Why 24/7 Tokenized Stocks Gap — and Liquidate People in Their Sleep

Tokenized stocks are one of the more genuinely useful things to arrive on-chain: a token like tAAPL or tTSLA that tracks a real equity and can be held, moved, or borrowed against at any hour. But that around-the-clock availability hides a structural mismatch that quietly liquidates people. This post explains the mismatch in plain English, shows why ordinary lending designs are fragile because of it, and lays out the two things that actually help — conservative, gap-aware loan limits and a risk layer that watches positions continuously.

The core mismatch: the token trades 24/7, the market doesn't

A tokenized stock can change hands at 3 a.m. on a Sunday. The company behind it cannot. The New York Stock Exchange and Nasdaq are open roughly 9:30 a.m. to 4:00 p.m. Eastern, Monday to Friday, with holidays off. That is about a third of the week. For the other two-thirds, the real price discovery that sets what Apple or Tesla is worth simply isn't happening.

So during the hours the underlying market is closed, an on-chain tokenized stock is trading against stale information. When the market reopens and absorbs everything that happened while it was shut — an earnings miss, a macro shock, a weekend headline — the price can jump in a single step rather than drift there smoothly. That single step is called a gap.

Why gaps specifically wreck loans

Lending against a volatile asset works on a simple promise: the collateral is worth comfortably more than the debt, and if it starts slipping toward the debt, the loan is unwound in small, orderly steps before it goes underwater. That promise quietly assumes prices move continuously — that you get a chance to act between "healthy" and "insolvent."

Gaps break that assumption. The price can cross the danger zone in one move while the borrower is asleep and the underlying market is closed. By the time anything can respond, the position may already be past the point where it should have been protected. This is not a tail risk you can engineer away; it is a property of the asset. Tokenized equities will gap. The only real questions are how much room you leave for it, and what is watching when it happens.

A concrete example

Suppose a protocol lets you borrow up to 80% of your collateral's value — generous, the kind of number that looks competitive in a table. You post $1,000 of a tokenized stock and borrow $800. On Friday afternoon everything looks fine. Over the weekend the underlying drops 15% on news. Monday's open reprices your collateral to $850. Your $800 debt is now backed by $850 — and after the liquidation penalty, you're effectively underwater. You never had a chance to add collateral or repay, because the move happened while the market was shut. A naive 80% loan didn't fail because the borrower was careless. It failed because the design assumed a world without gaps.

First answer: conservative, gap-aware loan limits

The first defense is unglamorous and the most important: leave real room. Two numbers govern this.

• Maximum loan-to-value (LTV) — the most you can borrow against your collateral when you open the loan. A 50% max LTV on $1,000 of collateral means you can borrow at most $500.
• Liquidation threshold — the point at which the loan is eligible to be unwound. It sits above the max LTV; the space between the two is your safety buffer, the room a gap can move into before anything is forced.

Most DeFi lending was tuned for crypto assets that trade continuously, so it could afford thin buffers. Tokenized stocks need the opposite: limits set deliberately low so that a realistic overnight or weekend gap lands inside the buffer instead of blowing through it. Tessera's parameters are tuned for exactly this — and they differ by asset, because a single stock gaps harder than a broad index.

Read tTSLA as the clearest statement of intent: you can borrow at most 40% of your collateral, and nothing is liquidated until 55%. That 15-point gap is room left on purpose for a weekend that goes wrong.

When a position does have to be unwound, the mechanics are kept orderly too. A liquidator earns a 5% base bonus (which ramps with depth, so distressed positions still get cleared), only 50% of a position can be closed in one step under normal conditions — a full close is reserved for positions that fall below a health factor of 0.95 — and there's a 100 USDC minimum debt so positions never fragment into uneconomical dust. None of this prevents a gap. It limits the damage when one arrives.

Second answer: a risk layer that never sleeps

Conservative limits buy room. They don't buy attention. The other half of the problem is that gaps happen precisely when no human is watching — overnight, over weekends, over holidays. So the protection has to be continuous and automatic, not a dashboard you're expected to babysit.

This is the role of the Watcher, Tessera's autonomous risk agent. It re-checks every borrower's health factor about every 10 seconds, around the clock. As a position drifts toward danger it sends plain-English alerts — not a wall of numbers — so you understand what's happening and why. And if you've opted in, it can auto-repay from USDC you pre-approved to pull a position back from the edge before it crosses the liquidation line, without you having to wake up and do it yourself.

An agent that can move money is only trustworthy if its limits are real and enforced by the chain, not by good intentions. So the Watcher is boxed in on every side:

• It never custodies your funds. It can only reduce your own debt, and only through an allowance you set and can revoke at any moment. Revoking that allowance is the single kill switch.
• An AI model never decides to move money. A deterministic core makes every financial decision from on-chain state; the language model only writes the human-readable alerts. The copy is AI; the choice to act is not.
• Hard on-chain caps bound it. Auto-repay is limited to 25,000 USDC per user per day and 10,000 USDC per user per transaction — enforced by the contract, not by the agent's own restraint.
• It can't disappear silently. The agent stamps an on-chain heartbeat, and a permissionless backstop is built so that — once enabled — anyone can step in if the agent goes quiet past the configured delay, meaning the system isn't dependent on one piece of software staying alive. On testnet the delay is 0 (backstop off, agent-only); enabling it is a mainnet gate. There is also an on-chain admin fail-safe.

Why both halves are needed

Conservative limits without a watcher leave you safe but unattended: the buffer is there, but nothing acts inside it until it's already breached. A watcher without conservative limits leaves you attended but exposed: something is paying attention, but a single gap can outrun any response because there was no room to begin with. The structural answer to 24/7 tokenized-stock lending is both at once — wide, gap-aware buffers that give a position room, and a continuous, bounded agent that uses that room before the market reopens and reprices everything.

Tessera: the safest place to borrow against tokenized stocks.