Credit that follows the market underneath.
Tokenized equities are a $2.8B market that mostly sits idle. Not for lack of demand. The reasons are structural, and both of them are fixable.
Where this stands
Worth saying first, because everything below reads differently without it. Acre launched at the end of August 2026. It holds a few hundred dollars, almost all of it our own. It has not been independently audited. Supply and borrow caps are deliberately small and will stay that way until it has been.
What follows is an argument about a market and a plan for addressing it, not a claim about a position we already hold.
The gap
Roughly $2.8B of tokenized equities exist across chains, held by more than a million addresses. Almost none of it is doing anything. That is not a demand problem: people who hold appreciating assets have always wanted to borrow against them rather than sell.
Two things stop it, and they are different in kind.
| Barrier | Consequence |
|---|---|
| Risk models built for continuously traded assets | Lending against them is either mispriced or avoided |
| Holders in apps that cannot call contracts | Excluded from every lending market by construction |
The first is a modelling failure. The second is a plumbing one. Acre addresses both, and they need entirely different work.
Pricing the calendar
Every lending protocol assumes it can sell collateral to cover a position that has gone bad. For a governance token that holds: there is a book at three in the morning on a Sunday. For a tokenized share of Apple it is false for sixty hours every week.
A tokenized share trades continuously. The equity underneath it does not. Between the Friday close and the Monday open there is no exchange to liquidate into, no price discovery, and no way for a borrower to add collateral if the underlying gaps. Protocols in this category inherit a risk model designed for assets that trade every second of their existence, and apply it to assets that do not.
Acre prices that directly. Borrowing capacity and rates follow the trading calendar of the asset underneath, tightening through Friday afternoon and across the weekend, and widening again at the open. Capacity also tightens around a company's earnings date, when single-name gap risk is highest.
One decision worth stating because it is counterintuitive: the liquidation threshold does not move with the clock. Only borrowing capacity does. A threshold that tightened at the close would mass-liquidate every borrower at four o'clock on a Friday, which is precisely the outcome the design exists to prevent.
Collateral is also scored as a portfolio rather than a list. Holding Apple, Microsoft, Alphabet, Amazon, Meta and NVIDIA is one bet on large technology, not six positions, and treating it as six would hand a concentrated portfolio the capacity of a diversified one.
The access problem
The second barrier is less discussed and probably larger. Most tokenized equities are held in consumer apps: embedded wallets, gasless trading, one-tap onboarding. Those apps can send a token. They cannot call an arbitrary contract.
Every lending market, including this one until recently, requires an approve followed by a deposit. That is two contract calls, and it excludes the majority of the supply in the category before anyone reaches a question about rates.
Acre now works by plain transfer. Every account has addresses derived from its own: send the loan asset to one and it becomes a supply position, send your position tokens back and the underlying returns, send the loan asset to a third and your debt falls. No connection, no approval, no contract call.
There is deliberately no borrow address. These addresses cannot tell who sent to them, which is exactly what makes them work, and for borrowing that property would be dangerous: anyone could send a trivial amount and put somebody into debt they never asked for. Lending works this way. Borrowing needs a wallet that can sign.
What is built
All of the following is deployed on Robinhood Chain. Addresses are listed in the documentation.
| Component | What it does |
|---|---|
| Session-aware lending market | Supply, borrow, repay and withdraw with terms that follow the trading calendar |
| Eight collateral assets | AAPL, MSFT, GOOGL, AMZN, META, NVDA, TSLA, SPCX, scored across correlation groups |
| Risk tier framework | Assets listed against published parameter sets rather than hand-tuned individually, with per-asset exposure caps |
| Partial liquidation | Timed auction in slices, with a discount that starts at zero and rises, rather than seizing a whole position at once |
| Auto-deleverage | A position reduces itself before the close, funded from a buffer the borrower sets aside |
| Deposit by transfer | Lending without a wallet connection, for holders in apps that cannot call contracts |
| Defined outcome pools | Fully collateralised protection: a floor under an asset for a fixed window, with someone else earning the premium |
| Supply rebate | A share of the protocol fee returned to suppliers holding ACRE, funded from the reserve rather than from emissions |
What comes next
In order of what unblocks the most, rather than what is most interesting to build.
An independent audit
The single largest constraint. Everything is covered by our own test suite and a set of stateful invariants, but those were written by the same people who wrote the contracts, which finds a great deal and does not find what an adversary finds. Until an audit exists the caps must stay small, and while they are small nobody can deposit meaningfully. This is a spending decision rather than an engineering one.
Higher caps, in steps
Raised against time running without incident and against audit findings being closed, not against a date. The current figures bound what a contract bug can take from anybody, which is the entire reason they exist.
Fixed-rate, fixed-term borrowing
Variable rates are a convention of decentralised lending rather than a preference anyone expressed. Somebody borrowing against Apple for three months should be able to know what it costs before they start. Nothing in the category offers this today.
More collateral, faster
The tier framework reduces listing an asset to a single decision against a published parameter set, recorded on chain for anyone who wants to argue with it. That is what makes going from eight assets to fifty a routine operation rather than fifty separate private judgements.
Earnings-aware capacity, extended
Capacity already tightens around a company's report. The same machinery can price other scheduled events where gap risk is concentrated and where a continuous market does not exist to absorb them.
What could go wrong
A plan that only describes the good outcome is not a plan.
The category might not grow. Tokenized equities may stay a curiosity. Acre is a bet that they do not, and being early to a market that never arrives is indistinguishable from being wrong.
The contracts are unaudited. They may contain a fault our tests do not reach. The caps limit the damage; they do not eliminate it, and nobody should deposit more than they would accept losing outright.
The approach is copyable. Nothing here is patented. A larger protocol could price the trading calendar too. The advantage is having built it first and understanding why each decision was made, which is real but not permanent.
Access may stay closed. Deposit by transfer works for lending. It cannot enable borrowing safely, and whether the apps holding most of this supply ever allow contract calls is not something we control.