Tokenization & RWA

What tokenization is

Tokenization is the process of representing a claim on an asset as a transferable token on a blockchain. The token is not the asset itself; it is an on-chain pointer — a record that says “the holder of this token is owed this much of that thing” — plus, crucially, a legal connection back to the real asset that the chain cannot itself enforce.

This is a different category from native crypto. Bitcoin and ether are natively digital: they exist only on their own chains, and ownership is enforced by the chain's consensus rules. A tokenized Treasury bill, by contrast, exists off-chain as a security held by a custodian; the token is a second representation of it. The chain enforces who can hold and transfer the token, but it cannot enforce that the underlying bill actually exists, that the custodian really has it, or that a court will honor the token-holder's claim. Those are off-chain facts.

The appeal is real and has little to do with crypto ideology. A token, unlike a paper certificate or a position in a broker database, is programmable, composable with DeFi, transferable 24/7 to anyone the rules allow, and dividable into fractions small enough to be useful. Whether that adds enough value to justify the legal and operational machinery it requires is the question this page keeps returning to.

What counts as an RWA

A real-world asset (RWA) is anything of value that exists off-chain: fiat currency, government bonds, corporate debt, equities, real estate, commodities, invoices, even music royalties or a share of a shipping container's cargo. The unifying property is that the asset predates and is independent of the blockchain — the token is a representation of it, not the thing itself.

RWAs fall on a spectrum of how easy they are to tokenize, and the spectrum mostly tracks how observable, liquid, and standardized the underlying already is:

• Cash and cash-equivalents — the easiest. A dollar is a dollar; a T-bill has a public price and a known maturity. This is where tokenization already works at multi-billion scale.
• Fixed income — Treasury bills, money-market fund shares, private credit, trade-finance invoices. Harder than cash but still well-defined; the largest non-stablecoin RWA category today.
• Equities — shares of a company. Legally heavy (securities law, transfer agents, corporate actions) and still mostly experimental on public chains.
• Real estate, art, and physical commodities — the hardest. Illiquid, jurisdictional, physically custodied, and valued by appraisal rather than market. Most of the “tokenize everything” pitches live here, and most have not worked at scale.

It is worth saying plainly: stablecoins are RWAs. A USDT or USDC token is a tokenized claim on off-chain dollar reserves held by a custodian. The crypto world spent years treating stablecoins as a separate thing called “stablecoins” and RWA as a newer, sexier category — but stablecoins are the original, the largest, and so far the only unambiguous success of the RWA thesis.

How it works

The pipeline from a real asset to a usable on-chain token is a chain of trust points, and each one is a place the “decentralized” framing quietly stops applying:

• The asset — the thing being represented (a T-bill, a loan, a building).
• The legal entity — usually an SPV (special-purpose vehicle), a trust, or a fund that legally owns the asset and is set up so that the token holders have a defined claim on it (see the legal wrapper below).
• Custody — a qualified custodian (a bank or trust company) or the issuer itself holds the actual asset.
• The oracle / attestation — someone reports the asset's existence, holdings, and price to the chain, either via an oracle feed or via off-chain attestations and audited reports.
• The token — the on-chain ERC-20 (or permissioned variant) that represents a share of the entity, with transfer rules coded into its contract.
• Transfer and settlement — secondary trading, redemptions, and the rails (DEXes, whitelisted pools, or off-chain OTC) that let the token actually move.

The blockchain governs only the last two links. Everything upstream — the asset, the legal entity, the custody, the reported value — is off-chain and trusted. This is the central fact of RWA tokenization: it does not remove intermediaries, it wires them into a token. Whether that is a worthwhile trade depends entirely on whether the on-chain layer earns its keep in speed, composability, or access.

The legal wrapper

The standard pattern is to put the asset into a special-purpose vehicle whose only business is owning that asset, and to issue the token as a claim on the SPV. The token holder is really a creditor, a limited partner, or a beneficiary of the SPV; the chain is a transfer ledger for those claims. This is why tokenization projects always involve lawyers, jurisdictions, and filing fees before they involve smart contracts.

The choice of jurisdiction matters enormously. Where the SPV is incorporated, what exemptions from securities registration it relies on, who can legally hold the token, and how redemptions are handled are all questions of national law. The same token can be a security in the United States, a commodity elsewhere, and an unregulated instrument in a third place. The chain does not resolve this; it simply has to be told (via the transfer whitelist) who is allowed in.

This is the part of the stack that crypto culture is most ambivalent about. Tokenizing a regulated asset reintroduces the exact trusted intermediaries — custodian, issuer, administrator, auditor, lawyer — that blockchains were built to route around. The honest answer is that for real-world assets there is no alternative: you cannot put a building on a blockchain, only a legal claim on one, and legal claims require legal machinery. The question is whether that machinery is made cheaper, faster, or more programmable by sitting on top of a token, not whether it can be eliminated.

The token standards

Most RWAs are fungible, so most use ERC-20 — the same standard that governs ordinary fungible tokens on Ethereum. But a plain ERC-20 is permissionless: anyone can transfer it to anyone. Securities and gated assets cannot work that way, so the RWA world has built permissioned variants on top:

• ERC-3643 (the “Permissioned Tokens” standard, formerly the T-REX protocol) — transfers are only valid if the recipient is on an on-chain identity whitelist managed by an identity service. This is the workhorse for tokenized securities today.
• ERC-1400 — a security-token standard that adds partitioning (separating tranches), forced asset recovery by the issuer, and document attestations. Less commonly deployed than ERC-3643 but influential in the design space.
• ERC-4626 — a standard for yield-bearing vaults: deposit an asset, receive a share token that accrues yield. Used for tokenized money-market and lending positions where the token itself should compound.
• ERC-20 + off-chain gating — many products skip the permissioned standards and use a plain ERC-20 with KYC at the on-ramp, relying on the issuer to refuse transfers to unauthorized wallets. Simpler to ship; weaker enforcement.

The standard matters more than it sounds because composability depends on it. A permissioned token will not flow into an ordinary liquidity pool or a lending protocol without breaking its transfer rules — which is the subject of the next section.

Permissioned vs. permissionless

This is the deepest tension in RWA tokenization, and it is structural, not a bug to be fixed. Securities law demands gated transfers: issuers must know who holds their token (KYC), screen for sanctioned counterparties (AML), and often restrict resale to verified or accredited investors. DeFi's entire value proposition is that it is permissionless — anyone can supply, borrow, or swap against any pool. The two are in direct conflict.

The result is that most real RWAs are permissioned tokens: only whitelisted addresses can receive them. That is the correct legal choice, but it severs the token from the thing that makes on-chain assets interesting. A tokenized T-bill that can only move between whitelisted wallets cannot be posted as collateral in a normal DeFi lending market, cannot sit in a Uniswap pool, and cannot be composed the way a stablecoin or a liquid staking token can.

The emerging middle path is “permissioned DeFi”: pools and lending markets that themselves check an on-chain identity whitelist, so permissioned tokens can be used composably but only among whitelisted parties. It works, but it narrows the audience to the same set of institutions that could have transacted off-chain anyway. The honest framing is that RWA composability is real but gated, and a gated marketplace is a smaller and less liquid one than the permissionless ideal DeFi was built around.

Stablecoins: the original RWA

Stablecoins are the proof that RWA tokenization works, and they are the benchmark every other RWA category is measured against. A stablecoin is a tokenized claim on off-chain fiat — usually dollars — held by an issuer/custodian. USDT and USDC together hold tens of billions in reserves and clear more daily volume than most of the rest of crypto combined. They are the single largest, most liquid, and most useful category of real-world asset on-chain, and they got there first because dollars are the easiest possible thing to tokenize: a unit is fungible, its price is fixed, and “redemption” is a wire transfer.

Even here, though, the link to the real asset is a custodian and attestations, not the chain. The chain records that address X holds 1000 USDC; it does not, and cannot, know that Circle actually has $1000 in a reserve account backing it. That knowledge comes from audited attestations and regulatory oversight off-chain.212 The 2021 New York Attorney General settlement with Tether and the CFTC order the same year were reminders that the integrity of the largest RWA in crypto rests entirely on off-chain reporting.177

The lesson the rest of the RWA market took from stablecoins is mixed. Stablecoins proved that the mechanics — issue a token, hold a reserve, let it circulate — work at scale. But they also proved that the hard part is not the token, it is the reserve: trust in the issuer, quality of the collateral, and the ability to redeem. Every other RWA category inherits exactly that dependency, usually with a harder underlying asset than dollars.

Tokenized treasuries & money funds

After a decade of false starts, tokenized government debt is the category that finally broke through. The products are simple in concept: a fund holds short-dated U.S. Treasuries, repos, and cash, and issues an ERC-20 token representing a share. Holding the token gives you exposure to T-bill yield — around 4–5% for much of 2023–2024 — without leaving the blockchain. Issuers include Ondo (USDY, OUSG), BlackRock (BUIDL), Franklin Templeton (BENJI), Hashnote, and Superstate.

Why this category and not equities or real estate? Three reasons line up. First, T-bills are the most observable, liquid, and standardized asset short of cash itself — the underlying has no pricing or custody ambiguity. Second, 2022–2024 took short rates to multi-year highs, so for the first time in a decade T-bill yield beat what DeFi lending and staking were paying; crypto-native holders wanted that yield without off-ramping through a broker. Third, institutions that were already comfortable with crypto infrastructure wanted a blockchain-native way to hold cash-equivalents on their balance sheets.

The tradeoff is the same as everywhere in RWA: these are securities, so transfers are gated to whitelisted (and often accredited) addresses, and the token is really a fund share with a custodian and an administrator behind it. The yield is real; the “on-chain” part is a settlement and transfer layer on top of an ordinary money-market fund, not a replacement for one.

Private credit on-chain

Private credit is the largest non-stablecoin RWA category by on-chain value, and it works differently from treasuries. Instead of a fund holding a public security, a pool of stablecoin holders funds loans to real-world borrowers — small-and-medium businesses, invoice financing, real-estate bridge loans, trade finance. Protocols like Centrifuge, Maple, Clearpool, and Goldfinch tokenize each loan pool; depositors receive a token representing their share of the pool's repayments.

The mechanics are appealing: stablecoin capital that would otherwise sit idle earns a spread, and borrowers get access to dollar funding that may be cheaper or faster than their local banking options. The yield, often mid-single to low-double digits, has drawn real demand.

But the risk profile is fundamentally an off-chain lending risk, not a crypto risk. A borrower defaulting on an invoice or a bridge loan is an off-chain event the chain cannot prevent, predict, or enforce against — it falls to the originator, the servicer, and the legal system. Several on-chain private-credit pools have taken losses when underlying borrowers defaulted, exactly as an off-chain lender would. The token does not diversify that away; it just makes the position transferable.

Real estate, art & commodities

These are the hardest RWAs, the ones that drove the original “tokenize everything” pitch, and the ones that have delivered the least so far.

• Real estate — the oldest tokenization idea: split a property into a thousand tokens and let anyone buy a slice. The appeal is obvious; the friction is everything else. Property is physically custodied, jurisdictionally tied, valued by appraisal, and legally expensive to transfer. Most attempts have produced a small number of demonstration properties with thin secondary markets.
• Art and collectibles — fractionalizing a painting or a watch overlaps with NFTs; the token represents a share of a physical asset held in storage. Valuation is subjective, custody is physical, and the secondary market is thin — the same problems as real estate, more acute.
• Commodities — the cleanest example is tokenized gold (PAXG, Tether Gold): one token equals one troy ounce in a vault, redeemable. Gold is fungible, widely priced, and cheap to store, so it works where art and real estate do not. It is still a small market compared to stablecoins.

The pattern across all three is the same: tokenizing an illiquid, physical, appraisal-valued asset does not make it liquid. The token is transferable, but if no one wants to buy the underlying, there is no buyer to transfer it to. For these assets the on-chain layer mostly adds cost and complexity on top of an unchanged, illiquid market.

The oracle & pricing problem

Every RWA token needs a price — the net asset value of what it represents. For a T-bill or a stablecoin that is easy: the price is public and continuously observable, and a price oracle can report it with little ambiguity. For private credit, real estate, art, or an invoice, the price is a model, an appraisal, or a servicer's report — an off-chain input that can be stale, wrong, or manipulated.

This is the oracle problem applied to asset valuation rather than to market prices, and it is harder in one specific way: a manipulated price feed can be detected and ignored, but a wrong appraisal of a building or a defaulted loan portfolio may simply be the only information anyone has. The chain will faithfully record whatever value the issuer reports; the chain cannot audit the building. RWA pricing therefore inherits the credit risk and the reporting honesty of the off-chain parties, and a token whose NAV is whatever the issuer says it is has a single, very human point of failure.

The mitigation the market has converged on is attestation: independent auditors, fund administrators, or on-chain reserve attestations that the holdings match the tokens. It helps, but it replaces “trust the chain” with “trust the auditor” — a return to the very intermediaries the technology was supposed to make optional.

The liquidity question

The most over-promised property of RWA tokenization is liquidity. The pitch says tokenization turns an illiquid asset into a liquid, 24/7, globally tradeable one. In practice, tokenizing an asset makes the token transferable; it does not create buyers.

Stablecoins and tokenized treasuries are genuinely liquid because their underlyings are liquid and because there is broad, symmetric demand to hold dollars and T-bills on-chain. For real estate, private credit, and art, the secondary market is thin: issuance is small, buyers are concentrated among the same institutional set, and AMM pools do not help much when the asset is gated, low-volume, and hard to price. A token in a market with no buyers is just a faster way to discover that no one wants your asset.

The honest read is that tokenization adds liquidity where some already exists (treasuries, stablecoins) and adds very little where there is none (real estate, art). The “global 24/7 market” for a fractionalized warehouse is, for now, mostly a marketing line. That could change if a deep, whitelisted institutional buyer base emerges, but that is a demand-side problem, not a token-design one.

Regulation & securities law

Most RWAs are securities under the Howey test: an investment of money in a common enterprise with an expectation of profit derived from the efforts of others.202 A tokenized fund share, a slice of a private-credit pool, and a fractional real-estate token all fit that description, which means they fall into the same regulatory framework as the equivalent off-chain product.

In practice that means one of a few exemption paths in the U.S.: Reg D 506(c) (sales to verified accredited investors only), Reg S (offshore, to non-U.S. persons), or Reg A+ (a mini-IPO with lighter reporting). Each comes with transfer restrictions, caps, and disclosure obligations. Other jurisdictions have their own regimes, and a product that is compliant in one is often not in another.

The chain's role here is narrow and mechanical: enforce a whitelist of allowed holders and a transfer lock-up if the exemption requires one. It cannot change who counts as an accredited investor, cannot make a security into a non-security, and cannot pick the jurisdiction — jurisdiction follows the investor, not the token. Issuers who treated tokenization as a way around securities law have been the ones who ran into enforcement; the ones who treated it as a new settlement layer on top of an ordinary securities offering are the ones shipping product today.

Risks & limitations

• Off-chain / counterparty risk. The custodian, the issuer, the administrator, the auditor, and (for private credit) the borrower are all off-chain trust points the chain cannot see. A tokenized T-bill inherits the failure modes of the fund holding it, plus the failure modes of the token contract on top.
• Smart contract risk. The token itself is code; a bug in the transfer logic, the whitelist, or the vault can lock or misallocate funds. A smart contract audit is necessary and not sufficient.
• Oracle / valuation risk. For anything but cash and public securities, the reported NAV is an off-chain input that can be stale or wrong, and the chain has no independent way to check it.
• Liquidity risk. Secondary markets for most RWAs are thin. Being able to transfer a token does not mean being able to sell it at a fair price, or at all.
• Legal and regulatory risk. Exemptions can be revisited, jurisdictions can reclassify instruments, and a product that is compliant today may not be after a rule change. The law, not the contract, defines what the token is.
• The “decentralization theater” risk. A permissioned token issued by a single entity, redeemable only through that entity, with a price set by that entity, is in practice a database entry with a blockchain UI. The token may still be useful — cheaper settlement, programmable transfers — but it is not decentralized, and calling it crypto does not change that.

History & milestones

The state & road ahead

RWA tokenization is the bridge between traditional finance and DeFi, and the state of that bridge depends a lot on which asset you look at. Stablecoins already proved the thesis at scale for the easiest asset (dollars). Tokenized treasuries and private credit are proving it for the next tier — observably priced, well-defined, with genuine institutional demand. Real estate, art, and the rest of the “tokenize everything” pitch remain mostly aspirational, blocked less by technology than by the fact that the underlying assets were illiquid and hard to price before anyone minted a token.

The optimistic read is that tokenization finally gives DeFi access to the vast pool of off-chain yield — Treasuries, private credit, commodities — and gives traditional issuers a faster, programmable settlement rail, and that the two converge over time. The pessimistic read is that the institutional version is mostly a permissioned database with a blockchain fee attached, and that the genuinely decentralized, composable version is blocked by the same securities law it always was.

The most likely truth is both at once, and which one dominates depends less on the token standards than on three off-chain questions: whether institutions adopt the rail in volume (they are starting to), whether permissioned DeFi can build enough liquidity to matter (small so far), and whether regulators let the composability layer grow or push tokenization back into closed, whitelisted systems. The technology is not the bottleneck; it has not been for years. The bottleneck is the legal wrapper, the custody, and the demand — exactly the parts that are not on the chain.

For now the practical upshot is simple: if you want on-chain exposure to dollars or T-bill yield, stablecoins and tokenized treasuries deliver it, with the usual caveats about who holds the reserve. If you are looking at a fractional token in a building or a painting, treat it as a small, illiquid, gated position in a hard-to-price asset — which is what it was before it was a token, and what it still is after.