centralized ones, use at least some of these features. Google Docs has consensus. Hashing is how passwords and pin codes are stored. It can also help with data compression. PKI secures the entire internet. But only the strange brew that was originally concocted in 2008 yields what almost everyone understands as a blockchain, and only a blockchain offers censorship-resistance.

Stripping out even one of these components, like permissionless consensus, doesn’t yield something that’s “almost a blockchain.” It yields a pointless, messy thing, in the same way that stripping out the cement out of concrete doesn’t lead to a different kind of building material, it leads to wet sand.

And part of you probably already knew this, because even permissionless systems aren’t foolproof. Ethereum reverted the chain due to the DAO hack. Sui seems to revert itself monthly. The chancellor was on the brink of a second bailout for Balancer across multiple chains not that long ago. It takes a fully decentralized and neutral stack, like only Bitcoin, Ethereum (and hopefully Solana someday) can claim to even hope to be immutable and censorship resistant.

The purveyors of corporate networks masquerading as chains have always known this. R3—a bank-owned fake-chain developer who was once heralded as the vanguard of the “blockchain, not Bitcoin” movement used to explicitly state this in its marketing materials. Its flagship product Corda wasn’t a blockchain because it didn’t need to be. But it got in trouble with the TradFi firms using it to do innovation theater, so it stopped. They wanted to keep the ruse going. It disappoints me greatly that now so do some crypto-native firms. They want to play make believe.

But the world is a harsh and dangerous place, one that is fragmenting before our eyes. Doing cryptographic cosplay won’t work. We all have our fantasies, and there seems to be a new corner of the crypto industry that believes if you pretend hard enough that a corporate database is like a blockchain, then it will be. But that’s not how things work.

I would love to have a full head of hair, but here we are. Wearing the most expensive wig won’t turn me into Brad Pitt. Similarly, using hash functions and cryptography won’t make a corporate database a blockchain. It won’t make the entries in that database tokens, or the code it operates smart contracts. It will make it a bad database.

The corporate database fantasy has gone on for too long, sucked up too many resources, enabled too many grifters, and distracted our industry from the real prize of credibly-neutral infrastructure with endogenous property rights. Systems that can’t screw you, no matter how badly corporations or governments would like them to. That’s the prize.

Just yesterday, Minneapolis Fed President Neel Kashkari reiterated

his belief that “crypto is utterly useless”. Bitcoin-hating FT columnist Jemima Kelly double-clicked on this and tweeted “spread the word.” They joined the likes of Nouriel Roubini and Paul Krugman in broadcasting the view that crypto has no benefit.

The remarkable thing about holding this view in 2026 is that we are long past the point of crypto’s potential being theoretical, as it was early on when some of us theorized institutional embrace followed by mass adoption. Different aspects of digital assets are being embraced by the largest organizations on earth.

This reality begs a series of questions:

If stablecoins are “utterly useless” then why are Square, Stripe, PayPal, and Visa—literally America’s most successful payment firms—making them a centerpiece of their growth strategy?

If tokenization is useless why are the DTCC, NASDAQ, and NYSE—the world’s most important capital market stack—diving in?

If DeFi is useless then why are Blackrock and Apollo—among the largest asset management and PE firms, holding $15 trillion in assets combined—investing directly in DeFi protocols like Uniswap and Morpho?

If Bitcoin is useless then why are the Harvard endowment, Millennium, and Mubadala holding significant stakes via ETFs? Harvard has the largest university endowment in the country. Mubadala is a top ten sovereign wealth fund. Millennium is a top 5 hedge fund.

In his interview, Kashkari started with the assumption that stablecoins have no domestic appeal. But if that’s true, then why are the largest American banks expending political capital to fight stablecoin yield? They went to the White House just yesterday to make their point!

The Clarity Act only applies to American issuers serving American holders. If there is no utility then why do they care?

Does Neel Kashkari know more about banking than Jamie Dimon and Brian Moynihan? Does Jemima Kelly know more about payments than the Collison brothers and the board of directors of Visa?

How is it that Roubini and Krugman, who have never managed money, are better at endowment investing than Harvard? Or better at Emirati sovereign wealth fund management than...well..the Emiratis?

Step aside Izzy, Ray, and Paul. You guys might have made billions of dollars for your investors over the decades, and you might have books written about you, but there’s a newspaper columnist out there who can manage a hedge fund better than you.

Or at least she must think she does. As do all of these other people. The world’s most accomplished leaders in a broad range of industries all think crypto is useful in their domain, but the haters still think they are wrong. Every single one of them!

Crypto skeptics must think they are the smartest freaking people on planet earth.

One of the most important lessons I try to impart on my students is the importance of being open-minded and questioning your own assumptions, particularly during our current moment of technological and social upheaval.

We spend a lot of time in the classroom talking about what is bad about crypto and what clearly doesn’t work. We debate the loftier promises, like Web3 and DePIN. I give entire lectures about the catastrophic failures (Luna), the dumb ideas (memecoins), and dangerous ones (DATs).

But nothing drives that lesson home better than the hubris of crypto-hating experts. I almost feel sorry for them.

But do they? Specifically: If a bank doesn’t have to pay much to attract deposits—possibly due to regulations that bar competition—does it pass the savings on to borrowers?

And does the banking industry do this as a whole?

Fortunately data on aggregate banking behavior is easy to come by, and we happen to be living in an era of relatively cheap deposits. So let’s see if the thesis holds up.

First, let’s consider US deposits in aggregate. Per the FDIC, the average yield that American banks pay for demand deposits (checking, savings, etc) is currently close to 1%. This is far lower than the overnight rate of 3.65% which is set by the Fed.

Bank accounts are currently paying just one-third the interest savers can get from a vanilla government money market fund, with less risk. Deposits are clearly cheap.

But you know what isn’t? Credit cards, which are the most widely accessible type of debt in America. Credit cards can only be issued by banks, the same firms currently enjoying historically cheap deposits. Alas, the average interest rate on cards is currently over 20%. That figure is near an all time high.

The rate that banks charge for credit card loans above the prime rate has climbed steadily for the past 11 years. There is over a trillion dollars in credit card debt, and the vast majority of it sits on commercial bank balance sheets.

When it comes to credit cards, cheap deposits do not translate to cheap credit.

Next, auto loans. There too, rates for new car loans are currently near the highest they’ve been in almost 20 years, at 7.22% on average. They’ve come down slightly from their post-COVID highs, but are clearly “not cheap” the way deposits are. In fact bank accounts paid savers over 2% more interest the last time auto loans were this expensive. There is approximately $1.5 trillion in auto loans outstanding, a third of which sits on commercial bank balance sheets.

When it comes to auto loans, cheap deposits do not lead to cheap debt.

Next, mortgages. This is the motherlode, since mortgage debt accounts for most of the debt taken on by consumers.

The average rate for a 30-year fixed-rate mortgage is currently around 6%. That number has come down somewhat from a few years ago, but remains above the average mortgage rate over the past twenty years. Pre-pandemic, the last time mortgage rates were this high was at the start of the Great Financial Crisis. Yields on bank savings accounts were approximately twice as high back then.

The residential mortgage market is massive, and banks account for a minority of loans, but there is still over two-trillion dollars worth sitting on their balance sheets.

When it comes to mortgages, cheap deposits don’t translate directly to more affordable loans.

We can play this game all the way on down for different kinds of credit including small business loans, but there’s actually a simpler way to test whether cheap deposits translate to cheap credit, because in situations when they don’t, the difference gets booked by banks as profit.

Thus the beauty of checking in on the net-interest income for banking today. This is an easily calculated and widely reported figure that tracks what banks pay to those who give them money (like depositors) and what they charge the people and companies they lend that money to.

If cheap deposits translate to cheap credit—as certain Wall Street executives and various bank lobby orgs have argued—then this number must be low.

It most certainly is not.

Per FDIC data, American banks pocketed approximately $750 billion in net-interest income last year. That’s three-quarters of a trillion dollars in money that banks earned by “not paying depositors much” but still “charging borrowers plenty.”

To put this number in perspective, it is more than the Mag 7 made in earnings during the same period, combined.

Nice gig if you can get it (but you can’t, since you don’t have a bank license). J.P. Morgan alone made close to $100B. If these numbers seem too good to be true, and you need further convincing, then check out what Chase pays on a savings account. And then consider what they charge on a popular credit card. That difference contributes to net-interest income.

(It also helps explain why Jamie Dimon is mad at Brian Armstrong for taking a hardline on stablecoin rewards).

Claude estimates that the GSIBs (aka Too Big to Fail Banks) accounted for a third of all net-interest income last year. That’s eight corporations making a quarter-trillion dollars while not paying savers much but also not providing cheap credit to consumers and businesses. No wonder bank stocks have been on a tear for years. (Fun fact: GSIBs always pay less for deposits than community banks thanks to the effective “infinite deposit insurance” they enjoy by way of presumed bailouts—bailouts funded by taxpayers).

A critic might argue net-interest income is simply compensation for the risks banks take, like that of default, and the operations they undertake, like underwriting bespoke loans. This is a valid counterargument.

Or rather, it would be, if the big banks did what many falsely assume they do, which is take deposits from ordinary people and make specialized loans back to other ordinary people or small businesses. That type of banking is important and deserves compensation—it’s worthy of a high net-interest margin. It’s also risky, so naturally we should want the orgs that do it to earn a high return. But it increasingly isn’t the business the banks are in, not at scale anyways.

For example, a simple thing big banks can do with customer deposits is to use them to buy Treasury bonds or Agency sponsored mortgage-backed securities. The 25 largest banks currently hold $3.5 trillion worth. These are government issued (or backed) securities that have minimal risk and require little underwriting. Of course a good chunk of it might be held for regulatory reasons or to meet capital requirements. But that doesn’t change the fact that when you lend money to a bank (by way of a cheap deposit) the bank could turn around and lend that money to the government to earn 4x as much.

Not exactly “God’s work” if you ask me, but great for bank shareholders. And banker bonuses.

Another common thing banks do with depositor funds is to just park them at the Fed. American banks currently hold $3 trillion in reserves. This is idle money that has been taken out of circulation and has no direct economic benefit—money not being loaned out. It’s also risk free. But thanks to an ouroboros monetary system, the Fed has to pay banks to take the money the Fed printed and gave them out of circulation. It currently pays a cushy 3.6%. Do things have to be this way for monetary policy to work? Perhaps. But that doesn’t change the fact that it’s yet another avenue where banks profit from cheap deposits but borrowers do not.

Ironically, both the interest on government securities that Treasury pays, and the interest on reserve balances that the Fed pays, are funded by taxpayers. So while banks pay us very little for our deposits, we pay them plenty for theirs.

I can go on, but I think you get the picture. Regardless of whether we look at the cost of different kinds of credit, or simply measure net interest income, the data tells a clear story: cheap deposits are just as likely to support bank profits as they are cheap credit. In the current rate environment, they tilt towards profits.

Now throw in the fact that virtually everyone is a saver, but not everyone is a borrower, and what you end up with is a great deal for bank shareholders, but not the rest of us.

We don’t know if yield-bearing stablecoins would actually deplete deposits or force banks to pay more yield to savers—anyone who claims certainty on this issue should not be trusted. Stablecoins are a new primitives and there are variables (such as offshore demand) that could lead to adoption increasing bank deposits.

But even if stablecoins competed with domestic deposits, it’s far from clear that borrowers would feel a material difference. Shareholders could feel the squeeze, but this is America. We don’t pass laws that outlaw competition and make consumers worst off to pad corporate profits. Or at least we shouldn’t.

Based on this math, I have a basic proposal for how to resolve the logjam over the stablecoin rewards issue in DC. If stablecoin issuers aren’t allowed to pay any, then banks should be subject to a windfall tax on net interest income. Why give them an exclusive right to cheap deposits if they won’t pass the savings on to borrowers?

To be clear I am not for such a tax. I am a capitalist and believe in competition being the thing that keeps profits reasonable, not regulation. But Wall Street doesn’t believe in capitalism. It certainly doesn’t every few years when it holds its hand out for taxpayer-funded bailouts, and it isn’t now in trying to outlaw innovation and consumer choice.

Live by the sword…

What I can offer is a literary explanation. Whatever the cause of the price declines, it makes for a great second act in the yet-to-be-written story of this industry, particularly if the third act is widespread adoption and success.

For the uninitiated, the second act of a classically structured story is both the longest and the most contentious. It is triggered by an “event” at the end of the first act and designed to explore the central tension of the story: loyalty vs. betrayal, hope vs. despair, life vs death. The second act also usually ends with a crisis, one that has an existential feel to it.

Sound familiar?

For the unfamiliar, Bitcoin is now trading below where it was before the US elected a “Bitcoin President.” Most other coins are far below their 2021 highs, despite a five-year period when most other assets like stocks and metals have skyrocketed. Industry leaders are quitting “to explore other technologies” and predictions of further collapse abound. We passed the Genius Act but may never get Clarity (or even clarity). Tokenization is finally happening, but reformed bros wonder out loud whether the winner will be Wall Street. Robinhood is launching its own layer-2 and Blackrock is entering DeFi, but nobody cares.

Things are so dark that people actually think permissioned blockchains—the stupidest idea in crypto other than algorithmic stablecoins—may actually work.

All of this makes for one hell of a redemption arc, if it ever happens. Imagine going from here to Bitcoin being a backup-reserve currency, Ethereum the global settlement layer, DeFi the core of finance, and stablecoins a disruptive force in payments and banking. Then imagine the industry finally cracking the code on DePIN, DeSo, and even Web3 (whatever that means).

Wouldn’t that be something? The fact that we went through the current era of fear, uncertainty, and doubt, would make those accomplishments that much sweeter. From FUD to Fortune has a nice ring to it.

Of course it’s also possible none of this happens and the story ends in tragedy. Many do. It’s possible that prices keep falling, the industry disintegrates, and the likes of Roubini and Krugman are proven right. But that doesn’t make for a good story. Experts Proven Right, If Early sounds rather lame if you ask me.

I for one am still hopeful. First (and foremost) because I think crypto solves important problems. Second, because I am a fan of both markets and literature, and have seen firsthand how much markets love a good story, especially ones with a wicked twist that proves most people wrong.

In the summer of 2008, the story in commodities was Peak Oil, or how production had already peaked and prices would only ever rise. It was a widely popular thesis that got increasing headline treatment as the price of oil almost tripled to $148 a barrel on the NYMEX. Four months later it was down to $32, thanks to a historic financial crisis that crushed demand. Prices would slowly recover over the next few years, then fall again (and stay down) as the shale boom proved Peak Oil laughably wrong.

How’s about that for a story?

But wait, because it gets even better. I was a trader back then but never believed in Peak Oil. So I shorted it after prices doubled to $100. And I naturally added to my short once they hit $125. So of course I got margined out at $145. I knew what it meant too, in real time. After I hung up with my broker (we still used phones back then) I called my buddy and told him to short, the top was in. Two days later it was.

There are many such stories across different markets, particularly during transitional periods. The year 2008 was also the peak of a historic and catastrophic housing bubble. Or was it? Take a look at this chart and you tell me. A lot of stupid ideas burned up during the dot com crash, ruining fortunes. Now all those ideas are wildly successful and some of the people who believed in them just a few years later are billionaires.

Men make plans, God laughs. So do markets. Equity volatility was dead in 2017. The S&P had its longest stretch without a 5% correction in half a century and the VIX made an all time low. There was talk of “a new normal” as once maligned vol selling strategies become widely popular. Then we had something called Volmageddon and the VIX exploded at record pace. A short vol ETF literally went to zero.

And you know what would have been a highly successful trade to put on the day after? Short vol. It worked well all the way to the COVID crash, at which point the right thing to have done smack dab in the middle of the greatest economic and employment collapse in history was to buy stocks. And Rolexes. And Bitcoin.

Now, guess which has performed the best since then. I’ll give you a hint, it ain’t a Submariner. Or the QQQ.

I can go on, but you get the picture. Most people filter their views through the lens of prices, but markets have a great sense of humor. And they love a great story. The simplest explanation for why crypto has crashed in the early days of 2026 is because it’s about to take over, then take off.

Or at least, that would be an epic story!

For example, the FT has a solid long read on this debate today, and includes this gem from Philipp Paech, described as “associate professor at the London School of Economics and former chair of the European Commission expert group on fintech”

“It’s not just a question of protecting an old business model . . . It is the question of whether we depend on the financial system. Stablecoins cannot finance our economy, because they are not allowed to lend to anyone”

So, the US government doesn’t count as a borrower? Even though it’s literally the single largest borrower on earth (and in history)?

The Federal government currently owes $40 trillion, money it borrows via the bond market. Stablecoins issuers lend most of their money directly to the largest borrower out there. This frees other capital to lend to others, and lowers the risk-free rate against which all other debt is benchmarked.

Next, the article cites Hilary Allen, professor at American University law school:

“When the panic comes and people are trying to exit out of these products and exit out of stablecoins, that’s where you’re going to have a run on the Treasury market”

This argument is doubly flawed. First, stablecoins don’t face the same run risk that traditional banks do because they are narrow banks. They can shrink their balance sheet as needed without gating deposits—unlike levered banks. Depositors intuitively know this. That’s why they run on traditional banks in the first place. It’s also why they won’t run on a properly regulated stablecoin.

Second, banks also hold treasuries. As a group, they are among the largest non-sovereign holders of US government debt. Why? Because capital requirements mandate them to do so, in part so they always have something liquid to sell (or borrow against) in the event of a run, which traditional banks are always more at risk of than narrow ones.

But Allen’s framing is even more inaccurate than that, because a Genius-regulated stablecoin issuer can only hold short term US debt, whereas banks often hold longer duration notes and bonds.

T-bills are both more liquid than the mix that banks hold and have little duration risk. We saw how deadly the latter could be to a traditional bank with the collapse of SVB—one of the largest bank failures in US history.

And yet here we are just a few years removed, and an otherwise intelligent academic is arguing “we need to make sure depositors are forced to use banks just like SVB, It’s giving them a fully-reserved alternative that doesn’t face the same duration and run risk that threatens the financial system.”

Lastly, Allen also worries about monetary policy. The potential growth of stablecoins, she argues:

“potentially disintermediating the banks, it’s potentially dislocating the availability of credit and then of course you’ve got more money outside of the banking system. That implicates central banking policy”

Funny how $5T moving into money market funds over the past 25 years didn’t do any of this, even though they are structurally very similar to stablecoins.

Also, we are just coming off an insane experiment (from a historical perspective) with zero and negative interest rates imposed by central banks. One of the overarching lessons was that the classic view of how rate policy translates through a levered banking system is wrong.

Part of the problem here is that levered banks don’t pay much interest in the first place. When interest rates were zero, Bank of America paid small businesses nothing. Now that rates are close to 4%, Bank of America still pays small businesses nothing (proving, as I wrote earlier, that BoA doesn’t care about them).

Stablecoin issuers will likely pass on all of their economics, one way or another. T-bill rates are tied to Fed Funds. If millions of depositors migrate from bank accounts to stablecoins, they’ll be holding an instrument that is likely more sensitive to monetary policy, not less.

Central banks will have more control, not less. At least as far as monetary policy is concerned. They will have less control over stablecoins when it comes to supervision and bailouts, but that’s because stablecoin balance sheets are more easily monitored than bank ones, and stablecoins are less run prone.

To be fair, any migration from bank accounts to stablecoins will have important impacts on the treasury market, monetary policy, and supervision. If the migration is long enough, they all will have to be evolved. And there are other—far smarter—academics working on these issues.

But that’s true for any kind of financial innovation. The problem with the stablecoin debate is that many academics seemed married to an overly-simplistic and often wrong bank-centric view of the world. Their biases fail a basic smell test of the first principles of finance that apply to any kind of instrument.

I’ve already written on how Bitcoin has discredited most of academic econ, since the prevailing view on campus remains that this multi-trillion dollar asset is either worthless, a fraud, or unimportant. We can now add stablecoins to this list.

As proof, he cites the destruction Tether and Circle could bring to a protocol like Aave—which is definitely a concern. He raises important issues and everyone should read his argument.

But here’s why I ultimately disagree and believe something closer to the opposite is likely to happen. A credibly neutral blockchain and sufficiently decentralized DeFi ecosystem will force asset issuers and FinTechs to be more cypherpunk than they’d want to be.

The reason why is moats. Or when it comes to most crypto businesses, the lack thereof.

My mental model of how crypto evolves increasingly revolves around what can be replicated and what cannot.

To start with an extreme example: Bitcoin can’t be replicated. Its superpower is having been invented at a time almost no one cared about it. A competing PoW currency invented today would immediately be ruined by farmers, investors, and regulators.

Something like a non-custodial wallet (such as Metamask) can easily be replicated. I think the wallet landscape will look drastically different in five years.

Ethereum is also very hard to replicate. Anyone can launch a new PoS smart contract platform of course, and many smart people have, but none come even close to Ethereum’s culture, leaders, token distribution, client diversity, and brand. Ethereum also had the luxury of being early, and there’s a reason why all the newer L1s try to win on features Ethereum doesn’t care about.

Aave is easier to replicate than Ethereum, but still hard as far as DeFi protocols are concerned. It has a proven track record and stable governance (recent controversies notwithstanding). It has survived bear markets and flash crashes without any major hiccups. There’s a reason nobody even tries to compete with it head on.

Stablecoin issuance, on the other hand, is easy to replicate. Narrow banks are all the same, to paraphrase Tolstoy, and laws like the Genius Act will drive even more conformity. Tether might have a liquidity advantage on some CEXs, and USDC might be popular in DeFi, but liquidity and volume are more fleeting than culture and history.

We’ve already seen this with perp exchanges, where market share has changed violently every few years. I suspect stablecoins will see something similar, particularly as the makeup of crypto flips from retail to institutional.

It’s also worth noting that Tether and Circle are corporations, whereas Ethereum and Aave are protocols. Corporations have fewer moats than protocols. We still use TCP/IP half a century later, but many web companies have come and gone during that time.

Gabe is correct that Circle freezing the USDC address on Aave would be catastrophic for DeFi. But it would be even more catastrophic for Circle, because nobody would use USDC in DeFi ever again. Such an extreme move might even lead to a run on USDC everywhere, killing Circle altogether.

Who wants a stablecoin that so wantonly hurts its own users? Particularly in a tokenized world where there are many identical products? Indeed, if Circle were to ever self-immolate by attacking DeFi in such a brazen way, it would open the door for one of its competitors to brand itself as the “neutral stablecoin for DeFi.”

But that’s why Circle wouldn’’t do it.

The thing many forget is that stablecoin issuers and other tokenization firms are for-profit entities. They need customers to succeed, and those customers are likely to be active in DeFi, because it’s objectively better. The other thing many forget is that the public nature of blockchains like Ethereum and protocols like Aave means that there will always be more competition for issuers than in TradFi. A simple baseline for any potential winner would be “don’t break DeFi.”

To be clear, these constraints don’t work on corporate databases masquerading as blockchains, like the one the NYSE and DTCC want to build on. Those networks will enshrine the issuers they like and have little competition. But that’s just another reason why they’ll fail.

I’m glad that Gabe brought this issue up, because centralized and off-chain assets will always present a form of tail risk for Ethereum and DeFi. We’d all be better off if we had more natively issued assets. But on a long enough timeline, I am confident crypto will have more of a lasting impact on Fintech than the other way around.

If public permissionless networks do re-architect the financial system, as I believe they will, then moats and business interests will keep the more centralized participants in that ecosystem honest. And if they don’t, then this whole debate was moot in the first place.

“And so when you think about that, that takes lending capacity out of the system. And that is the bigger concern that we’ve all expressed to Congress as they think about this, is that if you move it outside of the system, you’ll reduce lending capacity of banks that particularly hurt small medium-sized businesses…”

I was intrigued by the small business claim because that cohort is indeed uniquely dependent on bank credit. So I did this whole analysis on the current size of the small business loan market, the excess reserves that banks keep at the Fed, and their ability to withstand deposit losses and still cater to small businesses. Then I threw it out.

Not because the analysis wasn’t valid, I may still post it later. I threw it out because as part of that research I found a simple fact about Bank of America that proves Moynihan is full of it. His bank doesn’t remotely give a damn about small businesses. You could even argue it actively hurts them.

Here it is:

All of the debate and discourse around stablecoins and yield centers on the lending part of banking. That’s odd, because banks also borrow money. In fact they are the largest borrowers in the economy, currently borrowing $18T per Federal Reserve data. They borrow this money from me, from you, and from small businesses.

Our bank accounts are loans to the banks. If we didn’t lend them our money, they wouldn’t exist. Sometimes, they pay us interest on these loans. Increasingly, they don’t. And in the specific context of BoA and small businesses, they literally pay nothing. Bank of America borrows money from America’s small businesses for free.

I don’t know about you, but I think that’s bullshit.

Maybe you think I’m cherry picking here and just picked the simplest type of low dollar account to demonstrate them paying a pathetic 0.01% - the lowest allowable digit before zero. You are correct, people who hold deposits of a minimum of $20k for three months earn more interest. They are considered “Gold” clients and get an “interest rate booster.”

How much more?

Those who deposit over $100k—which is a lot of money for a small business—are in the Platinum bucket and get a bigger boost.

These numbers are so pathetically low that they are an insult to our collective intelligence. Bank of America pays nothing to borrow money from small businesses, regardless of how much it borrows.

We live in a time of high inflation. The least BoA could do is help small businesses keep up by paying something closer to it, say 2%. That rate would still be significantly lower than what it earns by holding T-bills or parking its money at the Fed. The bank would make plenty of money.

But it doesn’t.

Now, here’s where things get really interesting, because Moynihan was talking about lending to small businesses. Alas, most small businesses never take out a loan. I must have had a dozen different small businesses over the years and never borrowed a dollar. Think of your local restaurant, the bodega, your friend’s e-commerce startup, a freelance graphic designer’s solo business. All have bank accounts, almost none take out loans.

Thus the most likely relationship BoA has with any small business in America is to borrow money from it. For free. The bank only ever lends to a small fraction of them, and when it does, it most certainly doesn’t do it for free. The average small business loan in America currently costs anywhere from 6 to 11 percent interest, or 600x to 1100x what Bank of America pays to borrow money from the same exact company.

This of course only applies to the small businesses that apply for a loan and get it. Many don’t. If these businesses really need credit, their only recourse is to use a credit card. BoA currently charges 16% to 25% for those loans. That’s a nice business if you could get it.

A skeptic could argue that nobody has to keep their money at BoA or any of its competitors, the fact that people (and small businesses) do must mean that they get something in return.

To those people I say: Bank of America got $50 billion in direct bailout money after the 2008 financial crisis. It got over $100 billion in asset guarantees to buy Merrill Lynch. It got to participate in other Fed bailout programs designed specifically for big banks, to the tune of over a trillion dollars. It got arguably even more taxpayer money during COVID. It has privileged access to government-run payment systems. It can park endless excess reserves—reserves that it got from the Fed’s prolific printing—back at the Fed to earn 3.65%, risk free.

No small business got all of that. In fact no small business gets any of that, ever. And now, if Bank of America and the other big banks get their way, small businesses won’t be able to get interest on their stablecoins, either.

Decades later, I’m past the point of expecting anything different from the CEO of a big bank. Exploiting their privileged position in society to make great profits then having the gall to act like they are doing us a favor is what they do. I’m just disappointed so many other journalists, academics, and now members of Congress, fall for this nonsense.

Savers are people too. In fact they are the largest group of people, far larger than borrowers. So are people who need to make payments, payments that preferably don’t take 3 days and cost endless fees.

It’s time to stop screwing them.

And don’t worry, Brian Moynihan will be fine, he made over $35 million last year, and I’m certain his savings account pays a decent rate of interest.

The banking industry calls this the “stablecoin loophole” but Congress was clear on its intent as far as yield was concerned in the Genius Act. It’s also odd to call a practice that benefits ordinary Americans and puts more money in their pockets as they face affordability challenges a loophole.

That said, too much of this debate is derailed by unsubstantiated fears of what greater stablecoin adoption could mean for banking and credit. Here are the five biggest myths and why they are fals.

Myth #1: Stablecoin growth can only lead to shrinking bank deposits.

False. These aren’t always substitutes and stablecoin growth may actually increase U.S. bank deposits.

So far, stablecoins have increased domestic bank deposits. We know this because:

1. Most of the demand has come from abroad, likely from people who can’t get dollars otherwise.

2. All major issuers back their tokens with a mix of t-bills and bank deposits (Genius will further mandate this).

Put these two factors together and you end up with a scenario where every additional dollar of stablecoin issuance leads to more deposits at banks.

Even the portion of issuers’ reserves that consists of Treasury bonds will increase bank deposits. Dealing in government securities requires substantial banking activity, for purchase and sale, repo, FX, and so on. More stablecoins means more stablecoin-related securities transactions, which means more deposits at banks.

This shouldn’t come as a surprise to anyone. That stablecoins would lead to greater dollar adoption is a major reason why Genius found bipartisan support in the first place. More dollars held abroad means more deposits at U.S. banks. Demand will be even higher if stablecoin holders could earn rewards, leading to yet more deposits.

A skeptic might argue that this is only true so long as stablecoins don’t take off domestically. But that view is shortsighted because stablecoins are global. More demand at home means more liquidity, more awareness, greater innovation, greater adoption abroad.

In summary: Stablecoins increase demand for dollars everywhere, reward-bearing ones even more so. This might lead to higher bank deposits for the foreseeable future.

Myth #2: Domestic stablecoin adoption will hurt the banking industry’s ability to provide credit.

False. Competition for deposits doesn’t hurt bank lending. It hurts bank profitability. These are very different issues.

Banking in America is a highly profitable industry—more so today than during any other time in recent memory. Bank stocks were among the best performers last year. Net-interest margins—the difference between what banks pay depositors and charge borrowers—are at historic highs and rising.

All banks have to do to fight deposit competition—not just from stablecoins, but from anything—is pay savers a little more interest. This is the sort of thing that every other industry does regularly. Doing so might impact bank profits, but it doesn’t have to impact credit availability or the cost of borrowing.

Per an analysis conducted by the Financial Times, the banking industry pocketed an additional $1 trillion in net-interest profits during a recent two-year period. J.P. Morgan alone may have made $100 billion last year. That’s a lot of firepower to compete with whatever third-party rewards stablecoins pay consumers.

Also, American banks currently keep close to $3 trillion in reserves at the Fed, near a record amount, and more than they need to meet capital requirements. This is idle money that just sits there and earns them a cushy rate of return with no risk. This money could easily be tapped to offset any deposits that do leave due to stablecoins, allowing banks to provide the same amount of credit. The ongoing deregulation of banking means they’ll need these reserves even less in the years to come.

I don’t know about you, but I don’t think Congress should pass laws to protect bank shareholders and executives, and certainly not at a time when their stocks are at all-time highs.

In summary: banks can compete with stablecoins by paying higher interest to their customers. They can also offset lost deposits by reducing reserves at the Fed. Both actions might diminish their profitability, but they don’t have to impact lending.

Myth #3: Banks are the most important source of credit in America and must be protected from competition.

False: Banks account for approximately 20% of the credit provided to businesses and households in America. Any reduction in bank deposits is unlikely to lead to a proportional reduction in credit.

There are various nuances to this point, but on one thing the data is clear: banks in America do not provide most of the credit that matters to ordinary people, like mortgages. Something like half of all small business loans also come from non-banks.

Why does this matter? Because even if stablecoins hurt bank deposits—which is not a given—and even if banks can’t simply raise the interest they pay savers—which they clearly can—then it still doesn’t mean stablecoin adoption leads to higher borrowing costs.

Bank lending, particularly the kind that gets funded via deposits, is simply not that important. America is fortunate to have robust capital markets and large non-bank lenders. Think: money market funds, mortgage-backed securities, private credit funds, and insurance companies.

All of these lenders—which it’s worth reiterating account for the majority of lending—likely benefit from greater stablecoin adoption, including those that pay rewards. That’s because:

1. They all get to benefit from cutting-edge innovation in payments. The savings they realize by using stablecoins, instead of slow and expensive wires, may make their credit provisioning even more efficient and cheaper.

2. Their lending is mostly benchmarked to Treasury rates. The more demand stablecoins create for U.S. government debt, the cheaper non-bank credit might become.

Widespread adoption of stablecoins is likely to reduce the U.S. government’s borrowing cost. This in and of itself is a reason to embrace them: stablecoins save taxpayers money. But beyond that, it so happens most of the credit that gets created in America is benchmarked to Treasury rates. This leads to a possible scenario where stablecoins reduce bank deposits, but simultaneously reduce average borrowing costs.

In summary: Banks don’t need to be protected because they are not as important as they once were. There are scenarios under which money moving out of banks and into stablecoins makes mortgages and small business loans cheaper.

Myth #4: Community and regional banks are particularly vulnerable to stablecoin adoption

False. It’s the large “money center” banks that are more vulnerable.

A new kind of digital dollar that was invented to improve payments competes more with the large and global payments banks—the ones that cater to very large clients—than it does a small community bank that gives loans to farmers.

Thinking otherwise fails a basic common-sense test.

First, smaller banks already pay more interest to depositors. The academic literature has consistently explained this to be a natural result of the biggest banks providing certain services—like correspondent banking—that have little competition. Less competition means less yield.

Stablecoins are first and foremost payment instruments, so they are more likely to compete with wire transfers than basic savings accounts.

Second, survey data shows that community banks skew older in their consumer client base. This makes sense intuitively: younger depositors are more likely to use fintechs, or care about the latest tech gadgets that larger banks are more likely to support.

Do we really think that a middle-aged farmer in the Midwest is going to abandon the local bank that gave him his first mortgage and switch to a cryptocurrency offered by a startup?

The only reason this myth persists is because it’s pushed by an unholy alliance of large banks trying to protect their profits and crypto startups trying to sell smaller banks their services.

In summary: It’s far from clear what kind of financial institutions would be impacted most by stablecoin adoption, but common sense dictates the large “Too Big to Fail” banks are more likely to have to compete with new payment instruments.

Myth #5: Borrowers matter, savers don’t

False: Both groups are needed for a viable banking system, not to mention a robust economy.

Barring stablecoin issuers from sharing their economics is a tacit policy of “hurting American savers in order to benefit borrowers.” This is a weird policy choice. Saving is the flip side of borrowing. It’s also a good thing for everyone to do. Saving money leads to economic prosperity and helps families weather downturns.

The so-called “loophole” that the bank lobby keeps protesting against just gives people like my mother—a retiree who lives off her savings—an added opportunity to earn. She’s very unlikely to switch to stablecoins, but why even deprive her of the option?

That innovation and competition lead to companies fighting to provide better products and services to consumers is a bedrock of American dynamism. Why should the highly profitable banking industry be any different?

Their claim that savers don’t matter in this debate is even more infuriating when you consider the how most of the interest that stablecoin issuers could in theory pay out comes from taxpayers, by way of the government bonds issuers hold.

Why would Congress even consider a rule that insists taxpayer funds can only go to bank shareholders, but not savers? I’m pretty sure this isn’t what our representatives actually intend, but all the fearmongering emanating from the bank lobby has muddled the issue.

In summary: The stablecoin yield issue impacts a lot of people. What’s bad for borrowers might be good for savers, and lending is a two-sided market.

In conclusion

It would be weird if Apple lobbied Congress to outlaw better smartphones, as opposed to just improving current model. It would be strange if Ford and GM lobbied Congress to outlaw Tesla, as opposed to building desirable EVs. It would be wrong for me to campaign for my department chair to not offer any new classes, as opposed to trying to improve my own.

Digital currencies are no different. Most of the concerns raised by the banking industry on this topic are unproven and unsubstantiated. Congress has done a great job of putting American progress ahead of corporate interests so far; it shouldn’t stop now.

The crypto industry desperately needs greater regulation to go mainstream, and the Genius Act already resolved the stablecoin issue. It’s time to move on to more important things. American banking will be fine either way.