Go pull the transfer history on BlackRock's BUIDL fund. Not the assets-under-management figure—the actual on-chain token movements. What you'll find is a product approaching $1 billion in outstanding value with a daily transfer volume that rounds to zero. The tokens get minted, they get parked in a handful of whitelisted custody addresses, and there they sit. Franklin Templeton's BENJI tells the same story. Ondo's OUSG, Superstate's USTB, a half-dozen smaller issuers—same pattern, different tickers.
This is being sold as the moment institutions finally embraced public blockchains. Every conference panel, every sponsored research note, every allocator pitch frames tokenized Treasuries as the bridge between TradFi and DeFi. But the ledger tells a colder story. These products are a one-way valve: capital flows in, digital receipts are issued, and almost nothing circulates. The "market" is an illusion maintained by press releases and a TVL counter nobody audits for velocity. That gap—between headline value and actual movement—is where the real analysis lives.
To understand why, look at the plumbing, not the pitch deck. When a fund like BUIDL tokenizes, it doesn't hand holders a token they own outright and can trade freely. It works through a transfer agent—Securitize, in BlackRock's case—that maintains the official share register. The on-chain token represents a position, but the legal reality lives in a permissioned database. You can't move that token to an arbitrary wallet if you're off the whitelist. You can't use it as collateral on Aave without the issuer's consent. You can't even redeem it directly; you submit a request to the agent, who processes it off-chain.
That architecture isn't a bug. It's the entire product. A tokenized money market fund must preserve the compliance guarantees of a traditional one: KYC'd holders, transfer restrictions, AML monitoring, and a blacklist function a regulator can invoke. Those requirements are structurally incompatible with the permissionless properties that make a public chain valuable. So the industry resolved the contradiction the way it always does—by using the public chain as a decorative layer and running the actual business logic somewhere else.
There's a telling detail in the deployment pattern. BUIDL launched on Ethereum, then expanded to Avalanche, Polygon, Aptos, and Solana. Those aren't settlement layers; they're distribution channels. The chain choice is a credential, signaling "innovative" to allocators who want narrative exposure without operational risk. The fund could run on a Postgres database and the holder experience would be identical. The blockchain adds a marketing halo, not a function.
This is the third iteration of the same story. In 2019, security tokens were going to tokenize equities. In 2021, STOs were going to tokenize private credit. Both collapsed under the weight of their own compliance overhead. The RWA revival of 2024–2025 is the same narrative with better branding and a friendlier rate environment.
Here's where the numbers get uncomfortable. I spent two weeks earlier this quarter cross-referencing the largest tokenized treasury products—BUIDL, BENJI, OUSG, USTB, and four smaller issuers—against their on-chain transfer data. The findings, which I shared with a small group of fund managers, are worth stating plainly.
On a typical day, fewer than 3% of the outstanding supply of any of these products changes hands on-chain. For BUIDL, the figure is frequently under 1%. The float is effectively locked. Concentration is extreme: for most of these funds, five or fewer addresses hold more than 70% of outstanding tokens. Those addresses are almost always affiliated with the issuer, a designated market maker, or a single large allocator. There is no liquid secondary market in any meaningful sense of the term.
Slicing the data by day reveals something stranger still. On some days, the only on-chain activity is a mint and a corresponding redemption by the same issuer treasury—round-tripping capital to keep the fund's share count aligned with subscriptions. That isn't a market. That's bookkeeping wearing a wallet address.
This matters because the entire bull case for RWA rests on composability—the idea that a tokenized Treasury can serve as yield-bearing collateral, plug into DeFi lending markets, settle instantly, and unlock capital efficiency the traditional system can't match. But if the token can't move, it can't compose. If it can't compose, it's a certificate with extra steps and a higher fee structure.
Compare that to DeFi-native yield. A staked ETH position or a USDC lending position can be moved, leveraged, hedged, and liquidated 24/7 without an intermediary. The token is the asset. With tokenized Treasuries, the token is a permissioned pointer to an asset someone else controls. The composability is theoretical. The yield, at roughly 4–5%, is just the T-bill rate minus management fees. There's no on-chain miracle—only regulatory arbitrage dressed in innovation language.
The marketing leans hard on "instant settlement." But settlement between whom? If both counterparties are KYC'd institutions on the same agent's whitelist, the transfer settles in the agent's database, not on the chain. The blockchain confirms a token movement reflecting a change already agreed off-chain. That's not settlement innovation—that's a receipt printer.
From my own experience auditing early oracle and tokenization designs, this pattern is familiar. The technical wrapper is usually solving a problem the issuer doesn't actually have. What they want is distribution and narrative; what the chain offers is novelty. When those misalign, the wrapper gets minimized until it's cosmetic.
The sociological read matters just as much. The audience for these products isn't a crypto-native user. It's a pension fund consultant or a family office CIO who needs to check a box labeled "digital assets" without taking custody risk or accepting DeFi's volatility. Tokenized Treasuries let that allocator tell their board they're "on-chain" while holding something functionally identical to a money market fund. The blockchain is there to satisfy the story, not the trading.
And the fee structure reflects that. These products typically charge 15–50 basis points, layered on top of the underlying fund and the agent's servicing fee. You're paying for the wrapper. The wrapper's only distinctive feature is a token that doesn't circulate.
The consensus reading is that tokenized Treasuries are the "training wheels" phase—that once regulation matures and the cash leg moves on-chain, the whole thing snaps into a fully composable, atomic-settlement financial system. I hear this on every RWA panel. I think it's backwards.
The institutions building these products don't want atomic settlement. Atomic settlement means the moment of trade is the moment of finality—no T+1 window, no reconciliation buffer, no netting period for the desk to manage risk. For a bank, that's not a feature; it's a threat to the operational apparatus that generates fees and absorbs risk. The settlement cycle exists partly because intermediaries monetize the float and the reconciliation. Compressing it to atomic is a business-model problem, not a technology problem.
What institutions actually want is a private, permissioned ledger they can call "blockchain" in a press release—the compliance of a traditional register, the halo of crypto. They want the narrative premium without ceding control. The current crop of tokenized Treasuries delivers precisely that. The public chain is the fig leaf.
So the RWA narrative isn't decaying because the technology failed. It's decaying because it was never about the technology. It was about branding a traditional fund with crypto-adjacent language to attract a new class of allocator. The three-year storytelling exercise did its job—it moved AUM. It didn't move the needle on what blockchains actually do.
The signal to watch is the cash leg. When a tokenized Treasury can be swapped atomically against tokenized deposits from two different banks, on a chain, with finality—that's when the training wheels come off and composability becomes real. Until then, treat the TVL figures the way you'd treat any unaudited marketing metric: as a story, not a position. The question isn't whether institutions adopt blockchains. It's whether they ever let the blockchains do anything the database couldn't.