Stablecoins Are Becoming Shadow Banks — Without Looking Like Banks

Your wallet shows a very simple number:

1,000 USDC

It feels like $1,000.

You can send it across the world, move it between exchanges, use it inside DeFi, or transfer it at a time when ordinary banks are closed.

But there is no digital stack of Federal Reserve notes sitting inside Ethereum.

The token is only the visible end of a much larger financial machine.

Behind that $1 price sit reserve assets, banks, custodians, Treasury markets, institutional redemption channels, exchanges, and arbitrageurs.

The strange part is not that stablecoins imitate dollars.

It is how much machinery is required to make one token keep looking like one dollar.

The Dollar Is Somewhere Else

The simplest mental model of a fiat-backed stablecoin goes like this:

You give the issuer one dollar. The issuer puts that dollar somewhere safe. You receive one token. Later, you hand the token back and recover the dollar.

That picture is directionally useful, but the reserve is usually not a giant checking account full of idle cash.

Major issuers hold reserves in combinations of bank deposits, short-dated US Treasury securities, overnight Treasury repo, and other highly liquid cash-equivalent structures ^[6].

So the better mental model is:

token on-chain, reserve off-chain, redemption connecting the two.

The token can move thousands of times without any Treasury bill moving with it. What matters is that enough credible reserve assets sit behind the system and can support redemption when tokens eventually come back toward the issuer.

Your Token Pays 0%. The Reserve Does Not.

This creates one of the more unusual business models in modern finance.

Suppose an issuer supports $100 billion of stablecoins and much of the reserve earns roughly 4% in short-term government securities.

$100,000,000,000 × 4% = $4,000,000,000

That is $4 billion of gross annual interest before custody, compliance, distribution agreements, operating expenses, taxes, and other costs.

The holder of the basic payment token may receive none of that reserve yield directly.

At first, that sounds like a terrible deal.

But a stablecoin holder is often buying something other than an investment return.

They are buying liquidity.

A trader may want a dollar-like asset that can move between markets at 2:00 AM. A business may want blockchain settlement. A user in another country may value access to dollar-denominated purchasing power more than the interest on a small balance.

The economic exchange is therefore more interesting than "issuer wins, user loses."

The reserve generates income for the issuer and its distribution ecosystem. The token gives the holder a portable form of dollar liquidity.

So Why Is It Worth Exactly $1?

This is the part the wallet interface hides best.

Circle does not sit inside every exchange forcing USDC to trade at exactly one dollar. Tether does not manually move every USDT liquidity pool back to par.

There are really two markets.

In the primary market, eligible institutional customers can mint or redeem with the issuer under its rules. In the secondary market, everyone else trades tokens on exchanges, OTC desks, and decentralized liquidity pools ^[1].

PRIMARY MARKET

fiat ──► issuer ──► stablecoin
stablecoin ──► issuer ──► fiat


SECONDARY MARKET

buyers ⇄ exchanges / DEXs ⇄ sellers
             ~$1

Now suppose a stablecoin falls to $0.995 on the secondary market.

If an arbitrageur believes it can buy the token for $0.995 and reliably redeem it near $1 through the primary channel, the difference is an opportunity.

Buying below par pushes secondary demand upward. Redemption removes tokens from circulation. The gap tends to narrow.

The reverse can happen when a token trades above $1: qualified participants may mint near par and sell into the premium.

There are fees, timing constraints, banking risk, eligibility rules, and market risk, so this is not a free-money machine.

But the mechanism explains something important:

The peg is not embedded in the token.

It is an arbitrage relationship backed by confidence that redemption will work.

Why People Call This Shadow Banking

The phrase shadow banking is provocative, and it needs some care.

Stablecoin issuers are not simply commercial banks wearing crypto logos.

They do, however, share some economic features with money-market and other non-bank financial structures: they issue highly liquid, money-like claims; manage pools of short-term financial assets; depend on confidence in redemption; and can face run-like pressure if that confidence breaks ^[5].

The analogy also has limits.

A conventional bank may fund mortgages, business loans, and other long-duration credit with deposits. A major fully backed payment stablecoin is generally supposed to hold liquid reserve assets instead of running a large long-term lending book.

That difference matters.

The risk is less "your dollar funded a 30-year mortgage" and more:

Can this reserve be turned into redemption cash quickly enough, through functioning institutions, when everybody wants out at once?

That is closer to a liquidity-management problem than classic bank maturity transformation.

The Day a Bank Broke a Blockchain Dollar

March 2023 gave the market an unusually clean experiment.

Silicon Valley Bank failed.

Circle disclosed that about $3.3 billion of USDC reserves — roughly 8% of the reserve at the time — was held at SVB and had not been withdrawn before regulators took control ^[1].

Nothing was wrong with Ethereum.

USDC transfers still worked.

The token contract had not been hacked.

But suddenly the market had to ask a much older financial question:

Can the issuer actually get to the dollars behind the claim?

Primary-market redemptions were suspended over the weekend, uncertainty spread through the secondary market, and USDC traded as low as roughly $0.87 on some venues ^[1].

After US authorities announced that SVB depositors would have access to their funds, the uncertainty eased and USDC returned toward its peg.

The episode exposed the whole system in one shot:

A token on a decentralized blockchain lost its dollar peg because a conventional bank in California failed.

Backed Is Not the Same as Liquid

The SVB episode also shows why saying a stablecoin is "100% backed" is not the end of the analysis.

There are at least three different questions.

Solvency: Are there enough reserve assets to cover the outstanding claims?

Liquidity: Can those assets be converted into usable redemption cash fast enough?

Access: Are the banks, custodians, payment rails, and redemption channels actually operating when the market needs them?

A system can look solvent on a spreadsheet and still experience a market panic if access to the reserve becomes uncertain.

And that uncertainty matters because arbitrageurs do not trade against accounting statements alone.

They trade against the probability that redemption will work when they try it.

Then Crypto Started Buying Treasuries

Stablecoins have grown large enough for the reserve side of the system to matter beyond crypto.

The Federal Reserve put aggregate stablecoin market capitalization at roughly $320 billion in spring 2026 ^[7].

That means reserve management is no longer a niche problem.

Tether alone reported roughly $141 billion of direct and indirect US Treasury exposure in early 2026 ^[8].

That produces a wonderfully circular result:

people want digital dollars
        ↓
stablecoin supply grows
        ↓
reserve assets grow
        ↓
reserve managers buy short-term Treasuries
        ↓
US government debt supports more digital dollars

Crypto was often imagined as an escape from traditional finance.

One of its largest products now creates structural demand for one of traditional finance's most conventional assets: short-term US government debt.

The Dollar Escapes the Bank Account

The other macroeconomic consequence happens on the user side.

A dollar stablecoin can circulate outside the normal interface of a US bank account.

For people living with high inflation, capital controls, weak banking access, or repeated currency depreciation, that can be economically meaningful.

The user does not need to care about Treasury auctions or reserve funds.

They may simply want something denominated in dollars that can be held in a wallet and transferred across a public network.

This has echoes of older offshore-dollar systems such as the Eurodollar market, but the analogy should not be pushed too far. Stablecoins are technologically and institutionally different.

The useful similarity is simpler:

A dollar-denominated claim can circulate outside the domestic US banking interface while still depending heavily on US financial assets underneath.

That can widen access to dollar savings. It can also accelerate dollarization and make monetary policy harder for countries whose residents increasingly prefer a foreign unit of account.

Regulators Followed the Reserve

Once stablecoins became large enough, regulators increasingly focused on the same hidden machinery that keeps appearing throughout this article: reserves, custody, redemption, disclosure, and liquidity.

In the United States, the GENIUS Act, signed into law in July 2025, created a federal framework for payment stablecoins, including one-to-one reserve requirements using specified liquid assets and public reserve-disclosure requirements ^[4].

In the European Union, MiCA similarly establishes rules around issuance and redemption for relevant stablecoin categories and prohibits issuers from granting interest on covered e-money and asset-referenced tokens ^[3].

The details differ across jurisdictions and product types.

But the direction is revealing.

Policymakers are not regulating the token because it is an interesting piece of software.

They are regulating the financial machine behind it.

What If the Holder Wants the Yield?

The non-yielding payment stablecoin creates an unusual split.

The holder gets liquidity and transferability.

The reserve generates yield elsewhere.

That naturally raises another question:

Why not give the yield to the holder too?

Crypto already has products that try to do something like this through tokenized Treasury funds, synthetic-dollar structures, and yield-bearing wrappers.

But those products should not simply be called "better stablecoins."

Once a token passes investment income back to the holder, its economics — and often its legal treatment — begin to look different from a plain payment token.

The distinction matters precisely because the zero-yield stablecoin is doing a specific job.

It is trying to be useful as money-like infrastructure, not necessarily as an investment account.

The Token Hid the Machinery

A stablecoin succeeds when the user does not have to think about any of this.

The wallet says:

$1

Behind that one number may sit Treasury securities, repo markets, bank deposits, custodians, institutional redemption agreements, market makers, exchanges, regulations, and global demand for dollars.

The blockchain did not make those institutions disappear.

It gave them a cleaner interface.

That may be the strangest thing about stablecoins.

They look like crypto-native money, yet their stability depends on some of the oldest machinery in modern finance.

So where does the crypto system actually end — and the traditional dollar system begin?