Bequant — Institutional Trading Infrastructure & Prime Services

Exchange counterparty risk remains one of the most underpriced risks in institutional crypto. Off-exchange settlement via regulated custody is the solution most desks are missing.

The lesson nobody fully learned

FTX should have been the end of the story. A $8 billion hole in the customer ledger, a bankrupt exchange that had been acting as trading venue, custodian, lender, and clearinghouse all at once, and an entire generation of institutional allocators who'd been assured this couldn't happen. For about six months afterwards, every serious desk in crypto was having the counterparty risk conversation.

And then, for a lot of the industry, the conversation went quiet. Capital drifted back onto exchange balance sheets because exchange balance sheets are convenient. You want to trade on Binance, you need assets on Binance. You want to run a strategy across OKX, Bybit, and Deribit, you need balances sitting at all three. The operational logic pulled funds back into the same place from which FTX had just evaporated them.

Then Bybit got hit for $1.5 billion in early 2025 and the conversation restarted — with more urgency this time, because it wasn't just an insolvency, it was a live security failure on a venue most desks had assumed was safe. Sygnum's off-exchange custody volumes grew 900% year-on-year. Copper's ClearLoop network expanded. Fireblocks Off Exchange rolled out across more venues. Crypto.com went live with off-exchange settlement in March 2026. The pattern is now too clear to ignore: the desks that had done the work on off-exchange settlement before the incident were materially better off on the morning after.

If you're still running institutional size with your full trading balance sitting on exchange, you are carrying a risk that has a real probability and a catastrophic loss given default. The market has priced that correctly at least twice now.

What off-exchange settlement actually is

Off-exchange settlement (OES) lets you trade on an exchange without the exchange custodying your assets. The underlying holdings sit with a regulated third-party custodian. The exchange sees mirrored collateral credit — enough to let you trade at full size — and your actual assets never leave the custodian's segregated wallets.

When you execute a trade, P&L settles back to the custody account, not to the exchange balance. When the exchange fails, goes offline, gets hacked, or freezes withdrawals, your capital is sitting somewhere else entirely. You've separated the three things FTX had collapsed into one: price discovery on the exchange, custody at the custodian, and settlement between them on a defined legal and operational rail.

The mechanics vary by provider. Copper's ClearLoop uses a trust structure under English law with sufficient collateral maintained at the custodian to satisfy settlement obligations. Fireblocks uses Collateral Vault Accounts — MPC wallets that programmatically lock and mirror assets to the exchange. Sygnum Protect uses its Swiss banking licence. The legal wrappers and the tech differ, but the outcome is the same: trading access without custody transfer.

The risk you're actually hedging

This is worth being precise about, because "counterparty risk" gets used loosely in crypto. The specific exposures OES removes:

Exchange insolvency. The FTX scenario. Exchange goes bankrupt, customer assets are commingled or rehypothecated, you become an unsecured creditor in a multi-year bankruptcy process.
Exchange security failure. The Bybit scenario. Funds are stolen from exchange-controlled wallets through a compromise the exchange itself didn't detect in time. Whether the exchange chooses to reimburse, and how quickly, is not your decision.
Operational freezes. Regulatory action, banking partner withdrawal, API outages during volatile periods, or "temporary withdrawal suspensions" that last longer than they should.
Jurisdictional seizure. Assets on an exchange in a given jurisdiction are subject to that jurisdiction's legal processes. Segregated custody in a different legal venue breaks that link.

What OES doesn't remove: market risk, your own operational risk, smart contract risk on any DEX leg of your strategy, and custodian risk itself. You've swapped exchange counterparty risk for custodian counterparty risk. The trade only makes sense because regulated custodians — actual trust companies, banks, and licensed CASPs — have materially lower default probabilities than exchanges. Agio Ratings' Q1 2026 data puts the leading regulated custodians in the 0.39% to 0.84% range for 12-month probability of default. Exchanges, as a category, sit meaningfully higher.

You're not eliminating risk. You're moving it to a counterparty that charges less for it.

Why desks haven't made the switch

If the case is this clear, why isn't every institutional desk already running OES? Three reasons, in order of how often they come up:

Capital efficiency worries. The concern is that mirrored collateral ties up capital less efficiently than just holding the assets on-venue. In practice, modern OES networks settle in real time and support most major assets as eligible collateral — including tokenised MMFs and stablecoins that earn yield while posted. The capital efficiency loss versus exchange-held balances is marginal. The risk reduction is not.

Operational complexity. Setting up OES means onboarding a custodian, executing the legal documentation, integrating the APIs, and training the desk. It's real work — days, not hours. For a firm focused on P&L today, the project keeps getting pushed to next quarter. Until it can't be.

"Our exchanges are fine." The most dangerous reason, because it's the one that sounds most rational. Every firm that lost money on FTX or Bybit would have told you the week before that their exchanges were fine. Counterparty risk is always zero — until it isn't, and by then the withdrawal window has closed.

None of these are good reasons to carry the exposure. They're reasons the project hasn't happened yet.

What a sensible OES architecture looks like

For an institutional trading operation, the setup worth building toward has a few specific features:

Regulated custody at the core. Not a crypto-native startup with a "we custody funds" tagline — a licensed custodian subject to real supervision. Trust companies, banks, or MFSA/FCA/BaFin/FINMA-regulated CASPs. The custodian's default risk is now the residual risk in your architecture, so the quality of that custodian matters more than anything else in the stack.

Multi-venue coverage. One-venue OES is operationally fine but defeats much of the point. Institutional strategies run across exchanges; your settlement rail needs to as well. The major OES networks now cover most of the relevant CEXs.

Legal wrapper that survives a bad day. English law trusts, Swiss banking structures, MFSA-supervised segregation — whatever the jurisdiction, the critical question is: what happens to your assets if the custodian itself fails? The wrapper needs to answer that question clearly, in writing, before you sign.

Real-time settlement. P&L that sits in suspense for hours between trade and settlement reintroduces exactly the risk OES is meant to remove. Real-time or near-real-time rebalancing between exchange credit and custody collateral is the operational standard to aim for.

Integration with execution and risk. OES is not useful if it's a separate system from your trading. The collateral position, the exchange credit, the real-time margin, the hedge execution — all of that needs to sit on infrastructure that sees the full picture. Otherwise you're managing counterparty risk in one system and trading risk in another, and the gap between them is where mistakes live.

How Bequant fits into this

Bequant operates institutional prime services from two regulated bases — Bequant Pro in Malta, MFSA-regulated as a CASP with MiCA authorisation in process, and Bequant Prime in Seychelles. The DMA infrastructure and RiskQuant engine that power execution, lending, and market making also connect into the OES architecture clients need:

Custody-integrated execution. Client assets sit with institutional custody; Bequant's DMA routes orders across CEXs and DEXs against mirrored collateral rather than exchange-held balances. The same unified risk ledger that governs leveraged positions also tracks collateral utilisation across venues in real time.

Cross-venue settlement across CEXs and DEXs. Settlement logic that works on a single CEX is table stakes. Strategies running across multiple CEXs plus DEX legs need a settlement rail that spans the whole book. That's where the infrastructure investment actually pays off.

Regulated counterparty structure. For institutional allocators and token projects evaluating who to work with, the answer to "where do my assets sit and under what legal framework" needs to be clean. Operating from Malta and Seychelles with defined segregation and custody arrangements means the question has a crisp answer — not a white paper full of wallet addresses.

Prime services embedded, not bolted on. OES that sits alongside undercollateralised lending, DMA execution, market making, allocation of capital, and capital introduction is a different product from OES sold as a standalone custody tool. The same stack that hedges your counterparty risk is the stack you're trading through, financing through, and reporting from.

For desks running serious size in 2026, the operating question isn't whether off-exchange settlement is worth the setup effort. It's whether carrying full trading balances on exchange is a risk anyone on the investment committee would approve if it were presented honestly. Usually, the answer is no — and the project gets prioritised.