Katana’s Fisher says RWA utilization is mismeasured and closer to ~20%
He argues dashboards miss off-contract collateral use and that settlement speed, not composability, caps leverage loops.
Katana’s Matthew Fisher is pushing back on the popular claim that tokenized real-world assets are seeing “sub-1%” DeFi utilization, arguing the statistic is a measurement artifact. His framework puts effective utilization for assets “built to move” closer to ~20% once non-mobile supply and off-contract collateral usage are accounted for.
Key Takeaways
- Katana’s Matthew Fisher said tokenized-asset utilization looks close to ~20% after excluding non-mobile supply, adjusting for holder intent, and adding off-contract collateral usage.
- The widely repeated “sub-1% utilization” figure is framed as a market-wide read but is described as tracking only three tokenized money market funds totaling $7.2 billion with roughly $50 million deployed.
- Broader baskets produce materially higher readings, including 11.7% using DeFiLlama’s data and about ~19% from CoinShares’ $7.4 billion Q2 deposits versus RWA.xyz’s $38 billion total.
- Fisher argued the binding constraint for leverage traders is settlement and redemption timing, estimating a 4x position on a T+1 asset takes ~8 loops, or ~8 days to build and ~8 days to unwind.
Why the ‘Under 1%’ RWA Utilization Meme Keeps Showing Up
The “under 1%” line persists because it sounds like a clean verdict on tokenization: lots of assets minted, almost none doing work. Fisher’s point is simpler and more damaging to the meme. The market is arguing over a ratio without agreeing on what sits in the numerator or the denominator.
Fisher’s framing is that tokenized RWA “utilization” is being treated like a universal KPI when it is really a methodology label. He wrote that “depending on who you ask, the share of tokenized real-world assets actually being used in DeFi is under 1%, or 7%, or 11.7%, or close to 20%,” and that “none of them measures the same thing.”
That matters for traders because the utilization debate is often used as a proxy for whether RWA collateral is becoming real balance-sheet plumbing or staying a marketing layer. If the measurement is wrong, the market risks misreading adoption as stalled when it is merely happening in places dashboards cannot see.
The Numerator/Denominator Problem: $7.2B vs $50M, 11.7%, and ~19% in the Same Year
Fisher anchored the most-cited “sub-1%” figure to a narrow basket: three tokenized money market funds. He named BlackRock’s BUIDL, Circle’s USYC, and Franklin Templeton’s iBENJI, and put their combined size at $7.2 billion with roughly $50 million “deployed.” On that definition, utilization prints as tiny. The critique then gets repeated as if it describes the full tokenized RWA market.
Widen the basket and the number moves. Fisher said DeFiLlama’s data yields 11.7% utilization when more tokenized assets and venues are included. He also pointed to a separate “same market, same quarter” comparison that lands near the high teens: CoinShares’ $7.4 billion Q2 deposits count versus RWA.xyz’s $38 billion total, which he summarized as about ~19%.
The spread is the story. Under 1%, 11.7%, and ~19% can all be “right” if they are answering different questions. Fisher’s complaint is that the market keeps quoting the smallest number as if it is the definitive one.
He also attacked the denominator choice at the market level. The piece cites roughly $51 billion in tokenized RWAs on public blockchains, and Fisher cited Bernstein research that about 47% of that is private credit. His argument is that private credit is structurally a poor denominator for a composability-style metric because it “doesn’t move much in traditional finance either,” and tokenization does not change its redemption calendar or holder base.
If nearly half the market cap is an asset class that is not designed to be mobile collateral, then “how much of the $51 billion is actively deployed in DeFi” becomes a category error. It is measuring the wrong thing, then calling the result a failure.
What ‘Utilization’ Misses: Restricted Tokens, Parked Holders, and Invisible Collateral
Fisher broke low utilization into three buckets, and only one is a straightforward product failure.
First is “restricted by design.” He described early flagship products as shipping with transfer restrictions and whitelist gates that prevent them from functioning as collateral even if demand exists. He grouped BUIDL and Apollo’s ACRED into a “trophy tier” that generated headlines but arrived with enough gating to make DeFi-style reuse difficult. A participant in one such product described it to him as “a terrible product.”
Second is “parked by intent.” Fisher’s example was a foundation holding BUIDL on its balance sheet to attract BlackRock to deploy on its chain, which he characterized as “buying a headline, not a yield strategy.” In that framing, low onchain deployment is not a sign the asset cannot be used. It is a sign the holder does not want to use it.
Third is “used invisibly,” which is where the measurement problem becomes structural. Fisher listed off-contract collateral paths that do real work but do not show up in smart-contract TVL dashboards: BENJI functioning as off-exchange collateral via Binance and Franklin Templeton, BUIDL being accepted as derivatives margin, and Kraken taking tokenized equities as collateral for leveraged trades.
Those examples point to the same market-structure reality. If collateral utility is increasingly expressed through custodians and centralized venues, then onchain-only utilization metrics will systematically undercount adoption. The market will keep arguing about “idle” supply while the real activity sits in margin systems and custody agreements.
Signals to Watch: One-Shot Leverage Auctions, Stablecoin Rails, and Rulemaking That ‘Brings the Gates Down’
Fisher’s trader-relevant bottleneck is time, not just permissioning. He argued tokenized RWAs often settle T+1, T+2, or on redemption calendars, which forces leverage loops to run sequentially. His concrete example was a 4x position on a T+1 asset taking around 8 loops, implying roughly ~8 days to build and ~8 days to unwind.
He also warned that redemption calendars can turn looped positions into duration products during stress. Using Apollo’s ACRED as the example, he said he worked on bringing it onchain as collateral on Polygon PoS with Securitize, Gauntlet, and a Morpho-powered vault. He said it worked “to an extent,” but highlighted the need for on-demand liquidity for atomic redemption because redemptions are quarterly. He claimed that if a quarterly-redemption fund is looped 4x, fully unwinding can take up to a year.
One proposed path around sequential looping is to compress the cycle count. Fisher pointed to protocols like 3F, built on Morpho on Ethereum mainnet, that use an onchain auction where specialists front the full capital for target leverage in one shot, collapsing many settlement cycles into one. The text does not provide adoption metrics for 3F, so the open question is whether this structure scales beyond a concept demo.
On regulation, Fisher argued the sequencing that matters starts with settlement money. He said the GENIUS Act mattered more for enabling credible stablecoin rails than for tokenized securities directly, and he framed it bluntly: “You cannot have a collateral market without settlement money.” He also said CLARITY has stalled in the Senate, and suggested near-term progress may come via administrative rulemaking at the CFTC and SEC that reduces compliance anxiety and allows issuers to relax restrictive gates.
My Read: If the Market Can’t Measure RWA Collateral Use, It Will Misprice the Adoption Curve
The threshold that matters is not whether utilization prints at 1% or 20%. It is whether the market can separate three different states: assets that cannot move, assets that can move but are held for non-DeFi reasons, and assets doing collateral work off-contract. If those buckets stay blended, every dashboard chart becomes a narrative weapon instead of a signal.
The real test is whether settlement and redemption terms compress enough for leverage to behave like crypto margin instead of a slow duration trade. If one-shot auction leverage and stablecoin settlement rails turn T+1/T+2 loops into something closer to atomic, then “utilization” stops being a debate and starts being a risk parameter traders can actually price.
