Real World Assets On-Chain

The asset managers and banks are not running small experiments. They are moving their products on-chain and clearing significant volume through them. BlackRock's tokenized money market fund, BUIDL, holds $2.4 billion and filed two new fund structures with the SEC in May. Franklin Templeton runs its government money fund on-chain as BENJI; Fidelity, VanEck, UBS, and WisdomTree have each launched tokenized money market or Treasury funds, with WisdomTree's cleared by the SEC for intraday trading in February. Apollo brought private credit on-chain through ACRED, and Goldman Sachs and BNY Mellon built shared tokenized money market infrastructure. JPMorgan's blockchain unit, Kinexys, now processes more than $5 billion a day across its payment and tokenization rails.

The market structure underneath those products is moving on-chain too. In the first quarter of 2026, Nasdaq advanced plans to integrate tokenized securities into regulated market infrastructure, while Intercontinental Exchange, the NYSE's parent, is working with OKX on tokenized stocks, and Robinhood has begun distributing tokenized shares to users outside the U.S. Even the institution at the center of the system is responding: in December 2025, after the SEC granted no-action relief, the DTCC, which settles the vast majority of U.S. securities transactions through its depository, the DTC, announced a partnership to tokenize DTC-custodied Treasuries on the Canton Network, with a production pilot targeted for 2026.

To be clear, the existing system works. For large, liquid, domestic securities it's fast, cheap, and reliable, and the rails aren't broken. They were built for a specific set of conditions: domestic markets, business hours, large institutions, standardized instruments, and they strain wherever those conditions don't hold. The open question is who builds the infrastructure the migration runs on and where the durable value will accrue.

Everything we back comes down to one question: what does it cost two parties who don't fully trust each other to transact, and who keeps the margin that cost creates?

In capital markets that margin is large. It pays for the brokers, clearinghouses, custodians, transfer agents, and depositories that stand between a buyer and a seller precisely because no single ledger is trusted by all of them. The $17 to 24 billion the industry spends every year just on trade processing, the back-office and middle-office work of matching, confirming, and reconciling trades, is the most measurable piece of that margin, the cost of trust, billed back to the people transacting. It exists only because separate ledgers must be verified against one another.

Tokenization attacks that cost two ways. It compresses rent: when issuance, transfer, and servicing collapse into a smart contract, the reconciliation those intermediaries were paid for disappears. And it serves demand the old system couldn't reach. It reaches buyers across borders that were previously unserved, and it can give a private fund stake, normally locked for years, a secondary market where an LP can exit before the fund winds down. When both mechanisms hit the same asset, you get new buyers, not just a cheaper version of the market that already existed. The infrastructure we back lowers the cost of trust for the whole market, and no single player can own it. That is the bet in everything we invest in, and this is what that looks like in capital markets.

Tokenization means representing ownership of an asset (a Treasury bill, a share, a loan) as a token on a blockchain, so ownership lives on a shared, programmable ledger instead of in a custodian's database. While that sounds minor, it changes four things about an asset, and leaves three important ones untouched.

Continuous settlement. Markets run a six-and-a-half-hour weekday session and settle a full business day after the trade, in a world that trades around the clock. Extended-hours volume is already more than 11% of U.S. equity activity, and foreign investors who own close to 18% of U.S. equities are often awake when those markets are closed. On-chain settlement finalizes in seconds, around the clock, with no dependence on banking hours or correspondent chains. For a trader, it closes the T+1 gap and frees the margin posted against trades that haven't settled. A business built on instant settlement carries none of the buffers and workarounds that T+1 forces on everyone else, and that edge holds until the other side rebuilds on the same rails.

Programmable ownership. This is the property with no clean traditional analogue. When an asset is a programmable token, it can do several jobs at once inside a single transaction, with no chain of intermediaries to coordinate. A tokenized Treasury can sit as collateral against a stablecoin loan while still accruing its yield, and the position that creates can be put to work again elsewhere, all without the asset passing through an intermediary's books. Doing the same off-chain means a separate agreement and settlement cycle for each use, and usually moving the asset to another party along the way. That composability is why an asset on-chain can be worth more than the identical asset off-chain: you can do more with it.

Cost compression. Folding issuance, transfer, and servicing into smart contracts strips out layers of intermediaries. State Street estimates 40 to 50% lower issuance costs for investment-grade bonds, and in repo and collateral the savings are already concrete, with GFMA putting them at $150 to 300 million a year for a bank running $100 billion in daily repo. Those savings aren't fully visible yet: a 2026 ECB study found tokenized bonds lowered borrowing costs and improved liquidity but showed no clear operational cost savings at today's scale. The bigger effect is that once issuance gets cheap enough, asset classes that were never worth the overhead, like trade-finance receivables, emerging-market debt, and fractional real estate, become economical to serve.

Global access. Tokenization doesn't only make existing markets more efficient; it creates new buyers. A saver in Lagos or Jakarta who wants U.S. Treasuries or blue-chip stocks today has to clear correspondent banks, currency controls, and local brokers that often don't carry the product. Tokenization widens the addressable base for any asset to anyone with a wallet, at minimums as low as a few dollars, and tokenized equities are already held across more than 350,000 wallets, most of them investors who could never have bought those stocks through a domestic broker.

That is the bull case. Here is what tokenization does not change:

It does not create liquidity. A blockchain supplies the rails for trading, not the buyers and sellers. You can list a tokenized vineyard on an exchange, but that doesn't mean anyone wants it. A token trades only as actively as the asset behind it. Tokenized Treasuries trade deeply because they mirror one of the deepest markets on earth, while tokenized real estate, despite heavy issuance, shows very little secondary activity.

It does not remove credit, legal, or counterparty risk. A government bond on-chain is still bound by its repayment terms; a private loan on-chain is still bound by the borrower's cash flows and the chance of default, and when default comes, enforcement still runs through off-chain courts. Maple learned this in 2022, losing $36 million to a borrower default whose workout took months, because a real-world claim can't be liquidated with a line of code. Synthetic tokens that merely track a price expose the holder to the issuer's solvency on top of the asset's own risk. Tokenization improves transparency, but it does not make a bad asset good.

It does not, by itself, mean ownership. Most tokenized assets today are indirect claims: the token represents a claim on a fund or an SPV that holds the asset, not the asset itself. The structures sit on a spectrum: a synthetic token that only mirrors a price; a token backed one-to-one by shares a custodian holds on your behalf; and, rarest, a token recorded directly on the issuer's share register that makes you a shareholder of record. The gap is widest in the fastest-growing category: most tokenized equities today are synthetic or custodial exposure, not registered shares. Who this matters to depends on the holder. For retail buying through a broker it changes little, since they already hold in street name rather than on the register. It matters to anyone who needs direct title: an institution whose mandate requires segregated assets, or a holder who would rather own the share outright than hold a claim against the custodian or SPV. That last case is the one that matters most, because if the intermediary fails, a custodial or synthetic holder is left with a claim against it while a registered holder owns the asset directly, outside the intermediary's estate. The gap between exposure and ownership is the thing most investors underweight, and where much of the risk in this market sits.

Tokenization, then, is a real infrastructure upgrade, cheaper, faster, more global, but not alchemy: it doesn't manufacture demand, erase risk, or confer rights a legal structure withholds. So if the technology has been possible for years, why now?

None of this is a new idea. The first wave of tokenization arrived around 2017 and mostly failed, not because the technology didn't work but because the conditions around it weren't there. Three of those conditions are now in place.

Regulatory clarity. For years the binding constraint wasn't engineering, it was law: nobody could say whether a token was a security or whether a regulated institution could even touch it. That has changed. The GENIUS Act became federal law in July 2025, giving the U.S. its first statutory framework for stablecoins; the CLARITY Act, which splits oversight between the SEC and CFTC, passed the House in 2025 and cleared the Senate Banking Committee in May 2026, where it now awaits a floor vote; and the SEC under Chair Paul Atkins has reversed its enforcement-first posture, issuing new guidance and a token taxonomy. Europe's MiCA is already in force. A regulated institution can now look at this and know which rules apply.

Institutional permission. Until last year, an accounting rule known as SAB 121 effectively barred banks from custodying crypto by forcing them to carry customer assets against a full capital reserve. The SEC rescinded it in January 2025, and through 2026 regulators cleared the remaining barriers. Major banks are now building custody businesses of their own.

Infrastructure maturity. The plumbing that didn't exist in 2018 is now production-grade. Custody is regulated and insured at institutional scale, oracles bring prices and NAVs on-chain reliably, high-throughput chains have pushed transaction costs to cents, and compliance now lives at the token level, with allow-lists and transfer rules enforced by the contract itself so a regulated asset stays compliant as it moves between venues. None of this existed at the start of the last cycle.

With the rules in place, the institutions permitted, and the infrastructure built, the question shifts to which layers of this new stack capture the value.

If tokenization is mainly an infrastructure shift, the question for an investor is which parts of that infrastructure are hard to copy and which get competed away. Bringing a real-world asset on-chain takes six steps, from sourcing the asset to getting the token into a buyer's hands, and each step is its own business with its own economics.

Asset origination and underwriting. Someone still has to source the real asset and price the risk. This is the part blockchain changes least, since deciding whether an asset is sound takes human judgment and is not something you can program into a smart contract. The contrast is clearest in lending versus funds: more than $20 billion in loans now run through platforms like Figure and Maple, where sourcing and judging the loan is the real work, while for Treasuries the originator is just the asset manager tokenizing a fund it already runs. The businesses that last here are the ones whose underwriting is hard and whose deal flow is proprietary. Of the six layers, this is the one whose moat comes from judgment rather than technology.

Legal structuring. The token has to map to a real legal claim, an SPV, a trust, a fund wrapper, that defines what the holder actually owns. This is where the exposure-versus-ownership question gets settled, and it's the layer most investors skim past. The structuring work itself commoditizes: law firms and SPV factories drive the price down, and most issuance platforms fold it in anyway, with Securitize the largest for institutional funds. What doesn't commoditize is the legal architecture that becomes a standard, the wrapper other issuers copy because it's been tested and accepted. A standalone structuring shop competes on price with no real moat, and the standard-setting upside usually lands with the issuance platform that absorbs it.

Issuance. The most built-out layer: minting the token, running the primary sale, and keeping subscriptions, redemptions, and the cap table on-chain. Securitize administers the most assets; among the platforms building the next generation, Superstate, a 1kx portfolio company, is an SEC-registered transfer agent whose USTB fund holds about $967 million and is now managed by Invesco, and it issues other managers' funds through its FundOS platform, including for Coinbase Asset Management and Bitwise. Its Opening Bell product records public equities directly on the issuer's share register rather than as a backed claim, one of the few routes to real on-chain ownership instead of exposure. Issuance has durable economics (fees on assets administered, plus switching costs once a fund's cap table lives on your rails), but it drifts toward winner-take-most as standards converge. The defensible version pairs minting with something stickier: transfer-agent status, the register itself, or distribution.

Taken together, these first three layers are starting to fold into one. The platform that structures the wrapper increasingly mints the token and runs the fund administration too, so origination, structuring, and issuance often reach the market as a single integrated stack rather than three separate vendors. How far they collapse depends on the asset. For a money market fund, where origination is just a manager tokenizing a fund it already runs, the three are effectively one. For private credit, where sourcing and judging the loan is the hard part, origination stays its own business while structuring and issuance merge.

Compliance. Keeping a regulated asset compliant as it moves is two distinct jobs, and on a permissionless ledger both have to be built in rather than bolted on. The first is real-time enforcement. Predicate, a 1kx portfolio company, lets an issuer define rules such as sanctions screening, jurisdiction limits, and transfer restrictions, and enforce them at the smart-contract level, so a non-compliant transaction can't execute, and is already used by RWA-focused chains like Plume. The second is the back office. Cryptio, also in our portfolio, is the accounting and reconciliation system of record, turning on-chain activity into audit-ready books for institutions including SG Forge, Circle, and Gemini, having raised a $45 million Series B and processed more than $3 trillion in volume. Compliance is the most defensible layer in the stack: once an institution's policy engine and reporting are wired in, replacing them is painful, and the regulatory surface only grows.

Data and oracles. A token depends on off-chain facts, and something has to bring them on-chain reliably. While a market price is the simplest, a fund's NAV is more difficult. Behind the single number are choices that vary by issuer: whether positions are monitored in real time, how often the NAV is struck, and whether an independent party verifies it. Then there is proof that the underlying assets exist, the reserve attestation that separates a backed token from an unbacked claim. RedStone, a 1kx portfolio company, is the specialist for institutional NAV feeds, focused on the yield-bearing collateral that is hard to price well: more than $6 billion in total value secured across more than 100 chains, with feeds powering tokenized funds from BlackRock, Apollo, and VanEck. Oracles have real network effects and high switching costs, since they sit in the path of every transaction. They're also a prime target for attackers, since one bad feed can drain a protocol, so reliability is the moat. The layer is defensible through network effects, and a specialist can still dominate a segment general-purpose oracles serve poorly.

Distribution. The last mile: where an investor actually buys, holds, and uses the token. It's the most contested layer and the most valuable, because it owns the customer, and it runs through several channels that are easy to lump together but behave differently. Exchanges and brokerages like Coinbase, Kraken, and Robinhood reach retail and compete on listings and reach. Wallets serve people who custody their own assets, and increasingly sit embedded inside fintech and payment apps, where the product reaches the user directly. DeFi protocols like Morpho and Aave act as distribution channels too, giving a tokenized Treasury somewhere useful to sit as collateral. Above these sit vaults and on-chain asset managers, which aren't a separate channel so much as a vehicle: they package tokenized assets into strategies and push them out through the same wallets and exchanges to reach end users. Network effects are strong within each channel, which is why issuers, exchanges, and asset managers are all fighting to own the customer relationship. Of the six layers, distribution has the most distinct ways to win: each channel owns a different customer, so value spreads across several players rather than consolidating to one.

Which layer matters the most depends on the asset class.

The layer that looks commoditized for one asset is the bottleneck for another, so where we deploy capital is a judgment about layer and asset class simultaneously.

The stack points us toward the layers that are hard to replace, and away from the ones that aren't. But there's a harder question underneath, and it's the one that most worries us about backing crypto-native companies: as tokenization grows into a real share of capital markets, why wouldn't the incumbents (the NYSE, Nasdaq, the DTCC, the big banks) move to own the whole stack themselves?

We don't have a guarantee against this, and we treat it as the central risk to the thesis. The outcome we think most likely, though, is that incumbents capture part of the stack and rent the rest, and which part is fairly predictable. They are likely to keep the customer-facing layer: the venue, the brand, the client relationship, the DTC's place at the bottom of settlement. We don't expect a startup to out-distribute the incumbents or displace the DTC, and we don't bet that way. But the rails underneath that layer are usually rented rather than built. BlackRock's BUIDL carries BlackRock's name and BlackRock's assets, but it runs on a tokenization vendor's infrastructure rather than something BlackRock built.

Those rails are where we think the opportunity is, and they are hard for any one incumbent to own, for two reasons. A competitive incumbent can't: a bank won't run its compliance on a competitor's engine, and the NYSE won't price its tokenized assets off an oracle Nasdaq controls. A neutral incumbent like the DTCC could, since it competes with no one, and that is the case we watch most closely. But a neutral utility is built to run market infrastructure under its members' oversight, not to ship and iterate software, which is why it has operated on technology it didn't write for decades. The evidence fits: the DTCC built its tokenized-Treasury system on the Canton Network with Digital Asset and adopted Chainlink's standards rather than writing either itself. Invesco, Bitwise, and Coinbase Asset Management rent the same way on Superstate's rails.

We focus on three layers an incumbent can use but not own: compliance, data, and the issuance infrastructure that asset managers build on rather than compete with. These are also where the two mechanisms from the start of this piece get productized. Compliance compresses rent, turning the duplicated, manual work of screening and reconciliation into a protocol that runs once. Data and oracles are enablement, bringing valuation and proof-of-reserve on-chain so a tokenized asset can do things it never could in a custodian's database, like serve as live collateral. Issuance infrastructure does both, folding minting, transfer, and servicing into software while making true on-register ownership possible for the first time. A company in any of these can sell to every incumbent at once without threatening any of them, and that position is hard to dislodge once institutions depend on it. Whether that produces large independent companies or just acquisition targets is a fair question, but the record points to real independent scale: Circle, Fireblocks, and Chainlink have all gotten large without an incumbent owning them. We've invested in infrastructure the incumbents use, not banks or exchanges competing with them.

Distribution is the fourth, and we approach it differently. It's the most valuable layer and the most contested, and the one place we don't look for neutrality, because here owning the customer is the end goal. We won't back a venue going after the same retail customer as the NYSE or Coinbase head-on. The edge we look for is in DeFi, which we think gains more from tokenization than any other distribution channel. An exchange or a brokerage lets an investor buy and hold a tokenized asset; a DeFi protocol lets them use it, as collateral, as yield, as a building block in another position, the programmable-ownership property that makes an asset worth more on-chain than off. As tokenized Treasuries and credit move on-chain, the venues that can do something with them rather than only list them are the ones that gain the most. That is where distribution gets interesting to us, and will be the subject of its own thesis down the line.

The migration is real and underway. Tokenization makes settlement faster, ownership programmable, and access global, but it doesn't create demand, remove risk, or hand value to whoever mints the token. Value accrues to what is hardest to replace: the neutral rails everyone has to trust and no one can own, and the venues that can put a tokenized asset to work rather than only list it. The rails are compliance, data, and the issuance infrastructure asset managers build on. That is where we invest, and where we think the durable companies of this cycle will be built.

If you're building the rails beneath tokenized markets, or the venues that put those assets to work, we'd like to hear from you.